7.1.0.11

* fix tests on ghc 8.10.7

* fix test

* fix schema dump test

---------

Co-authored-by: Evgeny @ SimpleX Chat <[email protected]>

.github/workflows/build.yml

@@ -9,6 +9,10 @@ on:
       - "v*"
   pull_request:
 
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: ${{ !startsWith(github.ref, 'refs/tags/v') }}
+
 jobs:
 
 # =============================
@@ -62,8 +66,19 @@ jobs:
   build:
     name: "ubuntu-${{ matrix.os }}-${{ matrix.arch }}, GHC: ${{ matrix.ghc }}"
     needs: maybe-release
+    permissions:
+      contents: write
+      packages: write
     env:
       apps: "smp-server xftp-server ntf-server xftp"
+      opt: ${{ github.event_name == 'pull_request' && 'O0' || 'O2' }}
+      ghc_options: ${{ github.event_name == 'pull_request' && '--ghc-options=-O0' || '' }}
+      release: ${{ startsWith(github.ref, 'refs/tags/v') }}
+      postgres_dist: ${{ startsWith(github.ref, 'refs/tags/v') && 'dist-newstyle' || 'dist-postgres' }}
+      branch: ${{ github.base_ref || github.ref_name }}
+      cache_o0: ${{ github.event_name == 'push' && (github.ref == 'refs/heads/master' || github.ref == 'refs/heads/stable') }}
+      image_cache: ghcr.io/simplex-chat/simplexmq-build-cache:${{ matrix.os }}-${{ matrix.arch }}-ghc${{ matrix.ghc }}
+      source_hashes: dist-newstyle/sources.sha256
     runs-on: ${{ matrix.runner }}
     services:
       postgres:
@@ -142,6 +157,18 @@ jobs:
           sudo apt update
           sudo apt -y install postgresql-client-15
 
+      - name: Log in to GitHub Container Registry
+        if: matrix.should_run == true && env.release != 'true'
+        shell: bash
+        env:
+          GHCR_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        run: echo "$GHCR_TOKEN" | docker login ghcr.io -u "${{ github.actor }}" --password-stdin
+
+      - name: Compute image cache week
+        id: week
+        if: matrix.should_run == true && env.release != 'true'
+        run: echo "week=$(date -u +%G-W%V)" >> "$GITHUB_OUTPUT"
+
       - name: Build and cache Docker image
         if: matrix.should_run == true
         uses: simplex-chat/docker-build-push-action@v6
@@ -153,20 +180,14 @@ jobs:
           build-args: |
             TAG=${{ matrix.os }}
             GHC=${{ matrix.ghc }}
-
-      - name: Cache dependencies
-        if: matrix.should_run == true
-        uses: actions/cache@v4
-        with:
-          path: |
-            ~/.cabal/store
-            dist-newstyle
-          key: ubuntu-${{ matrix.os }}-${{ matrix.arch }}-ghc${{ matrix.ghc }}-${{ hashFiles('cabal.project', 'simplexmq.cabal') }}
+          cache-from: ${{ env.release != 'true' && format('type=registry,ref={0}-{1}', env.image_cache, steps.week.outputs.week) || '' }}
+          cache-to: ${{ github.event_name == 'push' && github.ref == 'refs/heads/master' && format('type=registry,ref={0}-{1},mode=max,ignore-error=true', env.image_cache, steps.week.outputs.week) || '' }}
 
       - name: Start container
         if: matrix.should_run == true
         shell: bash
         run: |
+          mkdir -p ~/.cabal
           docker run -t -d \
             --device /dev/fuse \
             --cap-add SYS_ADMIN \
@@ -177,21 +198,66 @@ jobs:
             -v ${{ github.workspace }}:/project \
             build/${{ matrix.os }}:latest
 
+      - name: Compute cache key prefixes
+        id: keys
+        if: matrix.should_run == true && env.release != 'true'
+        run: |
+          toolchain=$(docker exec builder dpkg-query -W -f '${db:Status-Status} ${binary:Package} ${Version}\n' 'binutils*' 'cpp*' 'gcc*' '*-dev' | grep '^installed ' | sha256sum | cut -c1-16)
+          target=${{ matrix.os }}-${{ matrix.arch }}-ghc${{ matrix.ghc }}-$toolchain
+          echo "deps=deps-$target" >> "$GITHUB_OUTPUT"
+          echo "dist=dist-$target" >> "$GITHUB_OUTPUT"
+
+      - name: Restore dependencies from cache
+        id: restore-deps
+        if: matrix.should_run == true && env.release != 'true'
+        uses: actions/cache/restore@v4
+        with:
+          path: ~/.cabal/store
+          key: ${{ steps.keys.outputs.deps }}-${{ env.branch }}-${{ github.sha }}
+          restore-keys: |
+            ${{ steps.keys.outputs.deps }}-${{ env.branch }}-
+            ${{ steps.keys.outputs.deps }}-
+
+      - name: Warn if dependencies not in cache
+        if: matrix.should_run == true && env.release != 'true' && steps.restore-deps.outputs.cache-matched-key == ''
+        run: echo "::warning title=Cache miss::${{ steps.keys.outputs.deps }}-${{ env.branch }}"
+
+      - name: Restore build directories from cache
+        id: restore-build
+        if: matrix.should_run == true && env.release != 'true'
+        uses: actions/cache/restore@v4
+        with:
+          path: |
+            dist-newstyle
+            dist-postgres
+          key: ${{ steps.keys.outputs.dist }}-${{ env.opt }}-${{ env.branch }}-${{ github.sha }}
+          restore-keys: |
+            ${{ steps.keys.outputs.dist }}-${{ env.opt }}-${{ env.branch }}-
+            ${{ steps.keys.outputs.dist }}-${{ env.opt }}-
+
+      - name: Warn if build directories not in cache
+        if: matrix.should_run == true && env.release != 'true' && steps.restore-build.outputs.cache-matched-key == ''
+        run: echo "::warning title=Cache miss::${{ steps.keys.outputs.dist }}-${{ env.opt }}-${{ env.branch }}"
+
+      - name: Set old timestamps on unchanged sources (GHC 8.10.7)
+        if: matrix.ghc == '8.10.7' && steps.restore-build.outputs.cache-matched-key != '' && hashFiles(env.source_hashes) != ''
+        run: scripts/ci/source-timestamps.sh restore ${{ env.source_hashes }}
+
       - name: Build smp-server, xftp-server (postgresql) and tests
         if: matrix.should_run == true
         shell: docker exec -t builder sh -eu {0}
         run: |
-          chmod -fR 777 ~/.cabal ./dist-newstyle || :; git config --global --add safe.directory '*'
-          cabal clean
+          chmod -fR 777 ~/.cabal ./dist-newstyle ./dist-postgres || :; git config --global --add safe.directory '*'
+          scripts/ci/prune-old-builds.sh dist-newstyle dist-postgres
           cabal update
-          cabal build --jobs=$(nproc) --enable-tests -fserver_postgres
+          cabal build --builddir=${{ env.postgres_dist }} --jobs=$(nproc) --enable-tests -fserver_postgres ${{ env.ghc_options }}
           mkdir -p /out
-          for i in smp-server xftp-server simplexmq-test; do
-            bin=$(find /project/dist-newstyle -name "$i" -type f -executable)
+          for i in smp-server xftp-server ntf-server simplexmq-test; do
+            bin=$(find /project/${{ env.postgres_dist }} -name "$i" -type f -executable)
             chmod +x "$bin"
-            mv "$bin" /out/
+            cp "$bin" /out/
           done
-          strip /out/smp-server /out/xftp-server
+          strip /out/smp-server /out/xftp-server /out/ntf-server
 
       - name: Copy simplexmq-test from container
         if: matrix.should_run == true
@@ -224,16 +290,17 @@ jobs:
           cat hashes.output >> "$GITHUB_OUTPUT"
 
       - name: Build everything else (standard)
+        id: build-standard
         if: matrix.should_run == true
         shell: docker exec -t builder sh -eu {0}
         run: |
-          cabal build --jobs=$(nproc)
+          cabal build --jobs=$(nproc) ${{ env.ghc_options }}
           mkdir -p /out
-          for i in ${{ env.apps }}; do
+          for i in smp-server xftp-server xftp; do
             bin=$(find /project/dist-newstyle -name "$i" -type f -executable)
             strip "$bin"
             chmod +x "$bin"
-            mv "$bin" /out/
+            cp "$bin" /out/
           done
 
       - name: Copy binaries from container and prepare them
@@ -278,6 +345,33 @@ jobs:
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
 
+      - name: Record source hashes (GHC 8.10.7)
+        if: matrix.should_run == true && matrix.ghc == '8.10.7'
+        shell: docker exec -t builder sh -eu {0}
+        run: scripts/ci/source-timestamps.sh record ${{ env.source_hashes }}
+
+      - name: Compute dependency cache key
+        id: hash-deps
+        if: matrix.should_run == true && env.release != 'true'
+        shell: bash
+        run: echo "hash=$(ls ~/.cabal/store/ghc-*/ | sha256sum | cut -c1-16)" >> "$GITHUB_OUTPUT"
+
+      - name: Save dependencies to cache
+        if: matrix.should_run == true && env.release != 'true' && steps.restore-deps.outputs.cache-matched-key != format('{0}-{1}-{2}', steps.keys.outputs.deps, env.branch, steps.hash-deps.outputs.hash)
+        uses: actions/cache/save@v4
+        with:
+          path: ~/.cabal/store
+          key: ${{ steps.keys.outputs.deps }}-${{ env.branch }}-${{ steps.hash-deps.outputs.hash }}
+
+      - name: Save build directories to cache
+        if: matrix.should_run == true && env.release != 'true' && steps.restore-build.outputs.cache-hit != 'true'
+        uses: actions/cache/save@v4
+        with:
+          path: |
+            dist-newstyle
+            dist-postgres
+          key: ${{ steps.keys.outputs.dist }}-${{ env.opt }}-${{ env.branch }}-${{ github.sha }}
+
       - name: Test
         if: matrix.should_run == true && matrix.arch == 'x86-64'
         timeout-minutes: 120
@@ -301,3 +395,75 @@ jobs:
             echo "All "$attempts" attempts failed."
             exit 1
           fi
+
+      - name: Remove -O2 build directories
+        id: remove-o2
+        if: env.cache_o0 == 'true' && !cancelled() && steps.build-standard.outcome == 'success'
+        shell: docker exec -t builder sh -eu {0}
+        run: rm -rf dist-newstyle dist-postgres
+
+      - name: Restore -O0 build directories from cache
+        id: restore-build-o0
+        if: ${{ !cancelled() && steps.remove-o2.outcome == 'success' }}
+        uses: actions/cache/restore@v4
+        with:
+          path: |
+            dist-newstyle
+            dist-postgres
+          key: ${{ steps.keys.outputs.dist }}-O0-${{ env.branch }}-${{ github.sha }}
+          restore-keys: |
+            ${{ steps.keys.outputs.dist }}-O0-${{ env.branch }}-
+            ${{ steps.keys.outputs.dist }}-O0-
+
+      - name: Set old timestamps on unchanged sources for -O0 build (GHC 8.10.7)
+        if: ${{ !cancelled() && matrix.ghc == '8.10.7' && steps.restore-build-o0.outputs.cache-matched-key != '' && hashFiles(env.source_hashes) != '' }}
+        run: scripts/ci/source-timestamps.sh restore ${{ env.source_hashes }}
+
+      - name: Build at -O0 for pull request caches
+        id: build-o0
+        if: ${{ !cancelled() && steps.remove-o2.outcome == 'success' }}
+        shell: docker exec -t builder sh -eu {0}
+        run: |
+          scripts/ci/prune-old-builds.sh dist-newstyle dist-postgres
+          cabal build --builddir=dist-postgres --jobs=$(nproc) --enable-tests -fserver_postgres --ghc-options=-O0
+          cabal build --jobs=$(nproc) --ghc-options=-O0
+
+      - name: Record source hashes for -O0 build (GHC 8.10.7)
+        if: ${{ !cancelled() && matrix.ghc == '8.10.7' && steps.build-o0.outcome == 'success' }}
+        shell: docker exec -t builder sh -eu {0}
+        run: scripts/ci/source-timestamps.sh record ${{ env.source_hashes }}
+
+      - name: Save -O0 build directories to cache
+        if: ${{ !cancelled() && steps.build-o0.outcome == 'success' && steps.restore-build-o0.outputs.cache-hit != 'true' }}
+        uses: actions/cache/save@v4
+        with:
+          path: |
+            dist-newstyle
+            dist-postgres
+          key: ${{ steps.keys.outputs.dist }}-O0-${{ env.branch }}-${{ github.sha }}
+
+# =============================
+#      Resolver test job
+# =============================
+
+# The SNRC resolver is Python, so this job needs none of the Haskell toolchain
+# above and runs independently of it.
+
+  resolver-test:
+    name: "resolver (python)"
+    runs-on: ubuntu-latest
+    steps:
+      - name: Clone project
+        uses: actions/checkout@v3
+
+      - name: Set up uv
+        # uv takes Python from .python-version, the version of the Dockerfile's runtime stage.
+        uses: astral-sh/setup-uv@c18668ad3cf93ea998bef934396af7bb5c839dc7 # v10.2.0
+
+      - name: Lint
+        working-directory: scripts/resolver/service
+        run: uv run --frozen ruff check .
+
+      - name: Test
+        working-directory: scripts/resolver/service
+        run: uv run --frozen pytest -v

.gitignore

@@ -1,4 +1,8 @@
 *.lock
+# the resolver's Docker image and CI install exactly these
+!/scripts/resolver/service/uv.lock
+# local overrides of scripts/resolver/.env.example
+/scripts/resolver/.env
 *.db
 *.db.bak
 *.session.sql
@@ -12,3 +16,4 @@ cabal.project.local~
 *.tix
 .coverage
 
+__pycache__/

cabal.project

@@ -31,7 +31,7 @@ source-repository-package
 source-repository-package
     type: git
     location: https://github.com/simplex-chat/direct-sqlcipher.git
-    tag: f814ee68b16a9447fbb467ccc8f29bdd3546bfd9
+    tag: 2330df3dc8c4b660674d3c02d76bb86f66c2abee
 
 source-repository-package
     type: git

cbits/sntrup761.c

@@ -719,23 +719,32 @@ Small_random (small * out, void *random_ctx, sntrup761_random_func * random)
 
 /* ----- Streamlined NTRU Prime Core */
 
+/* x^p-x-1 has a degree-19 factor mod 3, so a random g is not invertible in R3
+   with probability about 3^-19; KeyGen_attempts failures in a row mean a broken RNG */
+#define KeyGen_attempts 10
+
 /* h,(f,ginv) = KeyGen() */
-static void
+/* returns 0 if KeyGen succeeded; else -1 */
+static int
 KeyGen (Fq * h, small * f, small * ginv, void *random_ctx,
         sntrup761_random_func * random)
 {
   small g[p];
   Fq finv[p];
+  int i;
 
-  for (;;)
+  for (i = 0; i < KeyGen_attempts; ++i)
     {
       Small_random (g, random_ctx, random);
       if (R3_recip (ginv, g) == 0)
         break;
     }
+  if (i == KeyGen_attempts)
+    return -1;
   Short_random (f, random_ctx, random);
   Rq_recip3 (finv, f);          /* always works */
   Rq_mult_small (h, finv, g);
+  return 0;
 }
 
 /* c = Encrypt(r,h) */
@@ -884,18 +893,21 @@ typedef small Inputs[p];        /* passed by reference */
 #define PublicKeys_bytes Rq_bytes
 
 /* pk,sk = ZKeyGen() */
-static void
+/* returns 0 if KeyGen succeeded; else -1 */
+static int
 ZKeyGen (unsigned char *pk, unsigned char *sk, void *random_ctx,
          sntrup761_random_func * random)
 {
   Fq h[p];
   small f[p], v[p];
 
-  KeyGen (h, f, v, random_ctx, random);
+  if (KeyGen (h, f, v, random_ctx, random) != 0)
+    return -1;
   Rq_encode (pk, h);
   Small_encode (sk, f);
   sk += Small_bytes;
   Small_encode (sk, v);
+  return 0;
 }
 
 /* C = ZEncrypt(r,pk) */
@@ -960,19 +972,21 @@ HashSession (unsigned char *k, int b, const unsigned char *y,
 /* ----- Streamlined NTRU Prime */
 
 /* pk,sk = KEM_KeyGen() */
-void
+int
 sntrup761_keypair (unsigned char *pk, unsigned char *sk, void *random_ctx,
                    sntrup761_random_func * random)
 {
   int i;
 
-  ZKeyGen (pk, sk, random_ctx, random);
+  if (ZKeyGen (pk, sk, random_ctx, random) != 0)
+    return -1;
   sk += SecretKeys_bytes;
   for (i = 0; i < PublicKeys_bytes; ++i)
     *sk++ = pk[i];
   random (random_ctx, Inputs_bytes, sk);
   sk += Inputs_bytes;
   Hash_prefix (sk, 4, pk, PublicKeys_bytes);
+  return 0;
 }
 
 /* c,r_enc = Hide(r,pk,cache); cache is Hash4(pk) */

cbits/sntrup761.h

@@ -19,7 +19,8 @@
 
 typedef void sntrup761_random_func (void *ctx, size_t length, uint8_t *dst);
 
-void
+/* returns 0 on success, -1 if the RNG never produced an invertible polynomial */
+int
 sntrup761_keypair (uint8_t *pk, uint8_t *sk,
                    void *random_ctx, sntrup761_random_func *random);
 

contributing/CODE.md

@@ -7,6 +7,7 @@ This file provides guidance on coding style and approaches and on building the c
 When designing code and planning implementations:
 - Apply adversarial thinking, and consider what may happen if one of the communicating parties is malicious.
 - Formulate an explicit threat model for each change - who can do which undesirable things and under which circumstances.
+- Use the default cryptographic primitive for each purpose from the [primitive registry](../protocol/crypto-registry.md); any exception requires review, and the registry must be updated with every new primitive use or domain-separation string.
 
 ## Code Quality Standards
 

docs/smp-server-db-load.md

@@ -0,0 +1,200 @@
+# SMP server database load: root causes and fixes
+
+## Summary
+
+The recurring `pMsgFwdsOwn_pErrorsOther` spikes are a symptom of chronic disk saturation of the
+`smp_server` PostgreSQL database, not a proxy or forwarding bug. The baseline read load is the
+Prometheus scrape running whole-table `COUNT` scans of `msg_queues` (48.5M rows) every 30 seconds,
+compounded by the message-expiration sweep. The write load comes from indexing `updated_at`:
+`updateQueueTime` stamps it on the first send to each active queue per UTC day, and these non-HOT
+updates bloat the indexes to 104 GB and amplify WAL to about 425 GB per day (37% full-page images,
+~31% from `updateQueueTime`). Because the stamp is day-granular, this write burst clusters at 00:00 UTC,
+tipping the saturated disk into the observed spike and loading the pgBackRest backup repository.
+
+Code fixes replace the largest metric scan with an estimate and index the sweep and notifier counts.
+Reclaiming the index bloat, which also clears the remaining metric scan, and keeping it down are
+operator actions, because PostgreSQL never shrinks indexes automatically.
+
+## Environment
+
+The affected deployment runs the pure PostgreSQL backend (`store_queues: database`,
+`store_messages: database`), schema `smp_server`, `db_pool_size 20`, `prometheus_interval 30`,
+`[NAMES]` enabled, own-server proxying enabled, and forwards routed over a SOCKS proxy.
+
+`msg_queues`: 48.5M live rows, 12.3% dead tuples, 14 GB heap, 104 GB indexes, 118 GB total.
+
+## Evidence
+
+`pg_stat_statements` over a 5 minute window, ordered by disk reads (`track_io_timing` was off, so
+`shared_blks_read` is the disk proxy; the two `EXPLAIN ANALYZE` rows are manual and excluded):
+
+| query | calls | mean ms | reads | share of all reads |
+| --- | --- | --- | --- | --- |
+| `getEntityCounts` (the six-`COUNT` metric query) | 30 | 111,456 | 990 GB | 70.7% |
+| queue-record load (`SELECT recipient_id, …`) | 47,027 | 121 | 89 GB | 6.3% |
+| expiration sweep batch (`array_agg` in `expire_old_messages`) | 13 | 233,143 | 62 GB | 4.5% |
+| message peek (`DISTINCT ON (recipient_id)`) | 46,932 | 76 | 43 GB | 3.0% |
+| `updateQueueTime` (`UPDATE msg_queues SET updated_at`) | 172,885 | 4 | 7.7 GB | 0.5% |
+| `write_message` | 110,911 | 5 | 7.0 GB | 0.5% |
+
+`EXPLAIN (ANALYZE, BUFFERS)` of each `msg_queues` count inside `getEntityCounts`:
+
+| count | plan | time |
+| --- | --- | --- |
+| `queue_count` (`deleted_at IS NULL`) | Parallel Seq Scan | 30 s |
+| `notifier_count` (`deleted_at IS NULL AND notifier_id IS NOT NULL`) | Parallel Seq Scan | 32.6 s |
+| `ntf_service_queues_count` (`ntf_service_id IS NOT NULL AND deleted_at IS NULL`) | Parallel Seq Scan | 30.2 s |
+| `rcv_service_queues_count` (`rcv_service_id IS NOT NULL AND deleted_at IS NULL`) | Index Only Scan | 1.4 ms |
+
+The expiration sweep batch: 229 s for one 10,000-row batch, walking `msg_queues_pkey` and discarding
+809,729 rows by filter to find 10,000 expirable ones.
+
+`pg_stat_wal` over the measurement window: 27 TB of WAL accumulated, 37.2% of records full-page
+images, averaging about 425 GB per day; recent days reach ~470 GB (from the `archive-push` log) as
+the queue count grows.
+
+Per-statement WAL from `pg_stat_statements` (`wal_bytes`), leaf statements only (this instance runs
+`pg_stat_statements.track = all`, so the `write_message` wrapper double-counts the `INSERT` and
+`UPDATE` it runs, and is excluded):
+
+| category | share of WAL | main statements |
+| --- | --- | --- |
+| `msg_queues` updates | ~42% | `updateQueueTime` ~31%, `msg_can_write` flags ~9%, insert/delete ~2% |
+| `messages` insert and delete | ~34% | message `INSERT`, `DELETE` by `message_id` and `recipient_id` |
+| read path (hint bits, FPIs on SELECT) | ~24% | message peek ~15%, `msg_queue_size`, `getEntityCounts` |
+
+## Root causes
+
+1. Prometheus scrape. `getEntityCounts` (`QueueStore/Postgres.hs:154`) runs six `COUNT(1)` subqueries
+   every `prometheus_interval` (30 s), called from the metrics path (`Server.hs:698`). Four count
+   `msg_queues`; three of those seq-scan the 48.5M-row heap. At 70.7% of all reads it is the dominant
+   disk consumer, and it runs continuously (mean 111 s per scrape exceeds the 30 s interval).
+
+2. Message-expiration sweep. `expireMessagesThread` (`Server.hs:477`) calls the `expire_old_messages`
+   procedure (`server_schema.sql`), whose inner batch query selects expirable queues ordered by
+   `recipient_id`. With `oldQueue = 0` the `updated_at` filter matches everything, so the planner
+   walks the primary key and filters `msg_queue_expire`, discarding ~99% of rows per batch.
+
+3. Index bloat. 104 GB of indexes on a 14 GB table. `updated_at` is part of
+   `idx_msg_queues_updated_at_recipient_id` (`server_schema.sql:526`) and `updateQueueTime`
+   (`QueueStore/Postgres.hs:443`) updates it on the day's first send per queue (172,885 times in the
+   window above). Updating an indexed column prevents HOT updates, so every update writes a new row
+   version plus new entries in every index on the table, and the old index entries accumulate as bloat.
+
+4. WAL amplification. Only 15.6% of `msg_queues` updates are HOT, because its composite index covers
+   `updated_at` and `msg_queue_expire`, both changed on the hot path (cause 3); the rest rewrite every
+   index on the table. With the 104 GB of bloated indexes and frequent checkpoints (`max_wal_size 4GB`,
+   `checkpoint_timeout 5min`, `wal_compression off`), 37% of WAL records are full-page images, and the
+   server generates about 425 GB of WAL per day (27 TB accumulated; recent days near 470 GB).
+   `updateQueueTime`, stamping `updated_at` on the ~5.3M daily-active queues, is the single largest
+   statement at ~31% of WAL (see Evidence). pgBackRest `archive-push` ships all of it, loading the
+   primary disk and the backup repository. The load peaks
+   at 00:00 UTC because `getSystemDate` rounds `updated_at` to the UTC day, so the first send to each
+   active queue after midnight runs `updateQueueTime`, clustering these writes and the `archive-push`
+   volume in the 00h hour.
+
+## Index bloat and autovacuum
+
+Autovacuum is running (60 runs on this table) and is not misconfigured, but two limits apply.
+
+Autovacuum reclaims heap dead tuples and refreshes planner statistics; it does not shrink indexes.
+Only `REINDEX` or `pg_repack` rebuilds a bloated B-tree. Index bloat therefore has no automatic
+remedy.
+
+Default autovacuum pacing assumes moderate churn. `autovacuum_vacuum_scale_factor = 0.2` waits until
+20% of the table is dead (about 9.7M rows here) before vacuuming, and cost-based throttling caps its
+I/O, so a hot table stays around 12% dead. Per-table tuning makes it run sooner and faster.
+
+The amplifier is indexing `updated_at`, a column updated on every send, which the fixes and operator
+actions below address.
+
+## Implemented fixes (batch 1)
+
+Branch `sh/leaks-batch-1`, two commits.
+
+| commit | change | effect |
+| --- | --- | --- |
+| `88673eee` | migration `20260916_prometheus_indexes` adds partial indexes `idx_msg_queues_expire (recipient_id) WHERE deleted_at IS NULL AND msg_queue_expire` and `idx_msg_queues_notifier_active (notifier_id) WHERE deleted_at IS NULL AND notifier_id IS NOT NULL` | the sweep batch and `notifier_count` become index scans |
+| `2912be1e` | `getEntityCounts.queue_count` uses `pg_class.reltuples` (`QueueStore/Postgres.hs:154`) | removes the 30 s, 990 GB whole-table `COUNT` on every scrape |
+
+Verified: both indexes are created and used as index-only scans, the schema-dump test passes, and
+`smp-server` builds. `queueCount` is consumed only by metrics, logs, and display
+(`Server.hs:528,698,815,2369`), so an estimate is safe.
+
+Not fixed by new indexes:
+
+- `ntf_service_queues_count` seq-scans only because of index bloat; `idx_msg_queues_ntf_service_id`
+  already covers it. `REINDEX` restores the index-only scan, as proven by `rcv_service_queues_count`.
+- `rcv_service_queues_count` and the two `services` counts are already cheap.
+
+## Implemented fixes (batch 2)
+
+Migration `20260917_msg_queues_hot` and `MsgStore/Postgres.hs`.
+
+| change | effect |
+| --- | --- |
+| drop `idx_msg_queues_updated_at_recipient_id`, set `msg_queues` `fillfactor = 80` | `updateQueueTime` updates become HOT, so they rewrite no index entries; removes its index bloat and cuts its WAL |
+| remove the dead `updated_at > p_old_queue` filter and its parameter from `expire_old_messages`, matching the `CALL` (`MsgStore/Postgres.hs:113`) | the sweep query runs as an index-only scan on `idx_msg_queues_expire`; `updated_at` is no longer read on the hot path |
+| autovacuum reloptions on `msg_queues`, `messages` (and its TOAST), and `services` | keeps dead tuples and bloat low per table (see the table below) |
+
+Verified on PostgreSQL 16 with the daily-active pattern (~11% of queues updated per day): `updateQueueTime`
+goes from 0% HOT (indexed `updated_at`) to 100% HOT at `fillfactor = 80` (96.7% at 90), and per-update
+WAL drops about 4x. The HOT ratio depends on how many rows per page change before autovacuum reclaims
+space, so it is workload-dependent; `fillfactor = 80` reached 100% for this fraction, and updating a much
+larger share of queues at once would need a lower value. `fillfactor` applies to pages rewritten after
+the change, so the ratio ramps up as the heap turns over, or immediately after `pg_repack`.
+
+Dropping the index regresses the deprecated journal message store's expiration (`foldRecentQueueRecs`,
+`Journal.hs:434`) to a sequential scan; this is accepted because that store is being retired. The
+Postgres message store does not read `updated_at`.
+
+Autovacuum reloptions applied by the migration:
+
+| table | reloptions | reason |
+| --- | --- | --- |
+| `msg_queues` | `fillfactor 80`, `autovacuum_vacuum_scale_factor 0.02`, `autovacuum_analyze_scale_factor 0.01`, `autovacuum_vacuum_cost_limit 1000` | HOT headroom; vacuum/analyze at ~2%/1% churn instead of 20%/10%; faster vacuum |
+| `messages` (and TOAST) | `autovacuum_vacuum_scale_factor 0.02`, `autovacuum_analyze_scale_factor 0.01`, `toast.autovacuum_vacuum_scale_factor 0.02` | high insert and delete churn; 152 GB, mostly TOASTed message bodies |
+| `services` | `fillfactor 70`, `autovacuum_vacuum_threshold 1000`, `autovacuum_vacuum_scale_factor 0` | tiny, very hot table was autovacuumed tens of thousands of times; make updates HOT and vacuum it far less |
+
+## Operator actions
+
+1. Reclaim index bloat (online, no lock; needs free disk near the index size; run largest first):
+   ```
+   REINDEX INDEX CONCURRENTLY smp_server.idx_msg_queues_updated_at_recipient_id;
+   REINDEX INDEX CONCURRENTLY smp_server.idx_msg_queues_notifier_id;
+   REINDEX INDEX CONCURRENTLY smp_server.idx_msg_queues_ntf_service_id;
+   REINDEX INDEX CONCURRENTLY smp_server.idx_msg_queues_rcv_service_id;
+   REINDEX INDEX CONCURRENTLY smp_server.idx_msg_queues_sender_id;
+   REINDEX INDEX CONCURRENTLY smp_server.idx_msg_queues_link_id;
+   ANALYZE smp_server.msg_queues;
+   ```
+   Expect `pg_indexes_size('smp_server.msg_queues')` to drop from 104 GB to roughly 12 GB (measured
+   after reindexing). `ANALYZE` restores the index-only scan for `ntf_service_queues_count`. With the
+   batch-2 `fillfactor` the index bloat no longer rebuilds, so this is a one-time reclaim.
+
+2. Autovacuum reloptions are applied by the batch-2 migration (`msg_queues`, `messages` and its TOAST,
+   `services`), so no manual `ALTER TABLE` is needed. Reclaim the current `messages` bloat once with
+   `pg_repack smp_server.messages` (152 GB heap and TOAST, roughly 30-50 GB reclaimable); running
+   `pg_repack` on `msg_queues` also makes `fillfactor = 80` effective immediately rather than as pages
+   turn over. With `fillfactor = 80` the `msg_queues` index bloat no longer rebuilds, so scheduled
+   reindexing is rarely needed. To automate it anyway, run `reindexdb --concurrently` or `pg_repack`
+   from a scheduler; `pg_cron` runs each job inside a transaction and cannot execute
+   `REINDEX ... CONCURRENTLY`, so use `pg_cron` only for `VACUUM`/`ANALYZE` or a non-concurrent
+   maintenance-window `REINDEX`.
+
+3. Cut WAL and checkpoint pressure (all reloadable, no restart):
+   ```
+   ALTER SYSTEM SET wal_compression = 'lz4';
+   ALTER SYSTEM SET max_wal_size = '32GB';
+   ALTER SYSTEM SET checkpoint_timeout = '30min';
+   SELECT pg_reload_conf();
+   ```
+   Fewer checkpoints produce fewer full-page images; `wal_compression` shrinks those that remain.
+   Together these reduce the WAL written on the primary and shipped by `archive-push`. The message
+   insert and delete volume itself is fixed by traffic, but most of its WAL cost is amplification
+   (full-page images and hint-bit writes on bloated pages), which these settings and `REINDEX` reduce.
+   This is separate from pgBackRest repo-side compression, which does not reduce WAL on the primary.
+   Verify with `pg_stat_reset_shared('wal')`, wait, then re-check `wal_fpi` and `wal_bytes` in
+   `pg_stat_wal`.
+
+4. Enable `track_io_timing = on` so disk wait time is measurable in `pg_stat_statements`.

plans/2026-09-08-smp-server-dos-report.md

@@ -1,186 +0,0 @@
-# SMP server: memory leaks and CPU/RAM exhaustion paths
-
-Date: 2026-09-08
-Scope: server-side code at HEAD 200129af. Defensive review for the operator. All file:line references verified against the current tree.
-
-## Summary
-
-Scope is the PostgreSQL backend only (`store_queues: database`, `store_messages: database`), which selects `SSCDatabase` -> `PostgresMsgStore`. In that backend the queue store is created with `useCache = False` (`MsgStore/Postgres.hs:100`) and `getMsgQueue` is a no-op (`:172`), so there is no in-memory queue cache and no journal message files.
-
-Consequence: the queue-cache and journal issues below do not occur on this backend. Specifically, memory-leaks 1 and 2 (queue cache) require `useCache = True`, which only the journal-with-Postgres-queues mode (`store_messages: journal`) uses; and the SUBS journal-FD spike requires journal files. On PostgreSQL, SUBS is a single streaming DB query (`MsgStore/Postgres.hs:122`), with no FD/RAM spike.
-
-Real CPU/RAM issues on the PostgreSQL backend (all store-independent): unauthenticated RSLV outbound fan-out, the proxy session never torn down on forward timeout, the absence of any global connection/thread cap, and the service `subscriptions` map not pruned on ACK. Reproduction tests: `testRslvFanOut` and `testProxyForwardTimeoutStuckSession`. No fixes are applied. Verified against a local PostgreSQL (`server_postgres` build).
-
-## Memory leaks
-
-### 1. Queue cache retains deleted queues (not the PostgreSQL backend)
-
-Only with `useCache = True` (journal-with-Postgres-queues mode). The pure PostgreSQL backend uses `useCache = False`, so the `queues` cache is never populated and nothing leaks. Documented for completeness.
-
-`deleteStoreQueue` nulls the queue's `queueRec` TVar and cleans the `senders`, service, and notifier maps, but never removes the entry from the `queues` map.
-
-- STM: `QueueStore/STM.hs` (no `TM.delete rId queues`).
-- Postgres: `QueueStore/Postgres.hs:451-464` (sets `deleted_at`, nulls `queueRec`, never removed from the cache; `compactQueues` at `:148-152` deletes DB rows only).
-- Add sites: STM `QueueStore/STM.hs:126`; Postgres `:182` and `:245` (any load caches).
-
-Effect: each create+delete cycle leaves a permanent `STMQueue`/`JournalQueue` object with its TVars. A client looping NEW then DEL grows RAM without bound; freed only by restart. `getEntityCounts` counts `M.size queues` (`QueueStore/STM.hs:101`), so the reported `queueCount` also drifts up. Reproduced on STM, Journal, and Postgres+journal: 200 create+delete cycles retain 200 cache entries.
-
-Fix direction: drop the entry from the cache on delete in both stores. Safe because `loadQueue` filters `AND deleted_at IS NULL` (`QueueStore/Postgres.hs:231,286,309`), so a deleted queue is never reloaded.
-
-```haskell
-delete q = do
-  writeTVar qr Nothing
-  TM.delete rId $ queues st   -- add: drop deleted queue from the cache
-  TM.delete senderId $ senders st
-  ...
-```
-
-### 2. Postgres queue cache has no eviction (not the pure PostgreSQL backend)
-
-Only with `useCache = True` (journal-with-Postgres-queues mode). The pure PostgreSQL backend (`store_messages: database`) uses `useCache = False`, so there is no cache to grow. Documented for completeness.
-
-Every distinct queue touched by SUB/SEND/NSUB/GET is cached in `queues` via `cacheQueue` (`QueueStore/Postgres.hs:245`) and removed only on delete. There is no idle or size-based eviction, so the working set trends toward every queue accessed since boot, defeating the purpose of the database backing store. `idleQueueInterval` (`Env/STM.hs:233`, 4h) closes journal message-queue handles, not this cache. Requires the Postgres backend.
-
-### 3. Service subscriptions map not pruned on ACK
-
-A receiving-service client's `subscriptions` map gains a `Sub` per delivered queue at `Server.hs:1896` (initial scan) and `Server.hs:2073` (`newServiceDeliverySub` on each SEND to an associated queue). `acknowledgeMsg` (`Server.hs:1929-1951`) clears delivery state but never deletes the map entry. Entries are removed only on queue delete, unassociation, or disconnect. A long-lived service session grows toward the number of associated queues regardless of ACKs.
-
-### 4. Notification store grows without a subscriber (not fixed)
-
-`storeNtf` prepends to a per-notifier list with no bound (`NtfStore.hs:37-42`). The list is drained only by an active notifier subscription or the notification TTL (default 24h, `Env/STM.hs:238-243`). With no notifier subscribed, one entry accumulates per notification-flagged SEND. `deleteExpiredNtfs` also leaves the emptied notifier key in the outer map (`NtfStore.hs:57-61`).
-
-Not fixed here: the per-notifier list is TTL-bounded, and the empty-key retention is bounded by the number of active notifiers (freed on queue/notifier delete), so it is not truly unbounded. Removing the emptied key in `deleteExpiredNtfs` races with `storeNtf`, which reads the outer map outside a transaction; a safe fix requires making `storeNtf` fully transactional, which changes a hot path. Deferred as not worth that trade.
-
-### 5. Proxy client maps grow with distinct PRXY targets
-
-`smpClients` (keyed by `SMPServer`) and `smpSessions` grow per distinct proxied relay. On connect failure with `persistErrorInterval /= 0` (default 30s), the error var is kept in `smpClients` (`Client/Agent.hs:245-247`) and removed only lazily on re-query. Client-controlled `PRXY srv` targets make the target set attacker-chosen.
-
-### 6. Blocked writer threads on a stuck relay
-
-When a relay stops draining, the proxy client's `sndQ` fills and each further command forks a thread parked in `writeTBQueue` (`Client.hs:1367-1370`), each retaining a `Request`. This is the memory amplifier of the proxy stuck-session issue.
-
-### 7. sentCommands entries leak for unanswered commands
-
-`getResponse` sets `pending = False` on timeout but does not delete the `sentCommands` entry (`Client.hs:1372-1383`); removal happens only on a matching response (`Client.hs:704`). Commands a relay never answers stay for the connection lifetime.
-
-## CPU/RAM exhaustion paths
-
-Ranked by ease of trigger times impact.
-
-1. No global connection cap plus a 120s handshake hold. Sockets are accepted and handlers forked with no admission control (`Transport/Server.hs:176-184`); the SMP handshake may hang for `smpHandshakeTimeout` (`Main.hs:535`, 120s). Unauthenticated thread, FD, and buffer exhaustion.
-2. `PRXY` establishes an outbound TLS connection to an arbitrary client-named host and port (`Server.hs:1417-1429`), which is server-side request forgery plus unbounded cached outbound sessions. Gated by `allowSMPProxy` and optional `newQueueBasicAuth`; open on public relays.
-3. `NEW` creates a persistent queue with keys, a store-log record, and disk state (`Server.hs:1569-1628`), with no maximum queue count. On public or no-password servers this is unauthenticated and compounds leak 1.
-4. `RSLV` forks up to `serverResolverConcurrency` (1000) outbound HTTP requests per connection through a per-client counter with no global cap (`Server.hs:1279,1522`, `Env/STM.hs:256`). Only when `[NAMES]` is enabled. No result cache (`Names.hs:61-74`).
-5. `SUBS` on a recipient service folds over every queue of the service and reruns on every reconnect since `hasSub` is false on a fresh connection. On the PostgreSQL backend this is one streaming DB query joining each service queue with its head message (`MsgStore/Postgres.hs:122`): O(N) DB work, but no journal opens and no FD/RAM spike (the journal-open spike, `Journal.hs:449-467`, is journal-mode only). The store-independent cost that remains is the service `subscriptions` map growing (leak 3).
-6. `RFWD` runs one X25519 DH per request inline on the connection thread before validating the inner ciphertext (`Server.hs:2129`), unauthenticated (`Server.hs:1278`).
-7. Batch multiplier: one 16 KB block carries up to 255 transmissions (`Protocol.hs:2292`), multiplying every per-command cost. Unknown-queue transmissions still run a full signature verify via the dummy-verify path (`Server.hs:1345-1350`).
-
-Interaction: `RSLV` (limit 1000) and `PRXY`/`PFWD` (limit 32) share one `procThreads` counter (`Server.hs:1476-1489`), so an RSLV flood also starves proxying on the same connection.
-
-## Config-gated exposure to confirm
-
-- `newQueueBasicAuth`: gates NEW and PRXY. Absent on public servers by design.
-- `allowSMPProxy`: enables PRXY (hardcoded true in the default config path, `Main.hs:610`).
-- `[NAMES] enable`: gates RSLV.
-
-## Reproduction tests
-
-Two store-independent issues have sound reproductions on the PostgreSQL backend. Both are pending (`xit`) specs asserting the correct behavior, each verified to fail (reproduce) when activated (`it`).
-
-- `tests/RSLVTests.hs`
-  - `testRslvFanOut`: one connection must not fan out to many concurrent resolver requests. Verified: 64 concurrent from one connection.
-- `tests/SMPProxyTests.hs`
-  - `testProxyForwardTimeoutStuckSession`: after repeated forward timeouts the session must be dropped, so a later forward returns `NO_SESSION`. Verified: all 10 forwards return `BROKER TIMEOUT` (session never torn down; the operator-reported `pMsgFwdsOwn_pErrorsOther`).
-
-Running the Postgres tests requires a local PostgreSQL with roles `test_server_user` and `postgres` and database `test_server_db`, reachable via the default Unix socket (set `PGHOST`) and `localhost:5432`, plus a `-fserver_postgres` build.
-
-## Sanity of the reproductions
-
-Verified by tracing the exercised code path.
-
-- `testRslvFanOut`: sound, store-independent, timing-based. Full E2E: real RSLV commands to a real names-enabled server; a 3s resolver delay vs a 0.8s sample separates concurrent from sequential; 64 outbound requests are in flight from one connection. `managerConnCount = 10` does not serialize them (64 confirmed). Couples to a resolver-limit the fix must add and the test set.
-- `testProxyForwardTimeoutStuckSession`: sound, store-independent. Real PFWD path; 10 forwards on one session all return `BROKER TIMEOUT`, none `NO_SESSION`, so the session is never torn down. Asserts the desired behavior, so it fails now and passes once a teardown/circuit-breaker fix lands.
-
-Issues without a sound reproduction on the PostgreSQL backend:
-
-- Queue cache leak / no eviction (memory-leaks 1, 2): do not occur, `useCache = False`. The earlier `testDeleteQueueNoLeak`/`testPostgresCacheBounded` only reproduced in journal/cache modes and were removed.
-- SUBS journal-FD spike: journal-mode only; `getMsgQueue` is a no-op on PostgreSQL. The earlier `testSubsServiceScanBounded` was removed.
-- Global admission control: not testable black-box. `connect()` succeeds regardless of an app-level cap (kernel backlog), and an admission fix accepts-then-closes rather than refusing the connect, so a black-box client cannot distinguish fixed from unfixed. Needs a configurable connection cap the test can set or a server-exposed active-connection metric.
-- Service `subscriptions` map not pruned on ACK (memory-leaks 3): real and store-independent, but not reproduced here; needs a messaging-service (service-cert) client that SUBS then receives and ACKs, asserting the map does not grow.
-
-Note on issue 6: task 6 is "bound proxy writer threads to a stuck relay" (`nonBlockingWriteTBQueue` forking when the relay stops reading). The shipped test does not reproduce that; it reproduces the related, higher-impact and operator-observed "session not torn down on forward timeout". The writer-thread fork bomb needs a relay that accepts TCP but stops reading, and `nonBlockingWriteTBQueue` is not exported, so it is not reproduced here.
-
-## Fix directions (general)
-
-Sketches only, one per high-impact issue. Not applied.
-
-Issue 1, drop deleted queues from the cache (both stores):
-
-```haskell
-delete q = do
-  writeTVar qr Nothing
-  TM.delete rId $ queues st   -- add
-  ...
-```
-
-Issue 2, evict the Postgres queue cache working set (idle or bounded LRU):
-
-```haskell
--- periodic sweep, or on insert past a size bound:
-evictIdleQueues st now = atomically $
-  modifyTVar' (queues st) (M.filter (not . idleSince now))   -- keep only recently used
-```
-
-Issue 3, bound the SUBS scan: close each queue after peek and paginate; do not rerun the full scan on every reconnect:
-
-```haskell
-f' a (q, qr) =
-  (runExceptT (tryPeekMsg ms q) >>= f a (recipientId q) . ...) `finally` closeMsgQueue ms q
-```
-
-Issue 4, global RSLV concurrency cap (semaphore in NamesEnv, acquired around the outbound call):
-
-```haskell
-data NamesEnv = NamesEnv { ..., resolverSem :: TVar Int, resolverLimit :: Int }
-
-withResolverSlot NamesEnv {resolverSem, resolverLimit} =
-  bracket_ (atomically $ readTVar resolverSem >>= \n -> if n >= resolverLimit then retry else writeTVar resolverSem (n + 1))
-           (atomically $ modifyTVar' resolverSem (subtract 1))
-```
-
-Issue 5, global connection admission at accept (refuse beyond a cap):
-
-```haskell
--- in the accept loop: gate on a shared counter before forking the handler
-n <- atomically $ stateTVar activeConns (\n -> (n, n + 1))
-if n >= maxConns then close sock else forkHandler sock `finally` atomically (modifyTVar' activeConns (subtract 1))
-```
-
-Issue 6, bound proxy writers and add a forward-timeout circuit breaker:
-
-```haskell
--- fail fast instead of forking an unbounded blocked writer:
-nonBlockingWriteTBQueue q x = atomically (tryWriteTBQueue q x)
-  >>= (`unless` throwIO PCEResponseTimeout)   -- or count and drop
--- and drop the relay session after N consecutive forward timeouts (independent of received bytes)
-```
-
-Also: prune the service `subscriptions` entry on ACK (leak 3, `acknowledgeMsg`); the ntf empty-key (leak 4) needs a transactional `storeNtf` first.
-
-## Additional findings (verified 2026-09-09)
-
-Found by a subagent sweep of the PostgreSQL backend and store-independent hot paths, then each confirmed by reading the cited code. Not yet fixed or tested (except where noted).
-
-Attacker-triggerable (DoS):
-
-1. Batched crypto + per-transmission DB SELECT amplification (HIGH). `dummyVerifyCmd` runs a real Ed25519/Ed448 verify or attacker-chosen X25519 DH per transmission on an unknown queue (`Server.hs:1345-1350`). SEND is a non-batch party (`batchParty` = Recipient/Notifier only, `Protocol.hs:420-423`), so it uses the per-transmission path `mapM (verifyTransmission ...)` (`Server.hs:1167`): one crypto op plus one `getQueueRec` DB SELECT (no cache) per transmission. A 16 KB block holds ~130-250 transmissions (`Protocol.hs:2292`), so one write by any handshake-completing peer causes ~130-250 asymmetric verifications and ~130-250 individual SELECTs. Testable.
-2. `clientService` handshake grows the `services` table without bound (MED-HIGH). `getClientService` accepts a self-signed chain (`CCSelf`, `Transport.hs:755-768`) and verifies an attacker cert per connection; `getCreateService` inserts a `services` row on any new cert fingerprint (`QueueStore/Postgres.hs:469-482`). A fresh self-signed cert per handshake yields an unbounded persistent row per connection, pre-auth. Testable (count rows across N handshakes).
-3. Unbounded delivery threads to a slow consumer (MED). When a subscribed recipient's `sndQ` is full, `tryDeliverMessage` forks a bare `forkIO` parked on `writeTBQueue` (`Server.hs:2068-2072,2085-2105`), not gated by `procThreads`. One blocked thread per queue for a slow consumer subscribed to many queues.
-
-Operational / scale:
-
-4. DB pool is the throughput ceiling (HIGH for scale). Default `poolSize=10` (`Main/Init.hs:44`), one MVar-gated pool shared by all clients. Each SEND is 3 transactions: `getQueueRec` + `deleteExpiredMsgs` + `writeMsg`, and `expireMessagesOnSend=True` by default (`Main.hs:563`, `Server.hs:2004`) forces the middle one on every SEND to a non-empty queue. A second pool (`dbPriorityPool`, another `poolSize` backends) is opened but never used by the SMP server (no `withTransactionPriority _ True` in server code) -> 2xpoolSize backends, half idle.
-5. Prometheus scrape cost (MED-HIGH, if enabled). Every `prometheus_interval` (default 60s), `getDeliveredMetrics` folds over all clients x all subscriptions in memory (`Server.hs:723-734`) and `getEntityCounts` runs 6 `COUNT(1)` scans over `msg_queues`/`services` (`QueueStore/Postgres.hs:154-170`, `Server.hs:699`).
-6. Subscription serialization bottleneck (MED). All subscription churn funnels through one `subQ` + one `serverThread` mutating the single `queueSubscribers` Map-in-a-TVar (`Server.hs:263-360`, `Env/STM.hs:375-391`); a batched SUB of N queues is N separate STM writes to one contended TVar.
-
-Lower / noted (chain verified, not deep-dived): `expire_old_messages` does a full `COUNT(1) FROM messages` every 2h; legacy non-service `deliverNtfsThread` is O(ntf-clients x pending) every 1.5s; `PeriodStats` week/month `IntSet`s grow with distinct active queues; no server-side prepared statements; full-row queue decode per command.

plans/2026-10-08-fast-rotation-old-queue-drain.md

@@ -0,0 +1,62 @@
+# Fast rotation: draining the old receive queue
+
+Recipient only. The sender is unchanged.
+
+## Decisions
+
+- New `RcvSwitchStatus` value:
+
+  | constructor      | encoding        |
+  |------------------|-----------------|
+  | `RSReceivedQEND` | `received_qend` |
+
+  - Stored in `rcv_queues.switch_status`.
+  - `canAbortRcvSwitch` returns `False` for it.
+- `qEndMsg`, by the queue that delivered `QEND`:
+  - the queue named in `QEND`: the named queue is marked deleted, `ICDeleteRcvQueue` is enqueued.
+  - the new queue: the named queue is set to `RSReceivedQEND` and stays subscribed.
+  - `NSCCreate` and `SWITCH QDRcv SPCompleted`: sent when the named queue had no `RSReceivedQEND`.
+- Deletion of an `RSReceivedQEND` queue:
+  - trigger: `OK` response to an `ACK` on that queue, without a `GET` lock on it.
+  - action: the queue is marked deleted, `ICDeleteRcvQueue` is enqueued.
+- Bound, a `cleanupManager` step, for each connection with an `RSReceivedQEND` queue:
+  - condition: another receive queue of the connection has an active subscription.
+  - `delete_errors + 1 < deleteErrorCount`: `delete_errors` is incremented.
+  - otherwise: the queue is deleted as above; the first `DEL` error of `ICDeleteRcvQueue` removes it locally.
+- `ackSMPMessage` result, `type QueueDrained = Bool`:
+  - `OK`: `True`.
+  - `MSG`: `False`; the message is written to `msgQ`.
+- Apps: `received_qend` is added to `RcvSwitchStatus`.
+- Migration `20261008_rcv_switch_status`, SQLite and Postgres:
+  - up: index `idx_rcv_queues_switch_status` on `rcv_queues(switch_status)`.
+  - down: `received_qend` is set to `NULL`, the index is dropped.
+
+## Mechanics
+
+- `Agent/Protocol.hs`: `RSReceivedQEND`.
+- `Agent/Store.hs`: `canAbortRcvSwitch`.
+- `Client.hs`: `QueueDrained`; `ackSMPMessage` returns it.
+- `Agent/Client.hs`: `sendAck` returns `QueueDrained`.
+- `Agent/Store/AgentStore.hs`: `getEndedRcvQueueConnIds`:
+
+  ```sql
+  SELECT conn_id FROM rcv_queues WHERE switch_status = ? AND deleted = 0
+  ```
+
+- `Agent.hs`:
+  - `deleteRcvQueueAsync`: `setRcvQueueDeleted`, then `ICDeleteRcvQueue`.
+  - `qEndMsg`.
+  - `ackQueueMessage`.
+  - `cleanupManager`: `expireEndedRcvQueues`.
+- Apps:
+  - `apps/ios/SimpleXChat/APITypes.swift`
+  - `apps/multiplatform/common/src/commonMain/kotlin/chat/simplex/common/model/SimpleXAPI.kt`
+
+## Tests
+
+- `QEND` on the new queue while a message waits on the old queue:
+  - the message is delivered;
+  - the old queue is deleted after the `ACK` that returns `OK`.
+- The same with a recipient restart before the `ACK`.
+- Old server stopped: the old queue is deleted by the bound.
+- Downgrade: `received_qend` is cleared, other switch statuses are kept.

plans/2026-10-09-auth-error-delay.md

@@ -0,0 +1,52 @@
+# Delayed AUTH error responses
+
+## Decisions
+
+- Responses that include `ERR AUTH` are sent at the next multiple of the delay after the start time.
+- Delayed response sets:
+
+  | Source | Responses sent together | Start time |
+  |---|---|---|
+  | `receive`, failed verification | verification and parsing errors of one received block | block receipt |
+  | `client`, command processing | responses and messages of one batch | batch read from `rcvQ` |
+  | `RFWD`, forwarded command | the `RRES` of that command | start of forwarded command processing |
+
+- A thread created by `forkClient` writes the responses to the send queue.
+- Delay 0: the responses are written to the send queue directly.
+- `authDelayExceeded`: incremented when processing took longer than the delay.
+  - Prometheus counter `simplex_smp_auth_delay_exceeded`.
+  - Not saved in the stats backup.
+  - Not in the daily stats log.
+- Configuration:
+
+  | Item | Value |
+  |---|---|
+  | INI section | `TRANSPORT` |
+  | INI key | `auth_error_delay_ms` |
+  | default | 50 |
+  | `ServerConfig` field | `authErrorDelay`, microseconds |
+  | test servers | 20 ms |
+
+- AUTH timing test: elapsed time of 5 requests, test servers' delay.
+- Out of scope: NTF and XFTP servers.
+
+## Mechanics
+
+- `Server.hs`:
+  - `isAuthError`: response is `ERR AUTH`.
+  - `sendResponses`: responses with `ERR AUTH` to `sendAfterDelay`, others to the send queue.
+  - `sendAfterDelay`: writes responses to the send queue at the next multiple of the delay.
+  - `receive`: block receipt time from the `rcvActiveAt` reading.
+  - `client`: batch time.
+  - `processForwardedCommand`: start time.
+- `Server/Env/STM.hs`: `authErrorDelay` field, `defaultAuthErrorDelayMs`.
+- `Server/Stats.hs`: `authDelayExceeded` in `ServerStats` and `ServerStatsData`.
+- `Server/Prometheus.hs`: `simplex_smp_auth_delay_exceeded`.
+- `Server/Main.hs`: INI value.
+- `Server/Main/Init.hs`: INI template line.
+- `simplexmq.cabal`: `timeit` removed from the test suite.
+- `tests/SMPClient.hs`: `authErrorDelay = 20000`.
+- `tests/ServerTests.hs`:
+  - delayed AUTH for wrong key, absent queue, absent link and suspended queue; immediate `SOK`.
+  - AUTH timing test measures elapsed time.
+- `tests/SMPProxyTests.hs`: delayed AUTH to forwarded `SEND` and `LGET`.

protocol/agent-protocol.md

@@ -85,7 +85,7 @@ SMP agent protocol has 2 main parts:
 
 ![Duplex connection procedure](./diagrams/duplex-messaging/duplex-creating.svg)
 
-The procedure of establishing a duplex connection is explained on the example of Alice and Bob creating a bi-directional connection consisting of two unidirectional (simplex) queues, using SMP agents (A and B) to facilitate it, and two different SMP routers (which could be the same router). It is shown on the diagram above and has these steps:
+The procedure of establishing a duplex connection is explained on the example of Alice (the initiating party) and Bob (the joining party) creating a bi-directional connection consisting of two unidirectional (simplex) queues, using SMP agents (A and B) to facilitate it, and two different SMP routers (which could be the same router). It is shown on the diagram above and has these steps:
 
 1. Alice requests the new connection from the SMP agent A using agent `createConnection` api function.
 2. Agent A creates an SMP queue on the router (using [SMP protocol](./simplex-messaging.md) `NEW` command) and responds to Alice with the invitation that contains queue information and the encryption keys Bob's agent B should use. The invitation format is described in [Connection link](connection-link-1-time-invitation-and-contact-address).
@@ -208,7 +208,7 @@ This syntax of decrypted SMP client message body is defined by `decryptedAgentMe
 Decrypted SMP message client body can be one of 4 types:
 - `agentConnInfo` - used by the initiating party when confirming reply queue - sent in `agentConfirmation` envelope.
 - `agentConnInfoReply` - used by accepting party, includes reply queue(s) in the initial confirmation - sent in `agentConfirmation` envelope.
-- `agentRatchetInfo` - used to pass additional information when renegotiating double ratchet encryption - sent in `agentRatchetKey` envelope.
+- `agentRatchetInfo` - used to pass additional information when renegotiating double ratchet encryption - sent in `agentRatchetKey` envelope. A key sent in reply to another key includes the hash of that key; agents do not reply to such keys.
 - `agentMessage` - all other agent messages.
 
 `agentMessage` contains these parts:
@@ -233,7 +233,8 @@ connInfo = *OCTET
 agentConnInfoReply = %s"D" smpQueues connInfo
 smpQueues = length 1*newQueueInfo ; NonEmpty list of reply queues
 agentRatchetInfo = %s"R" ratchetInfo
-ratchetInfo = *OCTET
+ratchetInfo = [answeredKeyHash *OCTET] ; bytes after answeredKeyHash are ignored
+answeredKeyHash = %s"0" / (%s"1" shortString) ; "0" in a key that starts renegotiation, otherwise SHA-256 of the two raw public keys of the answered key
 
 agentMessage = %s"M" agentMsgHeader aMessage
 agentMsgHeader = agentMsgId prevMsgHash

protocol/crypto-registry.md

@@ -0,0 +1,189 @@
+Revision 1, 2026-10-01
+
+# SimpleX Network: cryptographic primitive registry
+
+The cryptographic primitives, constructions and domain-separation strings used by this repository, and which of them new code may use. For the threat model see [Security](./security.md).
+
+Module aliases follow the codebase: `C` = [Simplex.Messaging.Crypto](../src/Simplex/Messaging/Crypto.hs), `LC` = [Crypto.Lazy](../src/Simplex/Messaging/Crypto/Lazy.hs), `CR` = [Crypto.Ratchet](../src/Simplex/Messaging/Crypto/Ratchet.hs), `SL` = [Crypto.ShortLink](../src/Simplex/Messaging/Crypto/ShortLink.hs), `KEM` = [Crypto.SNTRUP761](../src/Simplex/Messaging/Crypto/SNTRUP761.hs) and [its bindings](../src/Simplex/Messaging/Crypto/SNTRUP761/Bindings.hs), `BBS` = [Crypto.BBS](../src/Simplex/Messaging/Crypto/BBS.hs). Lengths are in bytes.
+
+## Table of contents
+
+- [Policy](#policy)
+- [Defaults for new code](#defaults-for-new-code)
+- [Primitives](#primitives)
+- [SimpleX box constructions](#simplex-box-constructions)
+- [Domain separation and KDF inputs](#domain-separation-and-kdf-inputs)
+- [Nonces and IVs](#nonces-and-ivs)
+- [Implementation sources](#implementation-sources)
+- [xftp-web (TypeScript)](#xftp-web-typescript)
+- [Tests](#tests)
+
+## Policy
+
+- New code uses the default primitive for its purpose from [Defaults for new code](#defaults-for-new-code).
+- Using any other primitive, or a primitive marked *restricted* outside its listed purpose, requires a written justification in the pull request and review by a maintainer responsible for cryptography.
+- *Legacy-only* primitives stay only for the existing wire formats and stored data listed below. Do not add call sites.
+- Every new HKDF info string or hash prefix must be unique, start with `SimpleX`, and be added to [Domain separation and KDF inputs](#domain-separation-and-kdf-inputs).
+- Changing a primitive, key length, KDF input, info string or nonce construction changes the wire format. It requires a protocol version and an update of this registry and of the affected `protocol/` specification.
+
+Status values used below:
+
+| Status | Meaning |
+|---|---|
+| Default | Approved and preferred for new code for this purpose |
+| Approved | Approved for new code for this purpose when the default does not fit |
+| Restricted | Approved only for the listed purpose (external interoperability or a single protocol) |
+| Legacy-only | Kept for existing wire formats or stored data; no new call sites |
+
+## Defaults for new code
+
+| Purpose | Default | Functions |
+|---|---|---|
+| Signature | Ed25519 | `C.sign'`, `C.verify'` with `C.SEd25519` |
+| Key agreement | X25519 | `C.generateKeyPair`, `C.dh'` |
+| Post-quantum KEM | sntrup761 | `KEM.sntrup761Keypair`, `sntrup761Enc`, `sntrup761Dec` |
+| Hybrid DH + KEM secret | HKDF-SHA512 over `dh \|\| kemSecret` with a new info string, as in `CR.rootKdf` | `C.hkdf` |
+| Authenticated encryption | XSalsa20-Poly1305 [SimpleX secretbox](#simplex-box-constructions) | `C.sbEncrypt`/`C.sbDecrypt`, `LC.sbEncryptTailTag` for streams |
+| Authenticated encryption with a DH secret | XSalsa20-Poly1305 [SimpleX crypto_box](#simplex-box-constructions) | `C.cbEncrypt`/`C.cbDecrypt` |
+| Forward-secret symmetric session | HKDF-SHA512 key chain | `C.sbcInit`, `C.sbcHkdf` |
+| KDF | HKDF-SHA512 | `C.hkdf` |
+| Hash commitment, derived identifier | SHA3-256 | `C.sha3_256` |
+| Certificate fingerprint | SHA-256 X.509 fingerprint | `C.signedFingerprint`, `C.certificateFingerprint` |
+| Randomness | ChaChaDRG seeded from system entropy | `C.newRandom`, `C.randomBytes` and the `random*` helpers |
+
+## Primitives
+
+| Primitive | Status | Purpose and protocol | Implementation | Lengths |
+|---|---|---|---|---|
+| Ed25519 | Default | SMP/NTF/XFTP queue and file keys (agent default `rcvAuthAlg`, `sndAuthAlg` in [Agent/Env/SQLite.hs](../src/Simplex/Messaging/Agent/Env/SQLite.hs)), short-link signatures and owner auth (`SL.encodeSign`, `SL.newOwnerAuth`), XRCP identity and session keys ([RemoteControl/Invitation.hs](../src/Simplex/RemoteControl/Invitation.hs)), agent service request signatures, client/service TLS certificates ([Transport/Credentials.hs](../src/Simplex/Messaging/Transport/Credentials.hs)) | crypton `Crypto.PubKey.Ed25519` via `C.sign'`/`C.verify'` | pub 32, priv 32, sig 64; X.509 SPKI 44 |
+| Ed448 | Approved | Server CA and online certificates (default `signAlgorithm = ED448` in [Server/CLI.hs](../src/Simplex/Messaging/Server/CLI.hs)); accepted for SMP command signatures and TLS | crypton `Crypto.PubKey.Ed448`; server key generation by the `openssl genpkey` CLI | pub 57, priv 57, sig 114; SPKI 69 |
+| X25519 | Default | Per-queue e2e and server-to-recipient `crypto_box` keys, SMP session DH, command authenticator (SMP v7+), proxy (PRXY/PFWD), notification tokens, XFTP chunk transport, XRCP hello and announcements | crypton `Crypto.PubKey.Curve25519` via `C.dh'` | pub 32, priv 32, secret 32; SPKI 44 |
+| X448 | Restricted | E2E double ratchet only: X3DH and DH ratchet (`CR.RatchetX448`, E2E v3) | crypton `Crypto.PubKey.Curve448` | pub 56, priv 56, secret 56 |
+| sntrup761 | Default | PQ KEM in double ratchet (E2E v3, agent v5+) and XRCP v1 hello | Vendored C [cbits/sntrup761.c](../cbits/sntrup761.c) via FFI; randomness from ChaChaDRG through `haskell_rng_func` ([Bindings/RNG.hs](../src/Simplex/Messaging/Crypto/SNTRUP761/Bindings/RNG.hs)) | pk 1158, sk 1763, ct 1039, shared 32 |
+| XSalsa20-Poly1305, SimpleX crypto_box | Default | SMP message bodies (server to recipient), per-queue e2e envelope, confirmations, proxy forwarding, notifications, XRCP hello, XFTP chunk transport | crypton `Crypto.Cipher.XSalsa`, `Crypto.MAC.Poly1305`; `C.cryptoBox`, `LC.cbInit` | key 32 (DH secret), nonce 24, tag 16 |
+| XSalsa20-Poly1305, SimpleX secretbox | Default | SMP transport block encryption (SMP v11+), short-link data (SMP v15+), XFTP file encryption, local file encryption ([Crypto/File.hs](../src/Simplex/Messaging/Crypto/File.hs)), XRCP session | same; `C.sbEncrypt`, `LC.sbEncryptTailTag`, `LC.sbInit` | key 32, nonce 24, tag 16 |
+| crypto_box authenticator | Approved | Deniable sender command authorization with X25519 queue keys (SMP v7+); agent currently creates Ed25519 keys | `C.cbAuthenticate`, `C.cbVerify`: `crypto_box(sha512(msg))` | 80 = 64 + 16 |
+| AES-256-GCM, 16-byte IV | Legacy-only | Double ratchet header and body (E2E v3, [pqdr.md](./pqdr.md)) | crypton `Crypto.Cipher.AES`, `Crypto.Cipher.Types`; `C.encryptAEAD`, `C.decryptAEAD`, `C.initAEAD`; J0 = GHASH(IV) as NIST SP 800-38D defines for non-96-bit IVs | key 32, IV 16, tag 16; padded header 88 (PQ off) or 2310 (PQ on), `CR.paddedHeaderLen` |
+| AES-256-GCM, 12-byte IV | Restricted | WebRTC frame encryption in simplex-chat (`Simplex.Chat.Mobile.WebRTC`), for WebCrypto interoperability | `C.encryptAESNoPad`, `C.decryptAESNoPad`, `C.initAEADGCM`, `C.GCMIV` | key 32, IV 12, tag 16 |
+| HKDF-SHA512 | Default | All KDFs listed in [Domain separation](#domain-separation-and-kdf-inputs) | crypton `Crypto.KDF.HKDF` with `SHA512`; `C.hkdf` (extract + expand) | output per use, at most 255 * 64 |
+| SHA-256 | Default for X.509 fingerprints, otherwise Restricted | X.509 fingerprints (`C.KeyHash`, server identity, XRCP CA, service certificate hash, SMP v16+), XFTP chunk digests, agent message hashes, `requestCode`, ratchet key dedup hash, security code `codeAD = sha256(rcAD)` | crypton; `C.sha256Hash`, `LC.sha256Hash`, `getFingerprint ... HashSHA256` | 32 |
+| SHA-512 | Restricted | XFTP file digests and file description hash, input of the crypto_box authenticator | crypton; `C.sha512Hash`, `LC.sha512Hash` | 64 |
+| SHA-512 (inside sntrup761) | Restricted | sntrup761 internal hash | [cbits/sha512.c](../cbits/sha512.c) calls OpenSSL `SHA512()` from the system `libcrypto` (`extra-libraries: crypto`) | 64 |
+| SHA3-256 | Default | Short-link key `sha3_256(fixedData)` (SMP v15+), XRCP hybrid secret, service request binding (agent v8) | crypton; `C.sha3_256`, `KEM.kemHybridSecret` | 32 |
+| SHA3-384 | Restricted | Client-supplied sender ID: first 24 bytes of `sha3_384(corrId)`, computed by the agent (`prepareConnectionLink'`, `newRcvConnSrv`) and the server (`createQueue`) | crypton; `C.sha3_384` | 48, truncated to 24 |
+| Keccak-256 | Restricted | Label hash for SMP name queries (`NameQuery NQHash`), must match the on-chain registry | crypton `Keccak_256`; `labelHash` in [SimplexName.hs](../src/Simplex/Messaging/SimplexName.hs) | 32 |
+| MD5 | Legacy-only, non-security | XOR-aggregated queue ID hash `IdsHash` for service subscription reconciliation (SUBS, NSUBS, SOKS, ENDS) | crypton (`C.md5Hash`, `Protocol.queueIdHash`); SQLite UDF `simplex_xor_md5_combine` ([Agent/Store/SQLite.hs](../src/Simplex/Messaging/Agent/Store/SQLite.hs)); PostgreSQL pgcrypto `digest(..., 'md5')` in server and NTF schemas | 16 |
+| ChaChaDRG | Default | Keys, nonces, correlation IDs, random IDs, sntrup761 randomness | crypton `Crypto.Random`; `C.newRandom` seeds with `drgNew` from system entropy | n/a |
+| BBS+ over BLS12-381, SHA-256 suite | Restricted | Badge entitlement credentials and proofs ([Crypto/Entitlement.hs](../src/Simplex/Messaging/Crypto/Entitlement.hs)), XFTP storage-time extension | Vendored submodules libbbs and blst (C and assembly) via FFI; randomness from libbbs `getentropy`, `SecRandomCopyBytes` with the `commoncrypto` flag, `BCryptGenRandom` on Windows ([cbits/getentropy_win.c](../cbits/getentropy_win.c)) | sk 32, pk 96, sig 80, proof 272 + 32 per undisclosed message |
+| ECDSA with SHA-256 (ES256) | Restricted | APNS provider JWT ([Push/APNS.hs](../src/Simplex/Messaging/Notifications/Server/Push/APNS.hs)) | crypton `Crypto.PubKey.ECC.ECDSA.sign`; key read by cryptostore `Crypto.Store.PKCS8`; curve taken from the key file | signature DER `SEQUENCE {r, s}`, base64url |
+| TLS 1.3 / 1.2 | Restricted | SMP, XFTP, NTF, XRCP transport: `TLS_CHACHA20_POLY1305_SHA256` (1.3), `ECDHE_ECDSA_CHACHA20POLY1305_SHA256` (1.2), groups X448 and X25519, Ed448 and Ed25519 signatures (`defaultSupportedParams`) | `tls` 1.9 | n/a |
+| TLS for browsers | Restricted | XFTP server with HTTPS credentials (`defaultSupportedParamsHTTPS`: `ciphersuite_strong`, FFDHE, P-521, ECDSA and RSA signatures); SMP server HTTPS requires RSA-4096 (`checkHTTPSCredentials`) | `tls`, `warp-tls` | n/a |
+| X.509 | Restricted | Certificate chains, fingerprints, signed session keys (`C.signX509`, `C.verifyX509`) | `crypton-x509`, `crypton-x509-validation`, `cryptostore` | n/a |
+| SQLCipher | Restricted | Agent SQLite database at rest (`PRAGMA key`); cipher parameters are SQLCipher defaults, not set in this repository | `direct-sqlcipher` (git dependency, [cabal.project](../cabal.project)) | n/a |
+
+## SimpleX box constructions
+
+`C.cryptoBox` and `LC.sbInit_` compute, for a 32-byte key `k` and 24-byte nonce `n`:
+
+```
+k1 = HSalsa20(k, 0^16)
+subkey = HSalsa20(k1, n[0..16])
+stream = Salsa20(subkey, n[16..24])
+polyKey = stream[0..32]
+ct = msg XOR stream[32..]
+tag = Poly1305(polyKey, ct)
+```
+
+| Construction | Key `k` | Equivalent libsodium call | Layout |
+|---|---|---|---|
+| SimpleX crypto_box (`C.cbEncrypt`, `LC.cbInit`) | X25519 shared secret | `crypto_box_easy` (`crypto_box_beforenm` is `HSalsa20(dh, 0^16)`) | strict: `tag \|\| ct`; lazy tail-tag: `ct \|\| tag` |
+| SimpleX secretbox (`C.sbEncrypt`, `LC.sbInit`, `KEM.kcb*`) | 32-byte symmetric key | `crypto_secretbox_easy` with key `crypto_core_hsalsa20(0^16, k)`, not with `k` | same |
+
+Padding before encryption: `C.pad` prefixes a 2-byte big-endian length and fills with `#` (message at most 65533 bytes); `LC.pad` prefixes an 8-byte length. `*NoPad` variants skip padding.
+
+## Domain separation and KDF inputs
+
+HKDF is `C.hkdf salt ikm info len` (HKDF-SHA512).
+
+| Info string | Salt | IKM | Output (split) | Function | Use |
+|---|---|---|---|---|---|
+| `"SimpleXX3DH"` | 64 zero bytes | `dh1 \|\| dh2 \|\| dh3 [\|\| kemShared]` | 96: `hk`, `nhk`, root key (32 each) | `CR.pqX3dh` | Ratchet initialization, added in E2E v2; KEM secret from E2E v3 (current minimum) |
+| `"SimpleXVerifyCode"` | 64 zero bytes | same as `SimpleXX3DH` | 32: `rcVCPQ` | `CR.pqX3dh` | Verification code covering all handshake keys, new ratchets |
+| `"SimpleXRootRatchet"` | root key | `dh [\|\| kemShared]` | 96: root key, chain key, next header key | `CR.rootKdf` | DH/PQ ratchet step |
+| `"SimpleXChainRatchet"` | empty | chain key | 96: chain key, message key (32 each), message IV (16), header IV (16) | `CR.chainKdf` | Per-message keys |
+| `"SimpleXSbChainInit"` | SMP session ID | X25519 session secret | 64: two 32-byte chain keys | `C.sbcInit` from `Transport.blockEncryption` | SMP transport block encryption, SMP v11+ |
+| `"SimpleXSbChainInit"` | empty | XRCP hybrid secret (below) | 64: two 32-byte chain keys | `C.sbcInit` in [RemoteControl/Client.hs](../src/Simplex/RemoteControl/Client.hs) | XRCP v1 session |
+| `"SimpleXSbChain"` | empty | chain key | 88: chain key (32), secretbox key (32), nonce (24) | `C.sbcHkdf` | Each SMP block and XRCP message |
+| `"SimpleXContactLink"` | empty | link key (32) | 56: link ID (24), secretbox key (32) | `SL.contactShortLinkKdf` | Contact short links, SMP v15+ |
+| `"SimpleXInvLink"` | empty | link key (32) | 32: secretbox key | `SL.invShortLinkKdf` | Invitation short links, SMP v15+ |
+
+Other derivations:
+
+| Value | Construction | Function | Use |
+|---|---|---|---|
+| Service request binding | `sha3_256("SimpleXService" \|\| rcAD)` | `serviceReqBinding` in [Agent.hs](../src/Simplex/Messaging/Agent.hs) | Agent RPC, agent v8 |
+| BBS header | `"SimpleX badges v1"` | `entitlementBBSHeader` | Badge credentials |
+| XRCP hybrid secret | `sha3_256(x25519Dh \|\| kemShared)` | `KEM.kemHybridSecret` | XRCP v1 |
+| Short-link key | `sha3_256(smpEncode FixedLinkData)` | `SL.encodeSignFixedData`, checked in `SL.decryptLinkData` | SMP v15+ |
+| Sender ID | `take 24 (sha3_384 corrId)` | `prepareConnectionLink'`, `newRcvConnSrv` in Agent.hs; `createQueue` in [Server.hs](../src/Simplex/Messaging/Server.hs) | Client-supplied sender ID with link data, SMP v15+; the server rejects a mismatch to prevent an ID oracle |
+| Ratchet associated data | `pubKeyBytes sk1 \|\| pubKeyBytes rk1` (raw X448, 112) | `CR.pqX3dh` | AD of every ratchet AEAD |
+| Security code | `sha256(rcAD)` | `ratchetVerifyCodes` in [AgentStore.hs](../src/Simplex/Messaging/Agent/Store/AgentStore.hs) | Connection verification |
+| Contact request code | `sha256(smpEncode (k1, k2, kem, sndId))` | `requestCode` in Agent.hs | Contact request binding |
+| Command authenticator | `crypto_box(sha512(authorized))` | `C.cbAuthenticate` | SMP v7+ |
+| Queue IDs hash | `xor` of `md5(queueId)` | `Protocol.queueIdsHash` | Service subscriptions |
+
+`"SimpleXSbChainInit"` and `"SimpleXSbChain"` are shared by SMP block encryption and XRCP. Their outputs are separated only by the IKM (and the salt for `sbcInit`).
+
+## Nonces and IVs
+
+| Use | Construction |
+|---|---|
+| SMP command authenticator, proxied command | `C.cbNonce corrId`, where `corrId` is a random 24-byte nonce (`C.randomCbNonce`) |
+| SMP proxied responses | `C.reverseNonce` of the request nonce |
+| Server-to-recipient message body | `C.cbNonce msgId`; `msgIdBytes = 24` in [Server/Main.hs](../src/Simplex/Messaging/Server/Main.hs) |
+| Per-queue e2e envelope, short-link data, notifications, XRCP hello, XFTP chunk download | Random 24-byte nonce, sent with the ciphertext |
+| XFTP file, local encrypted file | Random key and nonce per file (`C.randomSbKey`, `C.randomCbNonce`) |
+| SMP block, XRCP session messages | Key and nonce from `C.sbcHkdf`, one per block |
+| Ratchet header, body | 16-byte IVs from `CR.chainKdf`; header IV is sent, body IV is not |
+| WebRTC frames | 12-byte `C.GCMIV` supplied by the caller in simplex-chat |
+
+## Implementation sources
+
+| Source | Version or pin | Provides |
+|---|---|---|
+| crypton | 0.34 | Ed25519, Ed448, X25519, X448, XSalsa20, Poly1305, AES-GCM, HKDF, SHA-2, SHA-3, Keccak, MD5, ChaChaDRG, ECDSA |
+| tls, crypton-x509, crypton-x509-validation, cryptostore | 1.9.0, 1.7.6, 1.6.12, 0.3.0.1 | TLS, X.509, PKCS#8 |
+| [cbits/sntrup761.c](../cbits/sntrup761.c) | Copy of draft-josefsson-ntruprime-streamlined-00 ([cbits/README.md](../cbits/README.md)) | sntrup761 |
+| [cbits/sha512.c](../cbits/sha512.c) and system OpenSSL `libcrypto` | system | SHA-512 for sntrup761 |
+| `cbits/libbbs` submodule ([simplex-chat/libbbs](https://github.com/simplex-chat/libbbs)) | `59a0f4bf` | BBS+ (`bbs_sha256_ciphersuite`), SHA-256, SHAKE256 |
+| `cbits/blst` submodule ([supranational/blst](https://github.com/supranational/blst)) | `db3defd0` | BLS12-381 (C and `build/assembly.S`, `-D__BLST_PORTABLE__`) |
+| direct-sqlcipher | git `f814ee68` | SQLCipher |
+| `openssl` CLI | system | Server CA and certificate key generation (`createServerX509_`) |
+
+## xftp-web (TypeScript)
+
+[xftp-web](../xftp-web/) reimplements a subset for the browser XFTP client. It has no AES-GCM, HKDF, SHA-3, MD5, X448 or sntrup761 code.
+
+| Primitive | Implementation | Haskell counterpart |
+|---|---|---|
+| Ed25519 sign, verify, keys | libsodium-wrappers-sumo (`crypto_sign_*`), [src/crypto/keys.ts](../xftp-web/src/crypto/keys.ts) | `C.sign'`, `C.verify'` |
+| Ed448 verify | `@noble/curves/ed448`, keys.ts `verifyEd448` | `C.verify'` |
+| X25519 | libsodium `crypto_scalarmult`, `crypto_box_keypair` | `C.dh'` |
+| SimpleX crypto_box, secretbox, tail-tag streaming | Salsa20 block function written in TypeScript, libsodium `crypto_core_hsalsa20` and `crypto_onetimeauth_*`, [src/crypto/secretbox.ts](../xftp-web/src/crypto/secretbox.ts) | `C.cryptoBox`, `LC.sbInit_` |
+| crypto_box authenticator | [src/protocol/client.ts](../xftp-web/src/protocol/client.ts) `cbAuthenticate`, `cbVerify` | `C.cbAuthenticate`, `C.cbVerify` |
+| SHA-256, SHA-512 | libsodium `crypto_hash_sha256`, `crypto_hash_sha512*`, [src/crypto/digest.ts](../xftp-web/src/crypto/digest.ts) | `C.sha256Hash`, `C.sha512Hash` |
+| Random | WebCrypto `crypto.getRandomValues`; libsodium for key generation | `C.randomBytes` |
+
+## Tests
+
+None of the tests below compares against published reference vectors (NIST, RFC 8032, RFC 7748, RFC 5869, NaCl, the sntrup761 draft or the BBS draft). Interoperability is tested between this repository's own implementations.
+
+| Area | Test module (hspec group) | Kind |
+|---|---|---|
+| Ed25519, Ed448, X25519 crypto_box, secretbox, lazy and tail-tag secretbox, AES-GCM 12-byte IV, X.509 key encoding, X.509 chains, sntrup761, BBS+, entitlements, padding | [tests/CoreTests/CryptoTests.hs](../tests/CoreTests/CryptoTests.hs) | Round-trip; fixed lengths for BBS+; fixed bytes for padding |
+| Double ratchet (X25519 and X448), PQ KEM agreement, AES-GCM 16-byte IV | [tests/AgentTests/DoubleRatchetTests.hs](../tests/AgentTests/DoubleRatchetTests.hs) | Round-trip; decoding of a stored v2 ratchet JSON |
+| Short links (HKDF info strings, SHA3-256 link key) | [tests/AgentTests/ShortLinkTests.hs](../tests/AgentTests/ShortLinkTests.hs) | Round-trip, tamper rejection |
+| File encryption | [tests/CoreTests/CryptoFileTests.hs](../tests/CoreTests/CryptoFileTests.hs) | Round-trip |
+| XRCP session (SHA3-256 hybrid, sb chain) | [tests/RemoteControl.hs](../tests/RemoteControl.hs) | End-to-end |
+| SMP authenticators, block encryption, `IdsHash` (MD5) | [tests/ServerTests.hs](../tests/ServerTests.hs) | End-to-end; with the PostgreSQL queue store this checks Haskell MD5 against pgcrypto |
+| Haskell and xftp-web: SHA-256/512, Ed25519, Ed448 identity proof, X25519, DER keys, crypto_box, secretbox, tail-tag, authenticator, file encryption | [tests/XFTPWebTests.hs](../tests/XFTPWebTests.hs) (`XFTP Web Client`) | Byte-identical cross-language output |

protocol/diagrams/xrcp/session.mmd

@@ -17,11 +17,11 @@ note over CC, HC: TLS handshake
 par
     note over CC: validate client X509 credentials
     CC->>CI: session code from tlsUnique
-    CI-->>CC: user confirmation
+    CI-->>CC: user confirmation (optional if known)
 and
     note over HC: validate server X509 credentials
     HC->>HI: session code from tlsUnique
-    HI-->>HC: user confirmation
+    HI-->>HC: user confirmation (optional if known)
 end
 
 note over CI, HI: 3. Session verification and protocol negotiation

protocol/pqdr.md

@@ -62,14 +62,16 @@ It is possible to reduce size overhead by using only one KEM agreement and makin
 
 ## Double ratchet with encrypted headers augmented with double PQ KEM
 
-Algorithm below assumes that in addition to shared secret from the initial key agreement, there will be an encapsulation key available from the party that published its keys (Bob).
+Algorithm below assumes that in addition to shared secret from the initial key agreement, there will be an encapsulation key available from the initiating party, that sent its keys in the connection invitation to the joining party (see [agent protocol](./agent-protocol.md)). Following the double ratchet specification, the pseudo-code below names the joining party Alice and the initiating party Bob.
+
+Unlike X3DH prekey bundles, these keys are not reusable keys published for any party to use. The initiating party generates two X448 key pairs and, optionally, a sntrup761 KEM key pair for each connection, sends the public keys in the invitation, and deletes the stored private keys once the ratchet is initialized; only the second X448 private key and the agreed KEM key pair remain in the ratchet state as its initial ratchet keys. The joining party generates its keys when joining and keeps only its KEM key pair, in the ratchet state. The exception is a contact address that publishes ratchet keys in its link data: all requesters use these keys until the address owner rotates them, and the owner keeps the private keys of a few recent generations.
 
 ### Initialization
 
 The double ratchet initialization is defined in pseudo-code. This pseudo-code is identical to Signal algorithm specification except for that parts that add post-quantum key agreement.
 
 ```
-// Alice obtained Bob's keys and initializes ratchet first
+// Alice (joining party) received Bob's keys in the invitation and initializes ratchet first
 def RatchetInitAlicePQ2HE(state, SK, bob_dh_public_key, shared_hka, shared_nhkb, bob_pq_kem_encapsulation_key):
     state.DHRs = GENERATE_DH()
     state.DHRr = bob_dh_public_key
@@ -89,7 +91,7 @@ def RatchetInitAlicePQ2HE(state, SK, bob_dh_public_key, shared_hka, shared_nhkb,
     state.HKr = None
     state.NHKr = shared_nhkb
 
-// Bob initializes ratchet second, having received Alice's connection request
+// Bob (initiating party) initializes ratchet second, having received Alice's first message
 def RatchetInitBobPQ2HE(state, SK, bob_dh_key_pair, shared_hka, shared_nhkb, bob_pq_kem_key_pair):
     state.DHRs = bob_dh_key_pair
     state.DHRr = None
@@ -210,6 +212,8 @@ The outer envelope contains the encrypted header (used as associated data for bo
 
 The message body is encrypted with AES-256-GCM using the message key derived from the sending chain key (`KDF_CK`). The associated data for body encryption is the concatenation of the ratchet associated data and the encoded encrypted header.
 
+`KDF_CK(CK)` is HKDF-SHA512 with empty salt, `CK` as input key material and info `"SimpleXChainRatchet"`, producing 96 bytes split into the next chain key (32 bytes), the message key (32 bytes), the message body IV (16 bytes, not transmitted) and `headerIV` (16 bytes). Both IVs are used as 16-byte AES-256-GCM IVs, not the 12-byte IVs recommended by NIST SP 800-38D, so the initial counter block is J0 = GHASH(IV || 0^64 || [128]_64) as defined there for non-96-bit IVs. WebCrypto and other conforming implementations compute it when given the full 16-byte IV; truncating the IV to 12 bytes produces different ciphertext.
+
 ```abnf
 encRatchetMessage = versionedLength encMessageHeader msgAuthTag encMsgBody
 ; encMessageHeader is used as associated data for body decryption: AD = rcAD || encMessageHeader
@@ -274,7 +278,7 @@ As SimpleX Messaging Protocol pads messages to a fixed size, using 16kb transpor
 
 Sharing the initial keys in case of SimpleX Chat it is equivalent to sharing the invitation link. As encapsulation key is large, it may be inconvenient to share it in the link in some contexts, e.g. when QR codes are used.
 
-It is possible to postpone sharing the encapsulation key until the first message from Alice (confirmation message in SMP protocol), the party sending connection request. The upside here is that the invitation link size would not increase. The downside is that the user profile shared in this confirmation will not be encrypted with PQ-resistant algorithm.
+It is possible to postpone sharing the encapsulation key until the first message from the joining party (confirmation message in SMP protocol). The upside here is that the invitation link size would not increase. The downside is that the user profile shared in this confirmation will not be encrypted with PQ-resistant algorithm.
 
 Another consideration is pairwise ratchets in groups. Key generation in sntrup761 is quite slow - on slow devices it can be as slow as 10-20 keys per second, so using this primitive in groups larger than 10-20 members would result in slow performance.
 

protocol/security.md

@@ -15,7 +15,7 @@ This document describes the cryptographic primitives and threat model for the Si
   - [SimpleX Messaging Protocol router that proxies the messages to another SMP router](#simplex-messaging-protocol-router-that-proxies-the-messages-to-another-smp-router)
   - [An attacker who obtained Alice's (decrypted) chat database](#an-attacker-who-obtained-alices-decrypted-chat-database)
   - [A user's contact](#a-users-contact)
-  - [An attacker who observes Alice showing an introduction message to Bob](#an-attacker-who-observes-alice-showing-an-introduction-message-to-bob)
+  - [An attacker who observes the initiating party showing an introduction message to the joining party](#an-attacker-who-observes-the-initiating-party-showing-an-introduction-message-to-the-joining-party)
   - [An attacker with Internet access](#an-attacker-with-internet-access)
 
 
@@ -37,6 +37,8 @@ This document describes the cryptographic primitives and threat model for the Si
   - AES-GCM AEAD cipher,
   - SHA512-based HKDF for key derivation.
 
+All primitives in use, their lengths, domain-separation strings and the policy for new code are listed in the [cryptographic primitive registry](./crypto-registry.md).
+
 
 ## Threat Model
 
@@ -168,7 +170,7 @@ This document describes the cryptographic primitives and threat model for the Si
 
 *cannot:*
 
-- impersonate a sender and send messages to the user whose database was stolen. Doing so requires also compromising the router (to place the message in the queue, that is possible until the Double-Ratchet advances forward) or the user's device at a subsequent time (to place the message in the database).
+- impersonate a sender and send messages to the user whose database was stolen, unless the database contains a new rotation queue that is not secured yet. Doing so requires also compromising the router (to place the message in the queue, that is possible until the Double-Ratchet advances forward) or the user's device at a subsequent time (to place the message in the database).
 
 - undetectably communicate at the same time as Alice with her contacts. Doing so would result in the contact getting different messages with repeated IDs.
 
@@ -190,15 +192,15 @@ This document describes the cryptographic primitives and threat model for the Si
 
 - cannot collaborate with another of the user's contacts to confirm they are communicating with the same user.
 
-### An attacker who observes Alice showing an introduction message to Bob
+### An attacker who observes the initiating party showing an introduction message to the joining party
 
 *can:*
 
-- Impersonate Bob to Alice.
+- Impersonate the joining party to the initiating party.
 
 *cannot:*
 
-- Impersonate Alice to Bob.
+- Impersonate the initiating party to the joining party.
 
 ### An attacker with Internet access
 

protocol/simplex-messaging.md

@@ -86,7 +86,7 @@ It's designed with the focus on communication security and integrity, under the
 
 It is designed as a low level protocol for other application protocols to solve the problem of secure and private message transmission, making [MITM attack][1] very difficult at any part of the message transmission system.
 
-This document describes SMP protocol version 20. Versions 1-5 are discontinued. The version history:
+This document describes SMP protocol version 23. Versions 1-5 are discontinued. The version history:
 
 - v1: binary protocol encoding
 - v2: message flags (used to control notifications)
@@ -108,6 +108,8 @@ This document describes SMP protocol version 20. Versions 1-5 are discontinued.
 - v19: service subscriptions to messages (SUBS, NSUBS, SOKS, ENDS, ALLS commands)
 - v20: public namespaces resolver (RSLV command, RNAME response) — direct or forwarded via PFWD
 - v21: server public information in handshake
+- v22: `RNAME` says whether a name can be registered, not only what it resolves to
+- v23: version in nonces of forwarded commands, random nonces for forwarded responses
 
 ## Introduction
 
@@ -1119,13 +1121,15 @@ When the client receives PKEY response it MUST validate that:
 - the fingerprint of the received certificate matches fingerprint in the router address - it mitigates MITM attack by proxy.
 - the router session key is correctly signed with the received certificate.
 
+`smpVersionRange` in `PKEY` is not signed.
+
 The proxy router may respond with error response in case the destination router is not available or in case it has an earlier version that does not support proxied commands.
 
 #### Send command via proxy
 
 Sender can send `SKEY` and `SEND` commands via proxy after obtaining the session ID with `PRXY` command (see [Request proxied session](#request-proxied-session)).
 
-Transmission sent to proxy router should use session ID as entity ID and use a random correlation ID of 24 bytes as a nonce for crypto_box encryption of transmission to the destination router. The random ephemeral X25519 key to encrypt transmission should be unique per command, and it should be combined with the key sent by the router in the handshake header to proxy and to the client in `PKEY` command.
+Transmission sent to proxy router should use session ID as entity ID and use a random correlation ID of 24 bytes as a nonce for crypto_box encryption of transmission to the destination router. When `smpVersion` in `PFWD` is 23 or higher, the first 2 bytes of this nonce are XOR-ed with `smpVersion`. The random ephemeral X25519 key to encrypt transmission should be unique per command, and it should be combined with the key sent by the router in the handshake header to proxy and to the client in `PKEY` command.
 
 Encrypted transmission should use the received session ID from the connection between proxy router and destination router in the authorized body.
 
@@ -1137,10 +1141,11 @@ commandKey = length x509encoded
 
 The proxy router will forward the encrypted transmission in `RFWD` command (see below).
 
-Having received the `RRES` response from the destination router, proxy router will forward `PRES` response to the client. `PRES` response should use the same correlation ID as `PFWD` command. The destination router will use this correlation ID increased by 1 as a nonce for encryption of the response.
+Having received the `RRES` response from the destination router, proxy router will forward `PRES` response to the client. `PRES` response should use the same correlation ID as `PFWD` command. The destination router will use this correlation ID increased by 1 as a nonce for encryption of the response. When `smpVersion` in `PFWD` is 23 or higher, the destination router uses a random nonce instead, and sends it before the encrypted response.
 
 ```abnf
-proxyResponse = %s"PRES" SP <encrypted padded(forwardedResponse, 16226)>
+proxyResponse = %s"PRES" SP [responseNonce] <encrypted padded(forwardedResponse, 16226)>
+responseNonce = %s"0" / (%s"1" 24*24 OCTET) ; from v23
 forwardedResponse = *OCTET ; client-encrypted SMP response, decrypted by client using per-command DH secret
 ```
 
@@ -1167,7 +1172,7 @@ The shared secret for encrypting transmission bodies between proxy router and de
 
 
 ```abnf
-relayResponse = %s"RRES" SP <encrypted(responseTransmission)>
+relayResponse = %s"RRES" SP [responseNonce] <encrypted(responseTransmission)>
 responseTransmission = fwdCorrId forwardedResponse
   ; fwdCorrId and forwardedResponse defined above in RFWD section
 ```
@@ -1451,53 +1456,150 @@ reads `NameRecord` from. The reference implementation forwards each RSLV to a
 companion REST resolver process (`scripts/resolver/snrc-resolve.py`) that
 queries the SNRC contract on Ethereum; alternative backings (different chains,
 DHT, etc.) are valid as long as they expose the documented HTTP shape (`GET
-/resolve/<name>` returning a `NameRecord` on 200, 404 / 400 for unknown names
-or TLDs, 502 for upstream RPC failures) or substitute a different transport
-while still returning a `NameRecord` matching the encoding below.
+/v2/resolve/<query>` returning a `NameRegistration` on 200 for every
+registration shape, 400 for unknown TLDs, 502 for upstream failures) or
+substitute a different transport returning the same JSON. The resolver API is
+versioned separately from this protocol: `/v1/resolve/<name>` returns a bare
+`NameRecord` and is what relays before v22 call as `/resolve/<name>`.
 
 #### Resolve name command
 
-The `RSLV` command carries the canonical fully-qualified name directly as the
-payload (not JSON):
+The `RSLV` command carries the query as text, not JSON. A client sends the
+hashed form only from v22, and the name itself below it:
 
 ```abnf
-rslv = %s"RSLV" SP domain   ; domain = canonical name as non-space bytes, consuming the remainder of the transmission
+rslv   = %s"RSLV" SP query
+query  = domain / hashed        ; hashed only from v22
+domain = 1*253 OCTET            ; the name as text
+hashed = "[" 64HEXDIG "]" tld   ; keccak-256 of the second-level label
+tld    = %s".simplex" / %s".testing"
 ```
 
 `domain` is the UTF-8 canonical fully-qualified name with the TLD always
 explicit (e.g. `privacy.simplex`, `test.testing`, `example.com`), bounded to
 253 bytes.
 
+**Hashed labels.** The query is a name, or the keccak-256 of a second-level
+label in ENS's bracketed form. A label can only be letters, digits and hyphens,
+so `[` tells the two apart and no tag is needed.
+
+Only a second-level name may be hashed. A name with subnames is sent as text: it
+is resolved rather than priced, and its record names it anyway. A web TLD has no
+registry to key a hash on.
+
+From v22 a client MUST send the hash. Older routers can only read the name, so a
+client on an older session sends the name. A router answering a hashed query
+does not know the label's length, so it cannot check a minimum-length policy
+either: the client does that, from the pricing it is sent.
+
+The same form reaches the backing resolver, which is what its HTTP API takes, so
+the query is one string end to end.
+
+A hashed query still answers with the name. The registrar records the plaintext
+label when a name is registered, keyed by the hash of that label, so a router can
+look up what the hash stands for without ever being told. The router is not
+trusted for it: a client MUST check that the record names the name it asked
+about, and reject the answer otherwise. A registry that does not record the
+label cannot answer a hashed query at all, and the router answers `ERR NAME
+RESOLVER` rather than a record it knows the client will reject. An unregistered
+name has no recorded label, so it cannot be looked up either.
+
 **Server-side validation.** The names router parses `domain` as a
 fully-qualified name (TLD required — bare labels are rejected) and forwards it
 to the configured backing resolver, which is the source of truth for which
 on-chain registry maps to each TLD.
 
-The names router responds with either an `RNAME` response carrying the resolved
-record, or an `ERR NAME` error whose subcode a client iterating across several
-configured servers can act on distinctly:
+The names router responds with either an `RNAME` response saying what it knows
+about the name, or an `ERR NAME` error whose subcode a client iterating across
+several configured servers can act on distinctly:
 
 | Response | Condition | Client action |
 |---|---|---|
-| `RNAME` | record resolved | use it |
-| `ERR NAME NOT_FOUND` | name not registered, unknown TLD, or malformed name | authoritative "no such name" — stop |
+| `RNAME` | the router read the registry | use it |
+| `ERR NAME NOT_FOUND` | below v22 only: every name that does not resolve. From v22 a router never sends it | stop, and do not read it as registrable |
 | `ERR NAME NO_RESOLVER` | this router has no resolver (names role not enabled) | skip this server, try the next |
-| `ERR NAME RESOLVER <detail>` | transient failure: backing resolver error (upstream 5xx, transport, timeout, decode) | transient — retry or surface, do not treat as "not found" |
+| `ERR NAME RESOLVER <detail>` | the router cannot state an answer completely: no registrar or price oracle for the TLD, an unreachable chain, a transport failure, a timeout, a registration it could not date or resolve | surface `<detail>`; retry only if it reads as transient |
 
 A client SHOULD NOT broadcast a `name` to further servers after a name-capable
 router has answered (`NOT_FOUND` or `RESOLVER`), since that router has already
 seen the lookup key; `NO_RESOLVER` discloses nothing about the name beyond the
 fact that this router cannot resolve, so iterating past it is safe.
 
-#### Name record response
+#### Name response
 
-The `RNAME` response carries a JSON-encoded record as the payload:
+`RNAME` answers both what a name resolves to and whether it can be registered.
 
 ```abnf
-rname = %s"RNAME" SP json-bytes   ; json-bytes consumes the remainder of the transmission
+rname = %s"RNAME" SP nameResponse
 ```
 
-`json-bytes` MUST be a UTF-8 JSON object with the following schema:
+`nameResponse` is a UTF-8 JSON object consuming the remainder of the
+transmission: the `registration` below, and `lastBlockTs`, the block timestamp the
+registry was read at. Money is US cents, times are seconds since the Unix epoch, and lengths are
+characters.
+
+| Field | JSON type | Constraints |
+|---|---|---|
+| `lastBlockTs` | number | block timestamp the answer was read at. Absent only from a v20/v21 router, which sent the record alone |
+| `registration` | object | below |
+
+A router reads the registry through a node of its own, which can lag. A name
+registered or changed after `lastBlockTs` still reads as it did before, so a client
+MUST NOT treat an answer as current without checking it.
+
+`registration`'s `type` selects which of the three answers it is.
+
+| `type` | Meaning |
+|---|---|
+| `registered` | held by someone until `expires`, renewable by its owner alone until `graceUntil`. It always carries `nameRecord`: where the owner set none, every field is unset and the resolver address is zero |
+| `available` | held by nobody and registrable now, at `pricing` |
+| `reserved` | held back by the registry and not registered. It carries no price, and a router MUST NOT quote one |
+
+| Field | On | JSON type | Constraints |
+|---|---|---|---|
+| `expires` | `registered` | number | absent only from a v20/v21 router, which sent the record alone |
+| `graceUntil` | `registered` | number | after `expires`; until here only the owner may renew. Absent on the same condition as `expires` |
+| `reservedReason_` | `registered` | string | a reason word, present only when the name is held back as well |
+| `nameRecord` | `registered` | object | the record, schema below |
+| `pricing` | `available` | object | `registrationPrices`, `basePrice` and `minLabelLength`, below |
+| `reservedReason` | `reserved` | string | a reason word |
+
+| `pricing` field | JSON type | Constraints |
+|---|---|---|
+| `registrationPrices` | object | label length, as a decimal string, to US cents per year, for the lengths the registry prices specially |
+| `basePrice` | number | US cents per year for every other length |
+| `minLabelLength` | number | characters; the registry refuses shorter labels |
+
+A reason word is `internal`, `trademark`, `community`, or a word a later version
+reserves under, at most 32 printable ASCII characters. A router truncates an
+unknown word to that and otherwise passes it through unchanged. A client MUST
+read a word it does not know as unknown and still treat the name as reserved.
+
+**Computing the price.** In US cents, for a duration in seconds:
+
+```
+price len duration = tier len * duration / 31536000
+tier len = the entry for len in registrationPrices, or basePrice when there is none
+```
+
+The registry's minimum registration is 730 days, a contract constant, so it is
+specified here rather than sent. `registrationPrices` omits any length below
+`minLabelLength`. A client MUST NOT show a quote for a label the registry
+refuses: a hashed query carries no length, so only the client can check it.
+
+Below v22, `RNAME` carries the bare record and nothing else, and every answer
+without one is `ERR NAME NOT_FOUND`. A v22 client reads such an answer as
+`registered` with no expiry, grace or reservation.
+
+From v22 a client MUST NOT read `ERR NAME NOT_FOUND` as "registrable": only
+`available` says that.
+
+A router that cannot state an answer completely MUST send `ERR NAME RESOLVER
+<detail>` rather than answer partially or guess. That covers a TLD with no
+registrar or price oracle, an unreachable chain, a timeout, a registration it
+could not date or resolve, and any status word it does not recognise.
+
+`nameRecord` MUST be a UTF-8 JSON object with the following schema:
 
 | Field | JSON type | Constraints |
 |---|---|---|
@@ -1514,34 +1616,16 @@ rname = %s"RNAME" SP json-bytes   ; json-bytes consumes the remainder of the tra
 | `owner` | string | `"0x"` followed by 40 lowercase hex characters (20 raw bytes) |
 | `resolver` | string | `"0x"` followed by 40 lowercase hex characters; the resolver contract address that produced the record |
 
-Text fields (`nickname`, `website`, `location`) use the empty string `""` as
-the "unset" sentinel: a backing resolver with no value for the field MUST emit
-an empty string, not JSON `null` and not an absent key. Link fields
-(`simplexContact`, `simplexChannel`) are arrays, primary link first, and use the
-empty array `[]` when unset. Coin fields (`eth`, `btc`, `xmr`, `dot`) use JSON
-`null` as the "unset" sentinel and MAY also be absent from the object entirely.
+Testnet-vs-mainnet status is derived from the queried TLD, not from the record.
 
-The backing resolver filters records that are expired or otherwise unavailable
-(the names router then returns `ERR NAME NOT_FOUND` to the client), so the wire
-format carries no expiry field. Testnet-vs-mainnet status is derived from the
-queried TLD rather than an in-record flag.
+Receivers MUST tolerate extra unknown fields; adding a required field is a
+breaking change requiring an SMP version bump. Receivers parse by key name, so
+peers MUST NOT rely on a byte-canonical form.
 
-Receivers MUST tolerate extra unknown fields (forward-compatibility for future
-field additions). Adding a required field is a breaking change requiring an
-SMP version bump.
-
-**Field order is not significant.** Receivers parse JSON by key name, so object
-key order, insignificant whitespace, and number formatting carry no meaning;
-records are interpreted by decoded value, never compared byte-for-byte. Peers
-MUST NOT rely on a byte-canonical form — a different resolver or server may emit
-the same record with different key order or spacing. This order-independence is
-what makes the format forward-compatible (see the unknown-field rule above).
-
-**Wire-size budget.** The names router caps the resolver response it will
-accept (`resolver_max_response_bytes`, ≤ 16000 bytes, the default) so the
-re-encoded `RNAME` stays within the SMP proxied transmission budget of 16224
-bytes; a response over the cap is rejected as `ERR NAME RESOLVER`. The link
-arrays are bounded by this overall budget rather than a fixed per-field count.
+The names router caps the resolver response it will accept
+(`resolver_max_response_bytes`, at most 16000 bytes) so the re-encoded `RNAME`
+stays within the SMP proxied transmission budget of 16224 bytes; a response over
+the cap is `ERR NAME RESOLVER`.
 
 ## Transport connection with the SMP router
 

protocol/xrcp.md

@@ -71,7 +71,7 @@ The session invitation contains this data:
 
 Host application decrypts (except the first session) and validates the invitation:
 - Session signature is valid.
-- Timestamp is within some window from the current time.
+- Timestamp of a multicast announcement is not earlier than 3660 seconds before and not later than 3600 seconds after the current time of the host. The host ignores announcements outside of this interval and continues listening.
 - Long-term key signature is valid.
 - Long-term CA and signature key are the same as in the first session.
 - Some version in the offered range is supported.
@@ -121,9 +121,15 @@ During TLS handshake, parties validate certificate chains against previously kno
 
 ### Session verification and protocol negotiation
 
-Once TLS session is established, both the host and controller devices present a "session security code" to the user who must match them (e.g., visually or via QR code scan) and confirm on the host device. The session security code must be a digest of tlsunique channel binding. As it is computed as a digest of the TLS handshake for both the controller and the host, it will validate that the same TLS certificates are used on both sides, and that the same TLS session is established, mitigating the possibility of MITM attack in the connection.
+Once TLS session is established, both the host and controller devices present a "session security code" to the user. The session security code must be a digest of tlsunique channel binding. As it is computed as a digest of the TLS handshake for both the controller and the host, it will validate that the same TLS certificates are used on both sides, and that the same TLS session is established, mitigating the possibility of MITM attack in the connection.
 
-Once the session is confirmed by the user, the host sends HELLO block to the controller.
+To verify the session, the user matches the codes (e.g., visually or via QR code scan) and confirms the session on the host device.
+
+The user MUST verify the first session with a new controller. The host stores the long-term identity of the controller (CA fingerprint and Ed25519 public key) only after this session is verified. A controller with the stored long-term identity is a known controller.
+
+The user MAY verify the subsequent sessions with a known controller. In a session that the host confirms without user action, the controller is authenticated only by its long-term identity stored in the first session (see [Threat model](#threat-model)). SimpleX Chat mobile apps require user verification of sessions with known controllers only when the user enables the "Verify connections" option.
+
+Once the session is confirmed, the host sends HELLO block to the controller.
 
 XRCP blocks inside TLS are padded to 16384 bytes.
 
@@ -305,7 +311,7 @@ to receive: (rcvKey', sk, nonce) = HKDF(rcvKey, "SimpleXSbChain", 88)
 - prevent host and controller devices from establishing the session
 
 *cannot:*
-- same as passive adversary, provided that user visually verified session code out-of-band.
+- same as passive adversary, provided that user visually verified session code out-of-band in the first session with the controller.
 
 #### An active adversary with the access to the network:
 
@@ -324,6 +330,15 @@ to receive: (rcvKey', sk, nonce) = HKDF(rcvKey, "SimpleXSbChain", 88)
 *cannot:*
 - connect to the host or make host connect to itself.
 
+#### An active adversary with the access to the network who also obtained the long-term private keys of a known controller (e.g., from the controller device):
+
+*can:*
+- make host connect to itself instead of the controller.
+- access any data of the controlled host application, within the capabilities of the provided API, if the host confirms the session without user action.
+
+*cannot:*
+- access host application data, provided that user visually verified session code out-of-band.
+
 #### Compromised controller device:
 
 *can:*

scripts/ci/prune-old-builds.sh

@@ -0,0 +1,9 @@
+#!/bin/sh
+set -eu
+
+version=$(awk '/^version:/ {print $2}' simplexmq.cabal)
+for dir in "$@"; do
+  if [ -d "$dir/build" ]; then
+    find "$dir/build" -mindepth 3 -maxdepth 3 -type d -name 'simplexmq-*' ! -name "simplexmq-$version" -exec rm -rf {} +
+  fi
+done

scripts/ci/source-timestamps.sh

@@ -0,0 +1,27 @@
+#!/bin/sh
+set -eu
+
+command=$1
+hashes=$2
+
+sources() {
+  git ls-files -z '*.hs' '*.hsc'
+}
+
+case "$command" in
+  record)
+    sources | xargs -0 sha256sum > "$hashes"
+    ;;
+  restore)
+    sources | xargs -0 touch
+    while read -r hash file; do
+      if [ -f "$file" ] && [ "$(sha256sum "$file" | cut -d' ' -f1)" = "$hash" ]; then
+        touch -d 2000-01-01 "$file"
+      fi
+    done < "$hashes"
+    ;;
+  *)
+    echo "usage: $0 record|restore <hashes file>" >&2
+    exit 1
+    ;;
+esac

scripts/resolver/.env.example

@@ -1,5 +1,8 @@
+# Copy to .env (next to docker-compose.yml) only to change a setting; without it
+# docker-compose.yml uses the values shown here.
+
 # ============================================================================
-# Required settings — the stack will not start without these.
+# Chain settings.
 # ============================================================================
 
 # Ethereum network: mainnet (the SNRC `.testing` contracts live on mainnet)
@@ -14,9 +17,16 @@ NETWORK=mainnet
 TRUSTED_NODE_URL=https://mainnet-checkpoint-sync.attestant.io
 
 # ============================================================================
-# Optional overrides — sensible defaults are baked into docker-compose.yml,
-# so leave these commented unless you need to change them.
+# Optional overrides — leave commented unless you need to change them.
 # ============================================================================
 
 # Nimbus NAT (default: any). For a stable public node set an explicit IP:
 #   NAT=extip:1.2.3.4        # your public IPv4: curl -s ifconfig.me
+
+# Resolver (defaults in the README's "Load and scaling" and "Logs" sections)
+#   SNRC_WORKERS=4                          # worker processes; default CPU count, at most 4
+#   SNRC_RPC_TIMEOUT=5                      # seconds per request to reth
+#   SNRC_LOG_FORMAT=json                    # default text (key=value)
+#   SNRC_LOG_COLOR=always                   # coloured `docker compose logs`
+#   SNRC_LOG_LEVEL=debug                    # also log health checks and idle connections
+#   SNRC_TRUSTED_PROXIES=172.16.0.0/12      # behind a reverse proxy on the host; see README

scripts/resolver/README.md

@@ -19,15 +19,15 @@ against **Ethereum mainnet** (where the `.testing` contracts live):
 
 ## 1. Configure
 
-Edit `.env` — the defaults work as-is; override only if needed:
+The defaults in `docker-compose.yml` work as-is:
 
 ```sh
 NETWORK=mainnet                                               # default
 TRUSTED_NODE_URL=https://mainnet-checkpoint-sync.attestant.io # default
 ```
 
-Everything else (NAT) has a working default baked into `docker-compose.yml`;
-uncomment the hints in `.env` only to override.
+To override these or NAT, copy `.env.example` to `.env` and edit it; `.env` is
+not committed.
 
 ## 2. Run
 
@@ -62,15 +62,26 @@ curl -s -X POST http://127.0.0.1:8545 \
 **2. resolver is healthy:**
 ```sh
 curl -s http://127.0.0.1:8000/health | jq
-# → {"ok": true, "rpc": "http://reth:8545", "registries": {"testing": "0x…", "simplex": ""}}
+# → {"ok": true, "rpc": "http://reth:8545", "registries": {"testing": "0x…", "simplex": ""},
+#    "blockNumber": 23400000, "chainLagSeconds": 12}
 ```
 
+`chainLagSeconds` is how far the node is behind the wall clock. A resolver that
+is reachable and answering can still be hours behind, and every name it reports
+is that stale. `null` means the node could not be reached.
+
 **3. resolver resolves a live name** (`foobar.testing` is a populated test name):
 ```sh
 curl -s http://127.0.0.1:8000/resolve/foobar.testing | jq
 # → {"name":"foobar.testing","nickname":"Foo","simplexContact":["https://smp16.simplex.im/a#…"], … }
 ```
 
+**4. the route your router will call** (check 3 passes on an older resolver too):
+```sh
+curl -s http://127.0.0.1:8000/v2/resolve/foobar.testing | jq
+# → {"type":"registered","expires":1780000000,"graceUntil":…,"nameRecord":{…}}
+```
+
 **Wire your smp-server:** in its `[NAMES]` section set
 `resolver_endpoint: http://127.0.0.1:8000` (no auth needed for loopback).
 
@@ -82,7 +93,7 @@ curl -s http://127.0.0.1:8000/resolve/foobar.testing | jq
 | reth p2p | `:30303` tcp/udp | Ethereum sync (open on firewall) |
 | nimbus p2p | `:9000` tcp/udp | beacon sync (open on firewall) |
 | nimbus REST | `127.0.0.1:5052` | beacon API |
-| **resolver** | `127.0.0.1:8000` | SNRC REST (`/resolve`, `/health`) |
+| **resolver** | `127.0.0.1:8000` | SNRC REST (`/v2/resolve`, `/resolve`, `/health`) |
 
 ## Caveats
 
@@ -110,7 +121,66 @@ standalone for local dev (no Docker), via [`uv`](https://docs.astral.sh/uv/):
 uv run scripts/resolver/service/snrc-resolve.py  # defaults to local reth + mainnet .testing
 ```
 
-### Response shape
+Tests and lint run from `scripts/resolver/service`:
+
+```sh
+uv run pytest
+uv run ruff check .
+```
+
+Dependencies are locked in `uv.lock`; the Docker image installs exactly those.
+
+Three routes, versioned separately from the protocol so each only changes when
+its own shape does:
+
+| Route | Called by | Answers |
+|---|---|---|
+| `/v2/resolve/<query>` | routers from SMP v22 | a `NameRegistration` |
+| `/resolve/<name>` | routers before SMP v22 | a name record, flat |
+| `/health` | anyone | readiness |
+
+`/v1/resolve/<name>` is an alias for `/resolve/<name>`.
+
+### v2: `/v2/resolve/<query>`
+
+The body is the SMP protocol's `NameResponse`, which the router decodes as is
+and forwards; translating the registry's model to it is this resolver's job. It
+is the registration below, plus `lastBlockTs`, the timestamp of the block it was read
+at: a node that lags answers with names it has not seen registered yet.
+Its `type` is `registered`, `available` or `reserved`, and the fields each one
+carries are specified once, in the **Name response** section of
+[`protocol/simplex-messaging.md`](../../protocol/simplex-messaging.md). It is
+the wire format, so it is documented with the wire.
+
+Two things follow from that and are worth stating here. Expiry and grace belong
+to the registration, not to the record: `expires` and `graceUntil` sit beside
+`nameRecord`, not inside it. And a name nobody holds is not an error: it
+answers 200 with `type: available` and its price, so no status code from this
+route means "not registered".
+
+| Status | Meaning |
+|---|---|
+| 200 | a registration: `registered`, `available` or `reserved` |
+| 400 | `tldNotConfigured`, `notFullyQualified` |
+| 502 | `noPriceOracle`, `labelNotRecorded`, an unreadable status, or `upstreamError` |
+
+Error bodies carry `name` and a fixed `error` code to branch on. Only
+`upstreamError` adds a `message`; the v1 route always adds one.
+
+`labelNotRecorded` means the registrar holds the name but never recorded its
+label, so a hashed query cannot be answered with a name. See
+[Querying by labelhash](#querying-by-labelhash).
+
+A subname reports the expiry and grace of the 2LD above it, since that is what
+bounds its lifetime. A subname nobody created reports as not registered.
+
+### v1: `/resolve/<name>`
+
+What routers before SMP v22 call. Its shape is unrelated to v2's: the record is
+flat, and `status`, `expires`, `graceEnds`, `reasonCode` and `reason` sit
+alongside its fields.
+
+#### v1 response shape
 
 ```jsonc
 {
@@ -119,28 +189,281 @@ uv run scripts/resolver/service/snrc-resolve.py  # defaults to local reth + main
   "simplexContact": ["https://smp16.simplex.im/a#…", "https://smp11…"],  // primary first, fallbacks after
   "simplexChannel": [],
   "eth": null, "btc": "bc1q…", "xmr": "4ANz…", "dot": "139G…",
-  "owner": "0xd83b…", "resolver": "0x80fa…"
+  "owner": "0xd83b…", "resolver": "0x80fa…",
+  "status": "registered",      // registered | grace | expired | unregistered | unknown
+  "expires": 1780000000,       // Unix seconds; when the registration ends
+  "graceEnds": 1787776000,     // expires + GRACE_PERIOD; last moment the owner can renew
+  "reasonCode": null,          // set when the name is held back as well
+  "reason": null               // set when the name is held back as well
 }
 ```
 
 `simplexContact`/`simplexChannel` are arrays (a name can advertise multiple SMP
-servers; clients try them in order). On-chain they're a single comma-separated
+servers; clients try them in order). On-chain they're a single `;`-separated
 text record; the resolver splits/trims/drops-empties. Address encodings are
 canonical per chain (EIP-55 / bech32 / SS58 / Monero-base58). Subnames work
 identically (`bar.foobar.testing`).
 
-### Status codes
+#### v1 registration status and expiry
+
+A response carries `status`, `expires` and `graceEnds` whenever the resolver
+read them, a successful resolve included, so a client that has just resolved a
+name already knows when it expires. Both timestamps are Unix seconds, and
+`null` when they could not be read.
+
+| `status` | Meaning |
+|---|---|
+| `registered` | live; `expires` is when that ends |
+| `grace` | lapsed, but only the previous owner may renew it, until `graceEnds` |
+| `expired` | lapsed and past grace; anyone may register it |
+| `unregistered` | never registered, and free to take |
+| `unknown` | no `SNRC_REGISTRAR_<TLD>` configured, so status could not be read |
+
+A reservation is orthogonal to the status: a name held back by the registry
+carries `reasonCode` and `reason` whether or not it is registered.
+
+`grace` and `expired` are told apart by the registrar's own `available(id)`
+rule, `expires + GRACE_PERIOD < now`. `GRACE_PERIOD` is read from the contract
+rather than assumed, and `now` is the latest block's timestamp rather than the
+host clock, which the registrar compares against too, so a machine with a wrong
+clock cannot misreport a registration. That rule alone is not enough: it also
+holds for a name nobody ever registered (`0 + GRACE_PERIOD < now`), so a zero
+expiry is what separates *never registered* from *registered and since
+released*.
+
+A subname reports the status of the 2LD above it, which is only as good as the
+name it sits under. A subname nobody created answers 404 `unregistered`.
+
+#### v1 errors
+
+Every non-2xx body carries two fields: `error` is a fixed code to branch on,
+and `message` is a sentence for a human. Match on `error`, never on `message`,
+which is free to change.
+
+```jsonc
+{"name": "nope.testing", "error": "unregistered",
+ "message": "this name has never been registered",
+ "status": "unregistered", "expires": null, "graceEnds": null}
+```
+
+The codes are `tldNotConfigured`, `notFullyQualified`, `unregistered`,
+`expired`, `noSuchRoute` and `upstreamError`. When the registration is what went
+wrong, `error` and `status` hold the same value, so one field is enough to read.
+
+`upstreamError` says only which exception type the RPC call raised. The text
+goes to the resolver's log instead, because `SNRC_RPC` can carry a provider key
+and the exception can carry the URL it failed on. It is also the answer
+when a registrar, controller or oracle address has no contract behind it: the
+empty reply is refused rather than read as zero, which would make every name
+look free.
+
+#### v1 status codes
 
 | Status | Meaning |
 |---|---|
-| 200 | resolved |
+| 200 | resolved (`status` is `registered` or `grace`, or `unknown` when no registrar is configured) |
 | 400 | TLD not configured, or not a fully-qualified name |
-| 404 | name has no resolver set on the registry |
+| 404 | `unregistered` |
+| 410 | `expired`: lapsed and past grace, so anyone may take it |
 | 502 | upstream RPC error / reth not synced |
 
-### Configuring registries
+### Querying by labelhash
 
-Defaults to mainnet `.testing` (`0x03f438…`); `.simplex` is unset until
-deployed. Override per TLD via env on the `resolver` service in
-`docker-compose.yml` (`SNRC_REGISTRY_TESTING` / `SNRC_REGISTRY_SIMPLEX`), or as
-env vars for the standalone script.
+A client asking whether a name is free is usually about to register it, and
+whoever runs the resolver could register it first. To avoid that, send the
+keccak hash of the label in ENS's `[<64 hex>]` form instead of the label:
+
+```sh
+# instead of /resolve/acme.testing
+curl -s "http://127.0.0.1:8000/resolve/[$(printf acme | keccak-256sum | cut -d' ' -f1)].testing"
+```
+
+namehash is `keccak(parent || keccak(label))`, so this reaches the same node and
+returns the same record. The registrar keys `nameExpires` and `reservedNames` on
+the labelhash too, so the status fields do not need the label either. The
+resolver learns the name only by guessing the label and hashing it.
+
+Only the second-level label is a registry key, and `status` decodes a bracket
+there at any depth. The record does not: a bracket is decoded only in a
+two-label name, so `sub.[<hash>].testing` is not a supported query. Subname
+labels stay text; a bracket label left of the 2LD is an ordinary label. Routers
+from v22 send every 2LD this way, so a registrable name normally never reaches
+this service.
+
+On v2, read `type`: only `available` means the name is free. On v1, read
+`status`: a name is free on `unregistered` (404) and on `expired` (410), every
+other status means somebody holds it, and a `reasonCode` means the registry will
+refuse it whatever the status says.
+
+The hash must be keccak-256. `openssl dgst -sha3-256` and `sha3sum` compute
+SHA3-256, a different function that returns 64 valid-looking hex characters
+pointing at the wrong node.
+
+The resolver lowercases the query before matching, so uppercase hex works too.
+Clients that refuse raw brackets in a path can percent-encode them as `%5B` and
+`%5D`.
+
+Brackets cannot collide with a real name: they are invalid in a normalised ENS
+name, and a `[<64 hex>]` label is 66 bytes against the registrar's
+`maxLabelLength` of 63. A plain `0x…` label is not treated as a hash, since that
+is an ordinary, registrable name.
+
+Only 2LDs can be queried this way, as only a 2LD can be raced for: subnames are
+created by the 2LD's owner. A bracket label in a subname is hashed as written,
+so it points at a node nobody can own. ENS tooling accepts the bracketed form at
+any depth; this resolver does not, on purpose.
+
+This hides interest in a name and nothing else: the registration itself is
+public, and commit-reveal covers that step. A short or well-known label is easy
+to guess by hashing candidates, and the reveal publishes the labelhash, so an
+operator who logged the query can match it to the name afterwards.
+
+### What a name costs
+
+The controller's `prices()` names the price oracle, so no extra configuration is
+needed beyond `SNRC_CONTROLLER_<TLD>`. A `SimplexPriceOracle` exposes its curve
+through `prices()`, in US cents per year, which is the unit this API carries.
+
+An ENS-shaped oracle exposes only `price1Letter()`..`price6Letter()`, in attoUSD
+per second, and charges a premium on a lapsed name that it does not expose. A
+quote from one is therefore only safe for a name that was never registered: an
+`expired` name gets no price rather than one below what the registrar charges.
+
+**Set `SNRC_CONTROLLER_<TLD>` wherever `SNRC_REGISTRAR_<TLD>` is.** Without a
+controller there is no oracle, so no name can be priced.
+
+**Upgrade this service before the routers that query it.** Routers from SMP v22
+call `/v2/resolve`, which an older resolver does not serve. Every name then
+answers `ERR NAME RESOLVER "HTTP 404"` until this service is upgraded, while
+`/health` still reports it as ready.
+
+### Why a name is reserved
+
+v2 carries the controller's reason as `reservedReason`. v1 carries the same word
+as `reasonCode`, plus `reason`, an English sentence for a human reading this API.
+Clients should branch on the word and phrase it themselves, in the user's
+language.
+
+| `reasonCode` | Meaning |
+|---|---|
+| `internal` | reserved for SimpleX |
+| `trademark` | reserved to protect a trademark |
+| `community` | reserved for the community |
+| `unknown` | a reason added to the contract after this resolver; still reserved |
+
+These are `SimplexController.Reason`, where 0 means not reserved. A controller
+from before the enum stores a boolean, whose `true` decodes as 1, which is why
+1 reads as `internal`, so nothing needs migrating.
+
+### Configuring addresses
+
+The resolver reads three contracts, each configured per TLD.
+
+The **registry** answers who owns a node, and `/resolve` reads the records from
+it. The **registrar** (ERC-721) holds `nameExpires` and `GRACE_PERIOD`, which
+is where every expiry field comes from, and `labelOf`, which is how a hashed
+query is answered with a name. A name registered without recording its label
+cannot answer one, and `/v2/resolve` refuses it rather than answer with a name
+the client will reject. With no registrar for a TLD, `/resolve` still works and
+reports `"status": "unknown"`. The **controller** holds
+`reservedNames`, which is where `reasonCode` comes from. With no controller a
+held-back name reads as not reserved, and no name can be priced.
+
+All three default to the mainnet `.testing` deployment. `.simplex` is unset
+until it is deployed.
+
+The controller default is the **proxy**, not `SimplexControllerImpl`. Storage
+lives in the proxy, so the implementation address answers nothing. The two
+deployment files use different names for that proxy:
+`deployments.mainnet.testing.json` records it under the ENS role name
+`ETHRegistrarController`, and `verification.mainnet.testing.json` calls it
+`SimplexControllerProxy`. Both are the same address, and it is the one used
+here.
+
+To override any of them, set `SNRC_REGISTRY_<TLD>`, `SNRC_REGISTRAR_<TLD>` or
+`SNRC_CONTROLLER_<TLD>` on the `resolver` service in `docker-compose.yml`, or
+as env vars when you run the script directly.
+
+### Load and scaling
+
+A lookup reads the chain in at most three rounds: the name's status and
+registry entries, then its record from its resolver, then prices for a name
+that is free. The node runs the calls of a JSON-RPC batch one after another,
+so the contract reads of a round go to the chain as one `eth_call` to
+[Multicall3](https://github.com/mds1/multicall). A node that is slow for a
+moment therefore delays a lookup a few times, not once per read. Connections to
+the node are kept open and reused.
+
+The resolver answers HTTP/1.1, so each smp-server keeps its connections to it
+open instead of connecting for every lookup. An idle connection is closed after
+60 s; the smp-server drops its own idle ones after 30 s, so it closes them first.
+
+Each access log line ends with how long the request took, so slow requests
+show up in `docker compose logs resolver`.
+
+| Variable | Default | Meaning |
+|---|---|---|
+| `SNRC_WORKERS` | CPU count, at most 4 | processes sharing the port. If one exits, the others stop, so the container restarts |
+| `SNRC_RPC_TIMEOUT` | `5` | seconds to wait for each request to the node; the smp-server gives up after 3 |
+| `SNRC_MULTICALL` | `0xcA11bde05977b3631167028862bE2a173976CA11` | Multicall3 address. If it does not answer, each round is sent as a plain batch and the log says so once |
+
+### Logs
+
+Each event is one line: UTC time, level, event name, then `key=value` fields.
+
+```
+2026-09-26T08:06:19.992Z INFO  request  client=203.0.113.7 worker=12 method=GET path=/v2/resolve/[4fdd…].testing status=200 bytes=701 ms=35
+2026-09-26T08:06:20.051Z WARN  upstream_error  name=foobar.testing error=ConnectionRefusedError message="[Errno 111] Connection refused"
+```
+
+| Event | Level | Meaning |
+|---|---|---|
+| `listening` | info | started: address, workers, RPC endpoint, registries, trusted proxies |
+| `request` | info | one answered request; `/health` only at debug |
+| `upstream_error` | warn | a read from the node failed; the caller got 502 |
+| `multicall_unavailable` | warn | Multicall3 did not answer; rounds are sent as plain batches |
+| `client_gone` | warn | the caller hung up before the answer, usually an smp-server past its timeout |
+| `http_error` | warn | a malformed request; idle keep-alive timeouts only at debug |
+| `request_failed` | error | a request failed unexpectedly, with its traceback |
+| `worker_exited`, `worker_failed` | error | a worker stopped; the others stop too, so the container restarts |
+| `stopping` | info | stopped by a signal |
+
+| Variable | Default | Meaning |
+|---|---|---|
+| `SNRC_LOG_FORMAT` | `text` | `text` (`key=value`) or `json`, one object per line |
+| `SNRC_LOG_COLOR` | `auto` | `auto` colours only a terminal, `always`, `never`. Docker output is no terminal, so set `always` for coloured `docker compose logs` |
+| `SNRC_LOG_LEVEL` | `info` | `debug` also logs health checks and idle connections closing |
+| `SNRC_TRUSTED_PROXIES` | none | addresses or CIDRs of reverse proxies whose `X-Forwarded-For` names the client |
+
+With docker compose, set these, and those in "Load and scaling", in `.env`
+(`.env.example` lists them).
+
+Behind a reverse proxy such as Caddy on the host, the resolver sees the Docker
+network's gateway rather than the client. Trust that gateway, and the logged
+client is the last `X-Forwarded-For` address no trusted proxy added, so a caller
+through the proxy cannot claim another address. Everything else on the host that
+connects to `127.0.0.1:8000`, the smp-servers included, arrives through the same
+gateway and could set the logged address too; run the proxy in the compose network
+and trust only its address to avoid that.
+
+The gateway is
+`docker network inspect resolver_default --format '{{(index .IPAM.Config 0).Gateway}}'`,
+with the network named after the compose project. It changes when the network is
+recreated, as `docker compose down` does, so trust one of, in `.env`:
+
+- `SNRC_TRUSTED_PROXIES=172.16.0.0/12`: Docker's default pools for compose
+  networks, whatever subnet it picks. This also trusts every other container on the
+  host. Add the gateway too if Docker ever gives the network a `192.168.x.x` one.
+- a pinned subnet, so the gateway stays fixed, and only that gateway:
+
+  ```yaml
+  networks:
+    default:
+      ipam:
+        config:
+          - subnet: 172.30.0.0/24
+            gateway: 172.30.0.1
+  ```
+
+  in `docker-compose.yml`, with `SNRC_TRUSTED_PROXIES=172.30.0.1`, on a subnet no
+  other network on the host uses.

scripts/resolver/docker-compose.yml

@@ -38,14 +38,14 @@ services:
       - sh
       - -c
       - |
-        if [ -d /home/user/nimbus-eth2/build/data/${NETWORK}/db ]; then
+        if [ -d /home/user/nimbus-eth2/build/data/${NETWORK:-mainnet}/db ]; then
           echo "Nimbus data dir already initialised — skipping checkpoint sync";
           exit 0;
         fi;
         /home/user/nimbus-eth2/build/nimbus_beacon_node trustedNodeSync \
-          --network=${NETWORK} \
-          --data-dir=/home/user/nimbus-eth2/build/data/${NETWORK} \
-          --trusted-node-url=${TRUSTED_NODE_URL} \
+          --network=${NETWORK:-mainnet} \
+          --data-dir=/home/user/nimbus-eth2/build/data/${NETWORK:-mainnet} \
+          --trusted-node-url=${TRUSTED_NODE_URL:-https://mainnet-checkpoint-sync.attestant.io} \
           --backfill=false
     restart: "no"
 
@@ -69,8 +69,8 @@ services:
           echo "Reth data already initialised — skipping snapshot download";
           exit 0;
         fi;
-        echo "Downloading Reth ${NETWORK} --minimal snapshot...";
-        reth download --datadir /data --chain ${NETWORK} --minimal && \
+        echo "Downloading Reth ${NETWORK:-mainnet} --minimal snapshot...";
+        reth download --datadir /data --chain ${NETWORK:-mainnet} --minimal && \
         touch /data/.snapshot-done && \
         echo "Snapshot download complete"
     restart: "no"
@@ -92,7 +92,7 @@ services:
     command: >
       node
       --datadir /data
-      --chain ${NETWORK}
+      --chain ${NETWORK:-mainnet}
       --minimal
       --authrpc.jwtsecret /jwt/jwt.hex
       --authrpc.addr 0.0.0.0 --authrpc.port 8551
@@ -117,8 +117,8 @@ services:
       - "9000:9000/udp"
       - "127.0.0.1:5052:5052"
     command: >
-      --network=${NETWORK}
-      --data-dir=/home/user/nimbus-eth2/build/data/${NETWORK}
+      --network=${NETWORK:-mainnet}
+      --data-dir=/home/user/nimbus-eth2/build/data/${NETWORK:-mainnet}
       --el=http://reth:8551
       --jwt-secret=/jwt/jwt.hex
       --non-interactive
@@ -131,6 +131,7 @@ services:
   # smp-server points its [NAMES] resolver_endpoint at this URL.
   # To change the host port, edit the LEFT side of the port mapping below.
   resolver:
+    image: snrc-resolve:local
     build:
       context: ./service
       dockerfile: Dockerfile
@@ -150,8 +151,24 @@ services:
       # only if you're deploying against a different network or contract.
       # SNRC_REGISTRY_TESTING: 0x...
       # SNRC_REGISTRY_SIMPLEX: 0x...
+      # Registrar and controller, same cascade. Without the registrar `status`
+      # is "unknown"; without the controller a reserved name is "unregistered".
+      # SNRC_REGISTRAR_TESTING: 0x...
+      # SNRC_REGISTRAR_SIMPLEX: 0x...
+      # SNRC_CONTROLLER_TESTING: 0x...
+      # SNRC_CONTROLLER_SIMPLEX: 0x...
+      # Set in .env (see .env.example); unset or empty means the resolver's default.
+      SNRC_WORKERS: ${SNRC_WORKERS:-}
+      SNRC_RPC_TIMEOUT: ${SNRC_RPC_TIMEOUT:-}
+      SNRC_LOG_FORMAT: ${SNRC_LOG_FORMAT:-}
+      SNRC_LOG_COLOR: ${SNRC_LOG_COLOR:-}
+      SNRC_LOG_LEVEL: ${SNRC_LOG_LEVEL:-}
+      SNRC_TRUSTED_PROXIES: ${SNRC_TRUSTED_PROXIES:-}
     ports:
       - "127.0.0.1:8000:8000"
+    # a line per lookup; the local driver keeps 5 rotated files of 20 MB
+    logging:
+      driver: local
     restart: unless-stopped
 
 volumes:

scripts/resolver/service/.dockerignore

@@ -0,0 +1,5 @@
+.venv
+**/__pycache__
+.pytest_cache
+.ruff_cache
+test_snrc_resolve.py

scripts/resolver/service/.python-version

@@ -0,0 +1 @@
+3.13

scripts/resolver/service/Dockerfile

@@ -1,9 +1,10 @@
 # syntax=docker/dockerfile:1.7
 # ---------- builder ----------
-# Use the official uv image (Astral) on top of a slim Python base.
-# uv resolves and installs the lockfile-free pyproject.toml in seconds and
-# produces a portable .venv we can copy into the runtime stage.
-FROM ghcr.io/astral-sh/uv:python3.13-bookworm-slim AS builder
+# The official uv image (Astral) on top of a slim Python base installs the
+# locked dependencies into a portable .venv for the runtime stage.
+# Pinned to the uv that wrote uv.lock (its bookworm images stopped at 0.9.30), on the
+# runtime's Debian release, so the .venv's Python path exists there.
+FROM ghcr.io/astral-sh/uv:0.12.16-python3.13-trixie-slim AS builder
 
 ENV UV_LINK_MODE=copy \
     UV_COMPILE_BYTECODE=1 \
@@ -12,17 +13,17 @@ ENV UV_LINK_MODE=copy \
 
 WORKDIR /app
 
-# Install deps first (separate layer) — script edits won't bust this cache.
-COPY pyproject.toml ./
+# Dependencies first (separate layer), exactly as locked: script edits don't bust this cache.
+COPY pyproject.toml uv.lock .python-version ./
 RUN --mount=type=cache,target=/root/.cache/uv \
-    uv sync --no-dev --no-install-project
+    uv sync --frozen --no-dev --no-install-project
 
 # Script is added after the dep layer for cache friendliness.
 COPY snrc-resolve.py ./
 
 # ---------- runtime ----------
 # Slim runtime — only the venv + script. No uv, no apt.
-FROM python:3.13-slim AS runtime
+FROM python:3.13-slim-trixie AS runtime
 
 ENV PYTHONUNBUFFERED=1 \
     PYTHONDONTWRITEBYTECODE=1 \
@@ -33,15 +34,15 @@ RUN groupadd --system --gid 10001 snrc && \
     useradd  --system --uid 10001 --gid snrc --no-create-home --shell /usr/sbin/nologin snrc
 
 WORKDIR /app
-COPY --from=builder --chown=snrc:snrc /app /app
+# owned by root, so the service cannot rewrite its own code
+COPY --from=builder /app /app
 
 USER snrc:snrc
 
 EXPOSE 8000
 
-# Liveness check hits the script's own /health route. ThreadingHTTPServer is
-# fast enough that 3s is generous for a localhost probe; restart if it stops
-# responding entirely.
+# Liveness check hits the script's own /health route. Three failures mark the
+# container unhealthy; compose does not restart it for that, but `docker ps` shows it.
 HEALTHCHECK --interval=30s --timeout=5s --start-period=10s --retries=3 \
     CMD ["python", "-c", "import urllib.request, sys; sys.exit(0 if urllib.request.urlopen('http://127.0.0.1:8000/health', timeout=3).status == 200 else 1)"]
 

scripts/resolver/service/pyproject.toml

@@ -9,5 +9,14 @@ dependencies = [
   "eth-hash[pycryptodome]>=0.7",
 ]
 
+[dependency-groups]
+dev = [
+  "pytest>=8",
+  "ruff>=0.6",
+]
+
 [tool.uv]
 package = false
+
+[tool.ruff.lint]
+select = ["E9", "F", "B"]

scripts/resolver/service/uv.lock

@@ -0,0 +1,156 @@
+version = 1
+revision = 3
+requires-python = ">=3.11"
+
+[[package]]
+name = "colorama"
+version = "0.4.6"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/d8/53/6f443c9a4a8358a93a6792e2acffb9d9d5cb0a5cfd8802644b7b1c9a02e4/colorama-0.4.6.tar.gz", hash = "sha256:08695f5cb7ed6e0531a20572697297273c47b8cae5a63ffc6d6ed5c201be6e44", size = 27697, upload-time = "2022-10-25T02:36:22.414Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/d1/d6/3965ed04c63042e047cb6a3e6ed1a63a35087b6a609aa3a15ed8ac56c221/colorama-0.4.6-py2.py3-none-any.whl", hash = "sha256:4f1d9991f5acc0ca119f9d443620b77f9d6b33703e51011c16baf57afb285fc6", size = 25335, upload-time = "2022-10-25T02:36:20.889Z" },
+]
+
+[[package]]
+name = "eth-hash"
+version = "0.8.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/3c/f5/c67fc24f2f676aa9b7ab29679d44f113f314c817207cd4319353356f62da/eth_hash-0.8.0.tar.gz", hash = "sha256:b009752b620da2e9c7668014849d1f5fadbe4f138603f1871cc5d4ca706896b1", size = 12225, upload-time = "2026-03-25T16:36:55.099Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/87/87/b36792150ca0b28e4df683a34be15a61461ca0e349e5b5cf3ec8f694edb9/eth_hash-0.8.0-py3-none-any.whl", hash = "sha256:523718a51b369ab89866b929a5c93c52978cd866ea309192ad980dd8271f9fac", size = 7965, upload-time = "2026-03-25T16:36:54.205Z" },
+]
+
+[package.optional-dependencies]
+pycryptodome = [
+    { name = "pycryptodome" },
+]
+
+[[package]]
+name = "iniconfig"
+version = "2.3.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/72/34/14ca021ce8e5dfedc35312d08ba8bf51fdd999c576889fc2c24cb97f4f10/iniconfig-2.3.0.tar.gz", hash = "sha256:c76315c77db068650d49c5b56314774a7804df16fee4402c1f19d6d15d8c4730", size = 20503, upload-time = "2025-10-18T21:55:43.219Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/cb/b1/3846dd7f199d53cb17f49cba7e651e9ce294d8497c8c150530ed11865bb8/iniconfig-2.3.0-py3-none-any.whl", hash = "sha256:f631c04d2c48c52b84d0d0549c99ff3859c98df65b3101406327ecc7d53fbf12", size = 7484, upload-time = "2025-10-18T21:55:41.639Z" },
+]
+
+[[package]]
+name = "packaging"
+version = "26.3"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/7d/fa/3944b40b07da9ce895c0e6303a5ab7d53da063554f534556b134a54d6093/packaging-26.3.tar.gz", hash = "sha256:94edc256424af38762eb31306eed28beb9f0efc50a8837492c9d6fd6004aed79", size = 313412, upload-time = "2026-08-04T18:15:28.737Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/63/34/ba1c580383c9eada3711951fef0795c80b829a078d72188184bcab9dd527/packaging-26.3-py3-none-any.whl", hash = "sha256:d7193f7c8e4e93f444fde0262bf90af30e16fa0ad0ad44cb553c87339b23cd1c", size = 129956, upload-time = "2026-08-04T18:15:27.159Z" },
+]
+
+[[package]]
+name = "pluggy"
+version = "1.6.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/f9/e2/3e91f31a7d2b083fe6ef3fa267035b518369d9511ffab804f839851d2779/pluggy-1.6.0.tar.gz", hash = "sha256:7dcc130b76258d33b90f61b658791dede3486c3e6bfb003ee5c9bfb396dd22f3", size = 69412, upload-time = "2025-05-15T12:30:07.975Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/54/20/4d324d65cc6d9205fabedc306948156824eb9f0ee1633355a8f7ec5c66bf/pluggy-1.6.0-py3-none-any.whl", hash = "sha256:e920276dd6813095e9377c0bc5566d94c932c33b27a3e3945d8389c374dd4746", size = 20538, upload-time = "2025-05-15T12:30:06.134Z" },
+]
+
+[[package]]
+name = "pycryptodome"
+version = "3.23.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/8e/a6/8452177684d5e906854776276ddd34eca30d1b1e15aa1ee9cefc289a33f5/pycryptodome-3.23.0.tar.gz", hash = "sha256:447700a657182d60338bab09fdb27518f8856aecd80ae4c6bdddb67ff5da44ef", size = 4921276, upload-time = "2025-05-17T17:21:45.242Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/04/5d/bdb09489b63cd34a976cc9e2a8d938114f7a53a74d3dd4f125ffa49dce82/pycryptodome-3.23.0-cp313-cp313t-macosx_10_13_universal2.whl", hash = "sha256:0011f7f00cdb74879142011f95133274741778abba114ceca229adbf8e62c3e4", size = 2495152, upload-time = "2025-05-17T17:20:20.833Z" },
+    { url = "https://files.pythonhosted.org/packages/a7/ce/7840250ed4cc0039c433cd41715536f926d6e86ce84e904068eb3244b6a6/pycryptodome-3.23.0-cp313-cp313t-macosx_10_13_x86_64.whl", hash = "sha256:90460fc9e088ce095f9ee8356722d4f10f86e5be06e2354230a9880b9c549aae", size = 1639348, upload-time = "2025-05-17T17:20:23.171Z" },
+    { url = "https://files.pythonhosted.org/packages/ee/f0/991da24c55c1f688d6a3b5a11940567353f74590734ee4a64294834ae472/pycryptodome-3.23.0-cp313-cp313t-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:4764e64b269fc83b00f682c47443c2e6e85b18273712b98aa43bcb77f8570477", size = 2184033, upload-time = "2025-05-17T17:20:25.424Z" },
+    { url = "https://files.pythonhosted.org/packages/54/16/0e11882deddf00f68b68dd4e8e442ddc30641f31afeb2bc25588124ac8de/pycryptodome-3.23.0-cp313-cp313t-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:eb8f24adb74984aa0e5d07a2368ad95276cf38051fe2dc6605cbcf482e04f2a7", size = 2270142, upload-time = "2025-05-17T17:20:27.808Z" },
+    { url = "https://files.pythonhosted.org/packages/d5/fc/4347fea23a3f95ffb931f383ff28b3f7b1fe868739182cb76718c0da86a1/pycryptodome-3.23.0-cp313-cp313t-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:d97618c9c6684a97ef7637ba43bdf6663a2e2e77efe0f863cce97a76af396446", size = 2309384, upload-time = "2025-05-17T17:20:30.765Z" },
+    { url = "https://files.pythonhosted.org/packages/6e/d9/c5261780b69ce66d8cfab25d2797bd6e82ba0241804694cd48be41add5eb/pycryptodome-3.23.0-cp313-cp313t-musllinux_1_2_aarch64.whl", hash = "sha256:9a53a4fe5cb075075d515797d6ce2f56772ea7e6a1e5e4b96cf78a14bac3d265", size = 2183237, upload-time = "2025-05-17T17:20:33.736Z" },
+    { url = "https://files.pythonhosted.org/packages/5a/6f/3af2ffedd5cfa08c631f89452c6648c4d779e7772dfc388c77c920ca6bbf/pycryptodome-3.23.0-cp313-cp313t-musllinux_1_2_i686.whl", hash = "sha256:763d1d74f56f031788e5d307029caef067febf890cd1f8bf61183ae142f1a77b", size = 2343898, upload-time = "2025-05-17T17:20:36.086Z" },
+    { url = "https://files.pythonhosted.org/packages/9a/dc/9060d807039ee5de6e2f260f72f3d70ac213993a804f5e67e0a73a56dd2f/pycryptodome-3.23.0-cp313-cp313t-musllinux_1_2_x86_64.whl", hash = "sha256:954af0e2bd7cea83ce72243b14e4fb518b18f0c1649b576d114973e2073b273d", size = 2269197, upload-time = "2025-05-17T17:20:38.414Z" },
+    { url = "https://files.pythonhosted.org/packages/f9/34/e6c8ca177cb29dcc4967fef73f5de445912f93bd0343c9c33c8e5bf8cde8/pycryptodome-3.23.0-cp313-cp313t-win32.whl", hash = "sha256:257bb3572c63ad8ba40b89f6fc9d63a2a628e9f9708d31ee26560925ebe0210a", size = 1768600, upload-time = "2025-05-17T17:20:40.688Z" },
+    { url = "https://files.pythonhosted.org/packages/e4/1d/89756b8d7ff623ad0160f4539da571d1f594d21ee6d68be130a6eccb39a4/pycryptodome-3.23.0-cp313-cp313t-win_amd64.whl", hash = "sha256:6501790c5b62a29fcb227bd6b62012181d886a767ce9ed03b303d1f22eb5c625", size = 1799740, upload-time = "2025-05-17T17:20:42.413Z" },
+    { url = "https://files.pythonhosted.org/packages/5d/61/35a64f0feaea9fd07f0d91209e7be91726eb48c0f1bfc6720647194071e4/pycryptodome-3.23.0-cp313-cp313t-win_arm64.whl", hash = "sha256:9a77627a330ab23ca43b48b130e202582e91cc69619947840ea4d2d1be21eb39", size = 1703685, upload-time = "2025-05-17T17:20:44.388Z" },
+    { url = "https://files.pythonhosted.org/packages/db/6c/a1f71542c969912bb0e106f64f60a56cc1f0fabecf9396f45accbe63fa68/pycryptodome-3.23.0-cp37-abi3-macosx_10_9_universal2.whl", hash = "sha256:187058ab80b3281b1de11c2e6842a357a1f71b42cb1e15bce373f3d238135c27", size = 2495627, upload-time = "2025-05-17T17:20:47.139Z" },
+    { url = "https://files.pythonhosted.org/packages/6e/4e/a066527e079fc5002390c8acdd3aca431e6ea0a50ffd7201551175b47323/pycryptodome-3.23.0-cp37-abi3-macosx_10_9_x86_64.whl", hash = "sha256:cfb5cd445280c5b0a4e6187a7ce8de5a07b5f3f897f235caa11f1f435f182843", size = 1640362, upload-time = "2025-05-17T17:20:50.392Z" },
+    { url = "https://files.pythonhosted.org/packages/50/52/adaf4c8c100a8c49d2bd058e5b551f73dfd8cb89eb4911e25a0c469b6b4e/pycryptodome-3.23.0-cp37-abi3-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:67bd81fcbe34f43ad9422ee8fd4843c8e7198dd88dd3d40e6de42ee65fbe1490", size = 2182625, upload-time = "2025-05-17T17:20:52.866Z" },
+    { url = "https://files.pythonhosted.org/packages/5f/e9/a09476d436d0ff1402ac3867d933c61805ec2326c6ea557aeeac3825604e/pycryptodome-3.23.0-cp37-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:c8987bd3307a39bc03df5c8e0e3d8be0c4c3518b7f044b0f4c15d1aa78f52575", size = 2268954, upload-time = "2025-05-17T17:20:55.027Z" },
+    { url = "https://files.pythonhosted.org/packages/f9/c5/ffe6474e0c551d54cab931918127c46d70cab8f114e0c2b5a3c071c2f484/pycryptodome-3.23.0-cp37-abi3-manylinux_2_5_i686.manylinux1_i686.manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:aa0698f65e5b570426fc31b8162ed4603b0c2841cbb9088e2b01641e3065915b", size = 2308534, upload-time = "2025-05-17T17:20:57.279Z" },
+    { url = "https://files.pythonhosted.org/packages/18/28/e199677fc15ecf43010f2463fde4c1a53015d1fe95fb03bca2890836603a/pycryptodome-3.23.0-cp37-abi3-musllinux_1_2_aarch64.whl", hash = "sha256:53ecbafc2b55353edcebd64bf5da94a2a2cdf5090a6915bcca6eca6cc452585a", size = 2181853, upload-time = "2025-05-17T17:20:59.322Z" },
+    { url = "https://files.pythonhosted.org/packages/ce/ea/4fdb09f2165ce1365c9eaefef36625583371ee514db58dc9b65d3a255c4c/pycryptodome-3.23.0-cp37-abi3-musllinux_1_2_i686.whl", hash = "sha256:156df9667ad9f2ad26255926524e1c136d6664b741547deb0a86a9acf5ea631f", size = 2342465, upload-time = "2025-05-17T17:21:03.83Z" },
+    { url = "https://files.pythonhosted.org/packages/22/82/6edc3fc42fe9284aead511394bac167693fb2b0e0395b28b8bedaa07ef04/pycryptodome-3.23.0-cp37-abi3-musllinux_1_2_x86_64.whl", hash = "sha256:dea827b4d55ee390dc89b2afe5927d4308a8b538ae91d9c6f7a5090f397af1aa", size = 2267414, upload-time = "2025-05-17T17:21:06.72Z" },
+    { url = "https://files.pythonhosted.org/packages/59/fe/aae679b64363eb78326c7fdc9d06ec3de18bac68be4b612fc1fe8902693c/pycryptodome-3.23.0-cp37-abi3-win32.whl", hash = "sha256:507dbead45474b62b2bbe318eb1c4c8ee641077532067fec9c1aa82c31f84886", size = 1768484, upload-time = "2025-05-17T17:21:08.535Z" },
+    { url = "https://files.pythonhosted.org/packages/54/2f/e97a1b8294db0daaa87012c24a7bb714147c7ade7656973fd6c736b484ff/pycryptodome-3.23.0-cp37-abi3-win_amd64.whl", hash = "sha256:c75b52aacc6c0c260f204cbdd834f76edc9fb0d8e0da9fbf8352ef58202564e2", size = 1799636, upload-time = "2025-05-17T17:21:10.393Z" },
+    { url = "https://files.pythonhosted.org/packages/18/3d/f9441a0d798bf2b1e645adc3265e55706aead1255ccdad3856dbdcffec14/pycryptodome-3.23.0-cp37-abi3-win_arm64.whl", hash = "sha256:11eeeb6917903876f134b56ba11abe95c0b0fd5e3330def218083c7d98bbcb3c", size = 1703675, upload-time = "2025-05-17T17:21:13.146Z" },
+]
+
+[[package]]
+name = "pygments"
+version = "2.21.0"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/49/2e/ced460408999b33da6b31b0021b0f37d329e202d4169aeb164493778f25b/pygments-2.21.0.tar.gz", hash = "sha256:610ca751c9bc2492b38eb9a38a7fbc93edbbb2d7182edaf34e66ae493dee5c8c", size = 5005329, upload-time = "2026-08-17T08:02:48.824Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/71/46/17f022dd3e953bf20a04a028a21ec746d942f8d2af30fa0f124fa0e6a684/pygments-2.21.0-py3-none-any.whl", hash = "sha256:2363c69b61c4a97c838da3b130dcd6468f4848992b21a82f2a63ec34377137d9", size = 1250147, upload-time = "2026-08-17T08:02:44.912Z" },
+]
+
+[[package]]
+name = "pytest"
+version = "9.1.1"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "colorama", marker = "sys_platform == 'win32'" },
+    { name = "iniconfig" },
+    { name = "packaging" },
+    { name = "pluggy" },
+    { name = "pygments" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/e4/47/b9efed96c114afcfa3c9d3fe98a76a1d14c74a9e266d397cf6eb64be5e01/pytest-9.1.1.tar.gz", hash = "sha256:1088fbde8f2b49d95a549a195707afa7a76a3ce9bcadc26b6d71f0ffda5fe313", size = 1636369, upload-time = "2026-06-19T10:58:32.857Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/24/25/1de2678b631f5a49215c6c96fff41ba892b0a34df68d6d80292b1b48aa7f/pytest-9.1.1-py3-none-any.whl", hash = "sha256:37a86b45efb9a47a61a36449063e8e18d0cab3161329fc099eb21783169c4f0c", size = 386536, upload-time = "2026-06-19T10:58:31.347Z" },
+]
+
+[[package]]
+name = "ruff"
+version = "0.16.9"
+source = { registry = "https://pypi.org/simple" }
+sdist = { url = "https://files.pythonhosted.org/packages/96/bf/c935ca98e73fe8ce65b87ef08a280c0c1e85295d569228c15e87d8fdfaf1/ruff-0.16.9.tar.gz", hash = "sha256:12b625c6cfba78d285d9f48eda5f053374f1e53cb10ef17342a383750db99161", size = 4948764, upload-time = "2026-09-24T20:37:49.416Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/0d/26/df51322b52ee1ada7eff2d071ea09d11d5c2d1dcc9f02594f5785c4d1635/ruff-0.16.9-py3-none-linux_armv6l.whl", hash = "sha256:95e6f022090368ab3b824c36276839c53b2adf1a3f4c09fefc33dfc400f6da96", size = 10082922, upload-time = "2026-09-24T20:37:13.045Z" },
+    { url = "https://files.pythonhosted.org/packages/a5/27/7bf51f5a7aa375e9f339280a303aab44ca75dc1525f1cdc5991761685b0f/ruff-0.16.9-py3-none-macosx_10_12_x86_64.whl", hash = "sha256:a5f27be168556594a86d2f415db0cf43f5291917849318f873c7e2791f7a8c67", size = 10236360, upload-time = "2026-09-24T20:37:16.049Z" },
+    { url = "https://files.pythonhosted.org/packages/b6/63/09659283f92f02dff45809da194a70da2f688d87c55d8875c4fae3536072/ruff-0.16.9-py3-none-macosx_11_0_arm64.whl", hash = "sha256:1632eb1d6197f33bd00b1acbc5b71009e89a8895c158e2d2b03a834fac964ab6", size = 9892940, upload-time = "2026-09-24T20:37:17.957Z" },
+    { url = "https://files.pythonhosted.org/packages/24/58/98de1b72ec172f5f8f1731236fe21585b3998bf7dfe9fcc44ae9ba626012/ruff-0.16.9-py3-none-manylinux_2_17_aarch64.manylinux2014_aarch64.whl", hash = "sha256:b3f951b14d865d5952c89d40a5ca07e87abe24fa5453299878411e127748fb1c", size = 10032114, upload-time = "2026-09-24T20:37:19.942Z" },
+    { url = "https://files.pythonhosted.org/packages/c8/7d/f1e17c54ab59d4bad1dce8ee3e22a7a1d0ef4745240decacdcf3832b5bb2/ruff-0.16.9-py3-none-manylinux_2_17_armv7l.manylinux2014_armv7l.whl", hash = "sha256:447fc07e1573afff7cb02803462b12b6c8ece7cf10e2cd78565fa6d7a1c0bf8d", size = 9910227, upload-time = "2026-09-24T20:37:21.872Z" },
+    { url = "https://files.pythonhosted.org/packages/34/19/436f647a65075bbd3bab2668b3bdaa5120559b294694018cdcefabbbf30b/ruff-0.16.9-py3-none-manylinux_2_17_i686.manylinux2014_i686.whl", hash = "sha256:8a3e039a6a40ed976c491722b60e0ae4a4aa1a86057f540ee7a37a5d19ae9120", size = 10547484, upload-time = "2026-09-24T20:37:24.229Z" },
+    { url = "https://files.pythonhosted.org/packages/03/59/38430a6bc2f6d8095447ac39625cf8b6e9344a47e6c26225c8ba1bff3ffb/ruff-0.16.9-py3-none-manylinux_2_17_ppc64le.manylinux2014_ppc64le.whl", hash = "sha256:4684dded7db60aa57cb118fa158630f5feade4af5782903b6053484bdf9bd129", size = 11412367, upload-time = "2026-09-24T20:37:26.307Z" },
+    { url = "https://files.pythonhosted.org/packages/c8/bd/bbb6d7fc7f208c8b8c50dd5dc8206e4cfdb1e7a8fb852606a3adf370c880/ruff-0.16.9-py3-none-manylinux_2_17_s390x.manylinux2014_s390x.whl", hash = "sha256:d29c934357e45642fda2f34c0b1f4025b4a6c01e15e4bf0016879d60078a142c", size = 10869787, upload-time = "2026-09-24T20:37:28.35Z" },
+    { url = "https://files.pythonhosted.org/packages/bc/b8/9c543074918061abbefc3bd139bee22de35abedb00dde0f2d27288838962/ruff-0.16.9-py3-none-manylinux_2_17_x86_64.manylinux2014_x86_64.whl", hash = "sha256:a21713e629d3e5bdb2f5c2def1cc7f04f47fa8e1a7eb0571b4a28e1da64bc728", size = 10406494, upload-time = "2026-09-24T20:37:30.624Z" },
+    { url = "https://files.pythonhosted.org/packages/35/7a/5a8851bd146e7ccf8fd4b003f6c75c11f8fbdb0e60673b45097986b6bf41/ruff-0.16.9-py3-none-manylinux_2_31_riscv64.whl", hash = "sha256:7baa24ef5fc8e77aa93879e1d3f43754a01ae488e869f1ae30cf431afd4d2452", size = 10590083, upload-time = "2026-09-24T20:37:32.439Z" },
+    { url = "https://files.pythonhosted.org/packages/87/f0/4c3467188f23f806960b46fa76575a7cd0514c9ba90562650b71efc96980/ruff-0.16.9-py3-none-musllinux_1_2_aarch64.whl", hash = "sha256:a41aac6230aadfaa133bdfa1614488531ffa3e0837567ae04c0da2058a9c0f9e", size = 10119151, upload-time = "2026-09-24T20:37:34.581Z" },
+    { url = "https://files.pythonhosted.org/packages/15/34/5a4def5adea572ce6aea0bb64f21f928ee317b80e0db747d7979b01d7261/ruff-0.16.9-py3-none-musllinux_1_2_armv7l.whl", hash = "sha256:c2529fb5896d49115b0e9aa8f887490b34bbe76baf879ec2264ac59406869ce7", size = 9911544, upload-time = "2026-09-24T20:37:36.796Z" },
+    { url = "https://files.pythonhosted.org/packages/62/5d/d15ebea7499eef9373318c0ee6ca127832927c6529731f6d48e18dce7ca9/ruff-0.16.9-py3-none-musllinux_1_2_i686.whl", hash = "sha256:41e3870277694177429b56406d65dfbdb2c2802c52b715edaf6a0b829c69d4ee", size = 10269884, upload-time = "2026-09-24T20:37:38.857Z" },
+    { url = "https://files.pythonhosted.org/packages/d1/56/c5d3cd119ded7a3c7aba0e961b69cb3df701c91662c98ad694467d060ce1/ruff-0.16.9-py3-none-musllinux_1_2_x86_64.whl", hash = "sha256:8adbe4e58af167f767d7b2ba5e83c42e878350796cf78c2f5e14ab9903a92588", size = 10749366, upload-time = "2026-09-24T20:37:41.042Z" },
+    { url = "https://files.pythonhosted.org/packages/ac/fe/734ec7527029ac757ecf821f143c9f3fcf69149c21f53a044a899b430f5a/ruff-0.16.9-py3-none-win32.whl", hash = "sha256:0e1dbc2073624dee6618d41d0098690a7244654af746704b64759e12b6b6b385", size = 10152355, upload-time = "2026-09-24T20:37:43.025Z" },
+    { url = "https://files.pythonhosted.org/packages/14/21/26e4643629b3ebb44f0a06f9c9a53058d63d989415f63a9a3c28e2ee7f22/ruff-0.16.9-py3-none-win_amd64.whl", hash = "sha256:6bd40fec8cd4c8a3d4dd589bd8ad4e6320c13c29234159bfd959a40d529d597b", size = 10592965, upload-time = "2026-09-24T20:37:44.944Z" },
+    { url = "https://files.pythonhosted.org/packages/51/60/5fb1a39dbb5ae314d5f59bc7348a63c1d5c20f3cd83914c4b5cb0be31d2d/ruff-0.16.9-py3-none-win_arm64.whl", hash = "sha256:ed1a252039200f57a59eebc063b54beabea67bfbaaca0eeaa7f54b5fbcda2284", size = 10458649, upload-time = "2026-09-24T20:37:46.882Z" },
+]
+
+[[package]]
+name = "snrc-resolve"
+version = "0.1.0"
+source = { virtual = "." }
+dependencies = [
+    { name = "eth-hash", extra = ["pycryptodome"] },
+]
+
+[package.dev-dependencies]
+dev = [
+    { name = "pytest" },
+    { name = "ruff" },
+]
+
+[package.metadata]
+requires-dist = [{ name = "eth-hash", extras = ["pycryptodome"], specifier = ">=0.7" }]
+
+[package.metadata.requires-dev]
+dev = [
+    { name = "pytest", specifier = ">=8" },
+    { name = "ruff", specifier = ">=0.6" },
+]

simplexmq.cabal

@@ -1,7 +1,7 @@
 cabal-version: 3.0
 
 name:           simplexmq
-version:        7.1.0.5
+version:        7.1.0.11
 synopsis:       SimpleXMQ message broker
 description:    This package includes <./docs/Simplex-Messaging-Server.html server>,
                 <./docs/Simplex-Messaging-Client.html client> and
@@ -194,6 +194,9 @@ library
           Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260411_service_certs
           Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260712_address_dr_rpc
           Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260823_snd_files_entitlement
+          Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260919_ratchet_verify_codes
+          Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260929_ratchet_indexes
+          Simplex.Messaging.Agent.Store.Postgres.Migrations.M20261008_rcv_switch_status
   else
       exposed-modules:
           Simplex.Messaging.Agent.Store.SQLite
@@ -248,6 +251,10 @@ library
           Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260411_service_certs
           Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260712_address_dr_rpc
           Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260823_snd_files_entitlement
+          Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260919_ratchet_verify_codes
+          Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260929_ratchet_indexes
+          Simplex.Messaging.Agent.Store.SQLite.Migrations.M20261005_commands_autoincrement
+          Simplex.Messaging.Agent.Store.SQLite.Migrations.M20261008_rcv_switch_status
           Simplex.Messaging.Agent.Store.SQLite.Util
   if flag(client_postgres) || flag(server_postgres)
       exposed-modules:
@@ -595,7 +602,7 @@ test-suite simplexmq-test
   default-extensions:
       StrictData
   -- add -fhpc to ghc-options to run tests with coverage
-  ghc-options: -Weverything -Wno-missing-exported-signatures -Wno-missing-import-lists -Wno-missed-specialisations -Wno-all-missed-specialisations -Wno-unsafe -Wno-safe -Wno-missing-local-signatures -Wno-missing-kind-signatures -Wno-missing-deriving-strategies -Wno-monomorphism-restriction -Wno-prepositive-qualified-module -Wno-implicit-prelude -Wno-missing-safe-haskell-mode -Wno-missing-export-lists -Wno-partial-fields -Wcompat -Werror=incomplete-record-updates -Werror=incomplete-patterns -Werror=incomplete-uni-patterns -Werror=missing-methods -Werror=tabs -Wredundant-constraints -Wincomplete-record-updates -Wunused-type-patterns -O2 -threaded -rtsopts -with-rtsopts=-A64M -with-rtsopts=-N1
+  ghc-options: -Weverything -Wno-missing-exported-signatures -Wno-missing-import-lists -Wno-missed-specialisations -Wno-all-missed-specialisations -Wno-unsafe -Wno-safe -Wno-missing-local-signatures -Wno-missing-kind-signatures -Wno-missing-deriving-strategies -Wno-monomorphism-restriction -Wno-prepositive-qualified-module -Wno-implicit-prelude -Wno-missing-safe-haskell-mode -Wno-missing-export-lists -Wno-partial-fields -Wcompat -Werror=incomplete-record-updates -Werror=incomplete-patterns -Werror=incomplete-uni-patterns -Werror=missing-methods -Werror=tabs -Wredundant-constraints -Wincomplete-record-updates -Wunused-type-patterns -O2 -threaded -rtsopts "-with-rtsopts=-A64M -N1"
   build-depends:
       base
     , aeson
@@ -633,7 +640,6 @@ test-suite simplexmq-test
     , stm
     , text
     , time
-    , timeit ==2.0.*
     , tls >=1.9.0 && <1.10
     , transformers
     , unliftio

src/Simplex/FileTransfer/Agent.hs

@@ -17,11 +17,16 @@ module Simplex.FileTransfer.Agent
     toFSFilePath,
     -- Receiving files
     xftpReceiveFile',
+    xftpPrepareReceiveFile',
+    xftpStartReceiveFile',
     xftpDeleteRcvFile',
     xftpDeleteRcvFiles',
     -- Sending files
     xftpSendFile',
     xftpSendDescription',
+    xftpPrepareSendFile',
+    xftpPrepareSendDescription',
+    xftpStartSendFile',
     deleteSndFileInternal,
     deleteSndFilesInternal,
     deleteSndFileRemote,
@@ -127,7 +132,12 @@ closeXFTPAgent a = do
     stopWorkers workers = atomically (swapTVar workers M.empty) >>= mapM_ (liftIO . cancelWorker)
 
 xftpReceiveFile' :: AgentClient -> UserId -> ValidFileDescription 'FRecipient -> Maybe CryptoFileArgs -> Bool -> AM RcvFileId
-xftpReceiveFile' c userId (ValidFileDescription fd@FileDescription {chunks, redirect}) cfArgs approvedRelays = do
+xftpReceiveFile' c userId vfd cfArgs approvedRelays = do
+  fId <- xftpPrepareReceiveFile' c userId vfd cfArgs approvedRelays
+  fId <$ xftpStartReceiveFile' c fId
+
+xftpPrepareReceiveFile' :: AgentClient -> UserId -> ValidFileDescription 'FRecipient -> Maybe CryptoFileArgs -> Bool -> AM RcvFileId
+xftpPrepareReceiveFile' c userId (ValidFileDescription fd@FileDescription {redirect}) cfArgs approvedRelays = do
   g <- asks random
   prefixPath <- lift $ getPrefixPath "rcv.xftp"
   createDirectory prefixPath
@@ -137,7 +147,7 @@ xftpReceiveFile' c userId (ValidFileDescription fd@FileDescription {chunks, redi
   lift $ createDirectory =<< toFSFilePath relTmpPath
   lift $ createEmptyFile =<< toFSFilePath relSavePath
   let saveFile = CryptoFile relSavePath cfArgs
-  fId <- case redirect of
+  case redirect of
     Nothing -> withStore c $ \db -> createRcvFile db g userId fd relPrefixPath relTmpPath saveFile approvedRelays
     Just _ -> do
       -- prepare description paths
@@ -149,8 +159,11 @@ xftpReceiveFile' c userId (ValidFileDescription fd@FileDescription {chunks, redi
       let saveFileRedirect = CryptoFile relSavePathRedirect $ Just cfArgsRedirect
       -- create download tasks
       withStore c $ \db -> createRcvFileRedirect db g userId fd relPrefixPath relTmpPathRedirect saveFileRedirect relTmpPath saveFile approvedRelays
-  forM_ chunks (downloadChunk c)
-  pure fId
+
+xftpStartReceiveFile' :: AgentClient -> RcvFileId -> AM ()
+xftpStartReceiveFile' c rcvFileEntityId = do
+  srvs <- withStore c (`startPreparedRcvFile` rcvFileEntityId)
+  lift $ forM_ srvs $ void . getXFTPRcvWorker True c . Just
 
 downloadChunk :: AgentClient -> FileChunk -> AM ()
 downloadChunk c FileChunk {replicas = (FileChunkReplica {server} : _)} = do
@@ -353,6 +366,11 @@ notify c entId cmd = atomically $ writeTBQueue (subQ c) ("", entId, AEvt (sAEnti
 
 xftpSendFile' :: AgentClient -> UserId -> CryptoFile -> Int -> Maybe Word32 -> AM SndFileId
 xftpSendFile' c userId file numRecipients storageHours = do
+  fId <- xftpPrepareSendFile' c userId file numRecipients storageHours
+  fId <$ xftpStartSendFile' c fId
+
+xftpPrepareSendFile' :: AgentClient -> UserId -> CryptoFile -> Int -> Maybe Word32 -> AM SndFileId
+xftpPrepareSendFile' c userId file numRecipients storageHours = do
   g <- asks random
   prefixPath <- lift $ getPrefixPath "snd.xftp"
   createDirectory prefixPath
@@ -360,12 +378,15 @@ xftpSendFile' c userId file numRecipients storageHours = do
   key <- atomically $ C.randomSbKey g
   nonce <- atomically $ C.randomCbNonce g
   -- saving absolute filePath will not allow to restore file encryption after app update, but it's a short window
-  fId <- withStore c $ \db -> createSndFile db g userId file numRecipients relPrefixPath key nonce Nothing storageHours
-  lift . void $ getXFTPSndWorker True c Nothing
-  pure fId
+  withStore c $ \db -> createSndFile db g userId file numRecipients relPrefixPath key nonce Nothing storageHours
 
 xftpSendDescription' :: AgentClient -> UserId -> ValidFileDescription 'FRecipient -> Int -> AM SndFileId
-xftpSendDescription' c userId (ValidFileDescription fdDirect@FileDescription {size, digest}) numRecipients = do
+xftpSendDescription' c userId vfd numRecipients = do
+  fId <- xftpPrepareSendDescription' c userId vfd numRecipients
+  fId <$ xftpStartSendFile' c fId
+
+xftpPrepareSendDescription' :: AgentClient -> UserId -> ValidFileDescription 'FRecipient -> Int -> AM SndFileId
+xftpPrepareSendDescription' c userId (ValidFileDescription fdDirect@FileDescription {size, digest}) numRecipients = do
   g <- asks random
   prefixPath <- lift $ getPrefixPath "snd.xftp"
   createDirectory prefixPath
@@ -376,9 +397,12 @@ xftpSendDescription' c userId (ValidFileDescription fdDirect@FileDescription {si
   liftError (FILE . FILE_IO . show) $ CF.writeFile file (LB.fromStrict $ strEncode fdDirect)
   key <- atomically $ C.randomSbKey g
   nonce <- atomically $ C.randomCbNonce g
-  fId <- withStore c $ \db -> createSndFile db g userId file numRecipients relPrefixPath key nonce (Just RedirectFileInfo {size, digest}) Nothing
+  withStore c $ \db -> createSndFile db g userId file numRecipients relPrefixPath key nonce (Just RedirectFileInfo {size, digest}) Nothing
+
+xftpStartSendFile' :: AgentClient -> SndFileId -> AM ()
+xftpStartSendFile' c sndFileEntityId = do
+  withStore c (`startPreparedSndFile` sndFileEntityId)
   lift . void $ getXFTPSndWorker True c Nothing
-  pure fId
 
 resumeXFTPSndWork :: AgentClient -> Maybe XFTPServer -> AM' ()
 resumeXFTPSndWork = void .: getXFTPSndWorker False
@@ -544,9 +568,10 @@ runXFTPSndWorker c srv Worker {doWork} = do
       notify c sndFileEntityId $ SFPROG uploaded total
       when complete $ do
         (sndDescr, rcvDescrs) <- sndFileToDescrs sf
+        flip catchAllErrors (\e -> logError $ "XFTP snd worker error: " <> tshow e) $ do
+          lift . forM_ prefixPath $ removePath <=< toFSFilePath
+          withStore' c $ \db -> updateSndFileComplete db sndFileId
         notify c sndFileEntityId $ SFDONE sndDescr rcvDescrs (sndFileExpiresAt chunks)
-        lift . forM_ prefixPath $ removePath <=< toFSFilePath
-        withStore' c $ \db -> updateSndFileComplete db sndFileId
       where
         addRecipients :: SndFileChunk -> SndFileChunkReplica -> AM SndFileChunkReplica
         addRecipients ch@SndFileChunk {numRecipients} cr@SndFileChunkReplica {sndChunkReplicaId, rcvIdsKeys}

src/Simplex/FileTransfer/Client/Main.hs

@@ -441,7 +441,6 @@ cliReceiveFile ReceiveOptions {fileDescription, filePath, retryCount, tempPath,
           | depth > 0 -> do
               CryptoFile tmpFile _ <- withExceptT cliCryptoError $ decryptChunks encSize chunkPaths key nonce $ \_ ->
                 fmap CF.plain $ uniqueCombine encPath "redirect.yaml"
-              forM_ chunks $ acknowledgeFileChunk a
               yaml <- liftIO $ B.readFile tmpFile
               whenM (doesPathExist encPath) $ removeDirectoryRecursive encPath
               innerVfd <- either (throwE . CLIError . ("Redirect: invalid file description: " <>)) pure $ strDecode yaml
@@ -450,7 +449,6 @@ cliReceiveFile ReceiveOptions {fileDescription, filePath, retryCount, tempPath,
         Nothing -> do
           liftIO $ printNoNewLine "Decrypting file..."
           CryptoFile path _ <- withExceptT cliCryptoError $ decryptChunks encSize chunkPaths key nonce $ fmap CF.plain . getFilePath
-          forM_ chunks $ acknowledgeFileChunk a
           whenM (doesPathExist encPath) $ removeDirectoryRecursive encPath
           liftIO $ do
             printNoNewLine $ "File downloaded: " <> path
@@ -479,12 +477,6 @@ cliReceiveFile ReceiveOptions {fileDescription, filePath, retryCount, tempPath,
       _ -> (`uniqueCombine` name') . (</> "Downloads") =<< getHomeDirectory
       where
         name' = T.unpack name
-    acknowledgeFileChunk :: XFTPClientAgent -> FileChunk -> ExceptT CLIError IO ()
-    acknowledgeFileChunk a FileChunk {replicas = replica : _} = do
-      let FileChunkReplica {server, replicaId, replicaKey} = replica
-      c <- withRetry retryCount $ getXFTPServerClient a server
-      withRetry retryCount $ ackXFTPChunk c replicaKey (unChunkReplicaId replicaId)
-    acknowledgeFileChunk _ _ = throwE $ CLIError "chunk has no replicas"
 
 printProgress :: String -> Int64 -> Int64 -> IO ()
 printProgress s part total = printNoNewLine $ s <> " " <> show ((part * 100) `div` total) <> "%"

src/Simplex/FileTransfer/Description.hs

@@ -29,6 +29,7 @@ module Simplex.FileTransfer.Description
     YAMLFileDescription (..), -- for tests
     YAMLServerReplicas (..), -- for tests
     validateFileDescription,
+    sharedDescriptionHash,
     groupReplicasByServer,
     fdSeparator,
     kb,
@@ -72,6 +73,7 @@ import Simplex.FileTransfer.Protocol
 import Simplex.Messaging.Agent.QueryString
 import Simplex.Messaging.Agent.Store.DB (Binary (..), FromField (..), ToField (..))
 import qualified Simplex.Messaging.Crypto as C
+import Simplex.Messaging.Encoding (Encoding (..))
 import Simplex.Messaging.Encoding.String
 import Simplex.Messaging.Parsers (defaultJSON, parseAll)
 import Simplex.Messaging.Protocol (XFTPServer)
@@ -227,6 +229,13 @@ validateFileDescription fd@FileDescription {size, chunks}
     chunkNos = map (\FileChunk {chunkNo} -> chunkNo) chunks
     chunksSize = foldl' (\(s :: Int64) FileChunk {chunkSize} -> s + fromIntegral (unFileSize chunkSize)) 0
 
+sharedDescriptionHash :: FileDescription p -> ByteString
+sharedDescriptionHash FileDescription {size, digest, key, nonce, chunkSize, chunks} =
+  C.sha512Hash $ smpEncode (unFileSize size, unFileDigest digest, C.unSbKey key, nonce, unFileSize chunkSize) <> foldMap chunkFields chunks
+  where
+    chunkFields FileChunk {chunkNo, chunkSize = cSize, digest = cDigest} =
+      smpEncode (fromIntegral chunkNo :: Int64, unFileSize cSize, unFileDigest cDigest)
+
 encodeFileDescription :: FileDescription p -> YAMLFileDescription
 encodeFileDescription FileDescription {party, size, digest, key, nonce, chunkSize, chunks, redirect} =
   YAMLFileDescription

src/Simplex/FileTransfer/Protocol.hs

@@ -80,7 +80,7 @@ import Simplex.Messaging.Protocol
     tParse,
   )
 import Simplex.Messaging.Transport (THandleParams (..), TransportError (..), TransportPeer (..))
-import Simplex.Messaging.Util ((<$?>))
+import Simplex.Messaging.Util (eitherToMaybe, (<$?>))
 
 xftpBlockSize :: Int
 xftpBlockSize = 16384
@@ -383,7 +383,7 @@ checkParty' c = case testEquality (sFileParty @p) (sFileParty @p') of
 xftpEncodeAuthTransmission :: ProtocolEncoding XFTPVersion XFTPErrorType c => THandleParams XFTPVersion 'TClient -> C.APrivateAuthKey -> Transmission c -> Either TransportError ByteString
 xftpEncodeAuthTransmission thParams@THandleParams {thAuth} pKey t@(corrId, _, _) = do
   let TransmissionForAuth {tForAuth, tToSend} = encodeTransmissionForAuth thParams t
-  xftpEncodeBatch1 . (,tToSend) =<< authTransmission thAuth False (Just pKey) (C.cbNonce $ bs corrId) tForAuth
+  xftpEncodeBatch1 . (,tToSend) =<< authTransmission thAuth False (Just pKey) (eitherToMaybe $ C.cbNonce $ bs corrId) tForAuth
 
 xftpEncodeTransmission :: ProtocolEncoding XFTPVersion XFTPErrorType c => THandleParams XFTPVersion p -> Transmission c -> Either TransportError ByteString
 xftpEncodeTransmission thParams t = xftpEncodeBatch1 (Nothing, encodeTransmission thParams t)

src/Simplex/FileTransfer/Server.hs

@@ -25,7 +25,6 @@ import Control.Monad.Reader
 import Control.Monad.Trans.Except
 import qualified Data.Aeson as J
 import Data.Bifunctor (first)
-import qualified Data.ByteString.Base64.URL as B64
 import Data.ByteString.Builder (Builder, byteString)
 import Data.ByteString.Char8 (ByteString)
 import qualified Data.ByteString.Char8 as B
@@ -63,7 +62,7 @@ import Simplex.Messaging.Crypto.Entitlement (Entitlement (..), EntitlementProof
 import qualified Simplex.Messaging.Crypto.Lazy as LC
 import Simplex.Messaging.Encoding
 import Simplex.Messaging.Encoding.String
-import Simplex.Messaging.Protocol (BlockingInfo, EntityId (..), RcvPublicAuthKey, RcvPublicDhKey, RecipientId, SignedTransmission, pattern NoEntity)
+import Simplex.Messaging.Protocol (BlockingInfo, CommandError (..), EntityId (..), RcvPublicAuthKey, RcvPublicDhKey, RecipientId, SignedTransmission, pattern NoEntity)
 import Simplex.Messaging.Server (controlPortAuth, dummyVerifyCmd, verifyCmdAuthorization)
 import Simplex.Messaging.Server.Control (CPClientRole (..))
 import Simplex.Messaging.Server.Expiration
@@ -594,7 +593,7 @@ processXFTPRequest ent HTTP2Body {bodyPart} = \case
               \used -> let used' = used + fromIntegral size in if used' <= quota then (True, used') else (False, used)
           receive = do
             path <- asks $ filesPath . config
-            let fPath = path </> B.unpack (B64.encode $ unEntityId senderId)
+            let fPath = path </> serverFileName senderId
             receiveChunk (XFTPRcvChunkSpec fPath size digest) >>= \case
               Right () -> do
                 stats <- asks serverStats
@@ -646,11 +645,12 @@ processXFTPRequest ent HTTP2Body {bodyPart} = \case
 
     ackFileReception :: RecipientId -> FileRec -> M s FileResponse
     ackFileReception rId fr = do
-      withFileLog (`logAckFile` rId)
-      st <- asks fileStore
-      liftIO $ deleteRecipient st rId fr
-      incFileStat fileDownloadAcks
-      pure FROk
+      pure $ FRErr $ CMD PROHIBITED
+      -- withFileLog (`logAckFile` rId)
+      -- st <- asks fileStore
+      -- liftIO $ deleteRecipient st rId fr
+      -- incFileStat fileDownloadAcks
+      -- pure FROk
 
 deleteServerFile_ :: FileStoreClass s => FileRec -> M s (Either XFTPErrorType ())
 deleteServerFile_ fr@FileRec {senderId} = do

src/Simplex/FileTransfer/Server/StoreLog.hs

@@ -17,6 +17,7 @@ module Simplex.FileTransfer.Server.StoreLog
     logDeleteFile,
     logBlockFile,
     logAckFile,
+    serverFileName,
   )
 where
 
@@ -24,6 +25,7 @@ import Control.Applicative (optional, (<|>))
 import Control.Concurrent.STM
 import Control.Monad.Except
 import qualified Data.Attoparsec.ByteString.Char8 as A
+import qualified Data.ByteString.Base64.URL as B64
 import qualified Data.ByteString.Char8 as B
 import qualified Data.ByteString.Lazy.Char8 as LB
 import Data.Composition ((.:), (.::.))
@@ -35,10 +37,11 @@ import Simplex.FileTransfer.Protocol (FileInfo (..))
 import Simplex.FileTransfer.Server.Store
 import Simplex.FileTransfer.Transport (XFTPErrorType (..))
 import Simplex.Messaging.Encoding.String
-import Simplex.Messaging.Protocol (BlockingInfo, RcvPublicAuthKey, RecipientId, SenderId)
+import Simplex.Messaging.Protocol (BlockingInfo, EntityId (..), RcvPublicAuthKey, RecipientId, SenderId)
 import Simplex.Messaging.Server.QueueStore (ServerEntityStatus (..))
 import Simplex.Messaging.Server.StoreLog
 import Simplex.Messaging.Util (bshow)
+import System.FilePath (takeFileName)
 import System.IO
 
 data FileStoreLogRecord
@@ -99,6 +102,9 @@ logBlockFile s fId = logFileStoreRecord s . BlockFile fId
 logAckFile :: StoreLog 'WriteMode -> RecipientId -> IO ()
 logAckFile s = logFileStoreRecord s . AckFile
 
+serverFileName :: SenderId -> FilePath
+serverFileName = B.unpack . B64.encode . unEntityId
+
 readWriteFileStore :: FilePath -> STMFileStore -> IO (StoreLog 'WriteMode)
 readWriteFileStore = readWriteStoreLog readFileStore writeFileStore
 
@@ -115,7 +121,9 @@ readFileStore f st = mapM_ (addFileLogRecord . LB.toStrict) . LB.lines =<< LB.re
       AddFile sId file createdAt expiresAt status
         | size file > 0 -> addFile st sId file createdAt expiresAt status
         | otherwise -> pure $ Left SIZE
-      PutFile qId path -> setFilePath st qId path
+      PutFile qId path
+        | takeFileName path == serverFileName qId -> setFilePath st qId path
+        | otherwise -> pure $ Left AUTH
       AddRecipients sId rcps -> runExceptT $ addRecipients sId rcps
       DeleteFile sId -> deleteFile st sId
       BlockFile sId info -> blockFile st sId info True

src/Simplex/FileTransfer/Types.hs

@@ -93,7 +93,8 @@ data RcvFile = RcvFile
   deriving (Show)
 
 data RcvFileStatus
-  = RFSReceiving
+  = RFSPrepared
+  | RFSReceiving
   | RFSReceived
   | RFSDecrypting
   | RFSComplete
@@ -106,6 +107,7 @@ instance ToField RcvFileStatus where toField = toField . textEncode
 
 instance TextEncoding RcvFileStatus where
   textDecode = \case
+    "prepared" -> Just RFSPrepared
     "receiving" -> Just RFSReceiving
     "received" -> Just RFSReceived
     "decrypting" -> Just RFSDecrypting
@@ -113,6 +115,7 @@ instance TextEncoding RcvFileStatus where
     "error" -> Just RFSError
     _ -> Nothing
   textEncode = \case
+    RFSPrepared -> "prepared"
     RFSReceiving -> "receiving"
     RFSReceived -> "received"
     RFSDecrypting -> "decrypting"
@@ -177,7 +180,8 @@ sndFileEncPath :: FilePath -> FilePath
 sndFileEncPath prefixPath = prefixPath </> "xftp.encrypted"
 
 data SndFileStatus
-  = SFSNew -- db record created
+  = SFSPrepared
+  | SFSNew -- db record created
   | SFSEncrypting -- encryption started
   | SFSEncrypted -- encryption complete
   | SFSUploading -- all chunk replicas are created on servers
@@ -191,6 +195,7 @@ instance ToField SndFileStatus where toField = toField . textEncode
 
 instance TextEncoding SndFileStatus where
   textDecode = \case
+    "prepared" -> Just SFSPrepared
     "new" -> Just SFSNew
     "encrypting" -> Just SFSEncrypting
     "encrypted" -> Just SFSEncrypted
@@ -199,6 +204,7 @@ instance TextEncoding SndFileStatus where
     "error" -> Just SFSError
     _ -> Nothing
   textEncode = \case
+    SFSPrepared -> "prepared"
     SFSNew -> "new"
     SFSEncrypting -> "encrypting"
     SFSEncrypted -> "encrypted"

src/Simplex/Messaging/Agent/Client.hs

@@ -272,7 +272,7 @@ import Simplex.Messaging.Protocol
     NetworkError (..),
     MsgFlags (..),
     MsgId,
-    NameRecord,
+    NameResponse,
     NtfServer,
     NtfServerWithAuth,
     ProtoServer,
@@ -312,7 +312,7 @@ import Simplex.Messaging.Session
 import Simplex.Messaging.SystemTime
 import Simplex.Messaging.TMap (TMap)
 import qualified Simplex.Messaging.TMap as TM
-import Simplex.Messaging.Transport (HandshakeError (..), SMPServiceRole (..), SMPVersion, ServiceCredentials (..), SessionId, THClientService' (..), THandleAuth (..), THandleParams (sessionId, thAuth, thVersion, serverInfo), TransportError (..), TransportPeer (..), shortLinksSMPVersion, newNtfCredsSMPVersion)
+import Simplex.Messaging.Transport (HandshakeError (..), SMPServiceRole (..), SMPVersion, ServiceCredentials (..), SessionId, THClientService' (..), THandleAuth (..), THandleParams (sessionId, thAuth, thVersion, serverInfo), TransportError (..), newNtfCredsSMPVersion)
 import Simplex.Messaging.Transport.Client (TransportHost (..))
 import Simplex.Messaging.Transport.Credentials
 import Simplex.Messaging.Util
@@ -1261,9 +1261,12 @@ sendOrProxySMPCommand c nm userId destSrv@ProtocolServer {host = destHosts} conn
           Left e -> throwE e
 
 ipAddressProtected :: NetworkConfig -> ProtocolServer p -> Bool
-ipAddressProtected NetworkConfig {socksProxy, hostMode} (ProtocolServer _ hosts _ _) = do
-  isJust socksProxy || (hostMode == HMOnion && any isOnionHost hosts)
+ipAddressProtected NetworkConfig {socksProxy, socksMode, hostMode} (ProtocolServer _ hosts _ _)
+  | isJust socksProxy = socksMode == SMAlways || if hostMode == HMPublic then allOnion else anyOnion
+  | otherwise = hostMode == HMOnion && anyOnion
   where
+    anyOnion = any isOnionHost hosts
+    allOnion = all isOnionHost hosts
     isOnionHost = \case THOnionHost _ -> True; _ -> False
 
 withNtfClient :: AgentClient -> NetworkRequestMode -> NtfServer -> EntityId -> ByteString -> (NtfClient -> ExceptT NtfClientError IO a) -> AM a
@@ -1496,7 +1499,7 @@ newRcvQueue_ c nm userId connId (ProtoServerWithAuth srv auth) vRange cqrd enabl
   let sessServiceId = (\THClientService {serviceId = sId} -> sId) <$> (clientService =<< thAuth thParams')
   when (isJust serviceId && serviceId /= sessServiceId) $ logError "incorrect service ID in NEW response"
   liftIO . logServer "<--" c srv NoEntity $ B.unwords ["IDS", logSecret rcvId, logSecret sndId]
-  shortLink <- mkShortLinkCreds thParams' qik
+  shortLink <- mkShortLinkCreds qik
   let rq =
         RcvQueue
           { userId,
@@ -1537,8 +1540,8 @@ newRcvQueue_ c nm userId connId (ProtoServerWithAuth srv auth) vRange cqrd enabl
       (Just ((ntfPublicKey, ntfPrivateKey), dhpk), Just (ServerNtfCreds notifierId dhk')) ->
         Just ClientNtfCreds {ntfPublicKey, ntfPrivateKey, notifierId, rcvNtfDhSecret = C.dh' dhk' dhpk}
       _ -> Nothing
-    mkShortLinkCreds :: THandleParams SMPVersion 'TClient -> QueueIdsKeys -> AM (Maybe ShortLinkCreds)
-    mkShortLinkCreds thParams' QIK {sndId, queueMode, linkId} = case (cqrd, queueMode) of
+    mkShortLinkCreds :: QueueIdsKeys -> AM (Maybe ShortLinkCreds)
+    mkShortLinkCreds QIK {sndId, queueMode, linkId} = case (cqrd, queueMode) of
       (CQRMessaging ld, Just QMMessaging) ->
         withLinkData ld $ \lnkId CQRData {linkKey, privSigKey, srvReq = (sndId', d)} ->
           if sndId == sndId'
@@ -1552,11 +1555,9 @@ newRcvQueue_ c nm userId connId (ProtoServerWithAuth srv auth) vRange cqrd enabl
       (_, Nothing) -> newErr "unexpected link ID"
       _ -> newErr "unexpected queue mode"
       where
-        v = thVersion thParams'
         withLinkData :: Maybe d -> (SMP.LinkId -> d -> AM (Maybe ShortLinkCreds)) -> AM (Maybe ShortLinkCreds)
         withLinkData ld_ mkLink = case (ld_, linkId) of
           (Just ld, Just lnkId) -> mkLink lnkId ld
-          (Just _, Nothing) | v < shortLinksSMPVersion -> pure Nothing
           (Nothing, Nothing) -> pure Nothing
           _ -> newErr "unexpected or absent link ID"
         newErr :: String -> AM (Maybe ShortLinkCreds)
@@ -1981,7 +1982,7 @@ decryptSMPMessage :: RcvQueue -> SMP.RcvMessage -> AM SMP.ClientRcvMsgBody
 decryptSMPMessage rq SMP.RcvMessage {msgId, msgBody = SMP.EncRcvMsgBody body} =
   liftEither $ parse SMP.clientRcvMsgBodyP (AGENT $ A_MESSAGE "decrypt message") =<< decrypt body
   where
-    decrypt = agentCbDecrypt (rcvDhSecret rq) (C.cbNonce msgId)
+    decrypt = agentCbDecrypt (rcvDhSecret rq) msgId
 
 secureQueue :: AgentClient -> NetworkRequestMode -> RcvQueue -> SndPublicAuthKey -> AM ()
 secureQueue c nm rq@RcvQueue {rcvId, rcvPrivateKey} senderKey =
@@ -2022,7 +2023,7 @@ getQueueLink c nm userId server lnkId =
 -- resolver) and falls back to a direct send when the proxy is unavailable
 -- (faster but exposes the client IP). Mode selection is delegated to
 -- `sendOrProxySMPCommand`, which honours the network config (SPMNever etc.).
-resolveName :: AgentClient -> NetworkRequestMode -> UserId -> SMPServer -> SimplexDomain -> AM NameRecord
+resolveName :: AgentClient -> NetworkRequestMode -> UserId -> SMPServer -> SimplexDomain -> AM NameResponse
 resolveName c nm userId server domain =
   snd <$> sendOrProxySMPCommand c nm userId server "" "RSLV" NoEntity resolveViaProxy resolveDirectly
   where
@@ -2079,9 +2080,9 @@ disableQueuesNtfs c = sendTSessionBatches "NDEL" (mkSMPTSession . snd) disableQu
     queueCreds :: DisableQueueNtfReq -> (SMP.RecipientId, SMP.RcvPrivateAuthKey)
     queueCreds (_, RcvQueue {rcvPrivateKey, rcvId}) = (rcvId, rcvPrivateKey)
 
-sendAck :: AgentClient -> RcvQueue -> MsgId -> AM ()
+sendAck :: AgentClient -> RcvQueue -> MsgId -> AM QueueDrained
 sendAck c rq@RcvQueue {rcvId, rcvPrivateKey} msgId =
-  withSMPClient c NRMBackground rq ("ACK:" <> logSecret' msgId) $ \smp ->
+  withSMPClient c NRMBackground rq ("ACK:" <> logSecret' (SMP.unMsgId msgId)) $ \smp ->
     ackSMPMessage smp rcvPrivateKey rcvId msgId
 
 hasGetLock :: SomeRcvQueue q => AgentClient -> q -> IO Bool
@@ -2276,7 +2277,7 @@ agentCbEncryptOnce clientVersion dhRcvPubKey msg = do
 
 -- | NaCl crypto-box decrypt - both for messages received from the server
 -- and per-queue E2E encrypted messages from the sender that were inside.
-agentCbDecrypt :: C.DhSecretX25519 -> C.CbNonce -> ByteString -> Either AgentErrorType ByteString
+agentCbDecrypt :: C.CbNonceI n => C.DhSecretX25519 -> n -> ByteString -> Either AgentErrorType ByteString
 agentCbDecrypt dhSecret nonce msg =
   first cryptoError $
     C.cbDecrypt dhSecret nonce msg

src/Simplex/Messaging/Agent/Env/SQLite.hs

@@ -162,6 +162,7 @@ data AgentConfig = AgentConfig
     connDeleteDeliveryTimeout :: NominalDiffTime,
     helloTimeout :: NominalDiffTime,
     quotaExceededTimeout :: NominalDiffTime,
+    commandQuotaRetryInterval :: RetryInterval,
     persistErrorInterval :: NominalDiffTime,
     initialCleanupDelay :: Int64,
     cleanupInterval :: Int64,
@@ -241,6 +242,7 @@ defaultAgentConfig =
       connDeleteDeliveryTimeout = 2 * nominalDay,
       helloTimeout = 2 * nominalDay,
       quotaExceededTimeout = 7 * nominalDay,
+      commandQuotaRetryInterval = RetryInterval {initialInterval = 30_000000, increaseAfter = 0, maxInterval = 3600_000000},
       persistErrorInterval = 3, -- seconds
       initialCleanupDelay = 30 * 1000000, -- 30 seconds
       cleanupInterval = 5 * 60 * 1000000, -- 5 minutes

src/Simplex/Messaging/Agent/Protocol.hs

@@ -79,6 +79,7 @@ module Simplex.Messaging.Agent.Protocol
     SMPConfirmation (..),
     AgentMsgEnvelope (..),
     AgentMessage (..),
+    RatchetInfo (..),
     RequestSignature (..),
     AgentMessageType (..),
     APrivHeader (..),
@@ -166,6 +167,8 @@ module Simplex.Messaging.Agent.Protocol
     ConnId,
     ConfirmationId,
     InvitationId,
+    ConnVerifyCodes (..),
+    ContactRequestBinding (..),
     MsgIntegrity (..),
     MsgErrorType (..),
     QueueStatus (..),
@@ -243,7 +246,6 @@ import Simplex.Messaging.Crypto.Ratchet
     RcvE2ERatchetParams,
     RcvE2ERatchetParamsUri,
     SndE2ERatchetParams,
-    RcvE2EPrivRatchetParams,
     pattern PQSupportOff,
     pattern PQSupportOn,
   )
@@ -404,7 +406,7 @@ data AEvent (e :: AEntity) where
   LINK :: ConnShortLink 'CMContact -> UserConnLinkData 'CMContact -> AEvent AEConn
   LDATA :: FixedLinkData 'CMContact -> ConnLinkData 'CMContact -> ConnectionRequestUri 'CMContact -> AEvent AEConn
   CONF :: ConfirmationId -> PQSupport -> [SMPServer] -> ConnInfo -> AEvent AEConn -- ConnInfo is from sender, [SMPServer] will be empty only in v1 handshake
-  REQ :: InvitationId -> PQSupport -> NonEmpty SMPServer -> ConnInfo -> Bool -> AEvent AEConn -- ConnInfo is from sender; Bool - rejection reason can be sent
+  REQ :: InvitationId -> PQSupport -> NonEmpty SMPServer -> ConnInfo -> ContactRequestBinding -> Bool -> AEvent AEConn -- ConnInfo is from sender; Bool - rejection reason can be sent
   SREQ :: InvitationId -> Maybe C.PublicKeyEd25519 -> MsgBody -> AEvent AEConn
   SSENT :: AgentMsgId -> Maybe SMPServer -> AEvent AEConn
   RJCT :: ConnInfo -> AEvent AEConn
@@ -609,6 +611,7 @@ data RcvSwitchStatus
   | RSSendingQADD
   | RSSendingQUSE
   | RSReceivedMessage
+  | RSReceivedQEND
   deriving (Eq, Show)
 
 instance StrEncoding RcvSwitchStatus where
@@ -617,12 +620,14 @@ instance StrEncoding RcvSwitchStatus where
     RSSendingQADD -> "sending_qadd"
     RSSendingQUSE -> "sending_quse"
     RSReceivedMessage -> "received_message"
+    RSReceivedQEND -> "received_qend"
   strP =
     A.takeTill (== ' ') >>= \case
       "switch_started" -> pure RSSwitchStarted
       "sending_qadd" -> pure RSSendingQADD
       "sending_quse" -> pure RSSendingQUSE
       "received_message" -> pure RSReceivedMessage
+      "received_qend" -> pure RSReceivedQEND
       _ -> fail "bad RcvSwitchStatus"
 
 instance ToField RcvSwitchStatus where toField = toField . decodeLatin1 . strEncode
@@ -926,7 +931,7 @@ data AgentMessage
   | -- AgentConnInfoReply is used by accepting party in duplexHandshake mode (v2), allowing to include reply queue(s) in the initial confirmation.
     -- It made removed REPLY message unnecessary.
     AgentConnInfoReply (NonEmpty SMPQueueInfo) ConnInfo
-  | AgentRatchetInfo ByteString
+  | AgentRatchetInfo RatchetInfo
   | AgentMessage APrivHeader AMessage
   | AgentServiceRequest (NonEmpty SMPQueueInfo) (Maybe RequestSignature) MsgBody
   | AgentServiceResponse MsgBody
@@ -937,7 +942,7 @@ instance Encoding AgentMessage where
   smpEncode = \case
     AgentConnInfo cInfo -> smpEncode ('I', Tail cInfo)
     AgentConnInfoReply smpQueues cInfo -> smpEncode ('D', smpQueues, Tail cInfo) -- 'D' stands for "duplex"
-    AgentRatchetInfo info -> smpEncode ('R', Tail info)
+    AgentRatchetInfo info -> smpEncode ('R', info)
     AgentMessage hdr aMsg -> smpEncode ('M', hdr, aMsg)
     AgentServiceRequest qs sig_ body -> smpEncode ('A', qs, sig_, Tail body)
     AgentServiceResponse body -> smpEncode ('P', Tail body)
@@ -946,13 +951,23 @@ instance Encoding AgentMessage where
     smpP >>= \case
       'I' -> AgentConnInfo . unTail <$> smpP
       'D' -> AgentConnInfoReply <$> smpP <*> (unTail <$> smpP)
-      'R' -> AgentRatchetInfo . unTail <$> smpP
+      'R' -> AgentRatchetInfo <$> smpP
       'M' -> AgentMessage <$> smpP <*> smpP
       'A' -> AgentServiceRequest <$> smpP <*> smpP <*> (unTail <$> smpP)
       'P' -> AgentServiceResponse . unTail <$> smpP
       'J' -> AgentRejection . unTail <$> smpP
       _ -> fail "bad AgentMessage"
 
+data RatchetInfo = RatchetInfo {answeredKeyHash :: Maybe ByteString}
+  deriving (Eq, Show)
+
+instance Encoding RatchetInfo where
+  smpEncode RatchetInfo {answeredKeyHash} = smpEncode answeredKeyHash
+  smpP = do
+    answeredKeyHash <- smpP <|> pure Nothing
+    _ <- A.takeByteString
+    pure RatchetInfo {answeredKeyHash}
+
 -- internal type for storing message type in the database
 data AgentMessageType
   = AM_CONN_INFO
@@ -1372,6 +1387,15 @@ type ConfirmationId = ByteString
 
 type InvitationId = ByteString
 
+data ConnVerifyCodes = ConnVerifyCodes
+  { codeAD :: ByteString,
+    codePQ :: Maybe ByteString
+  }
+  deriving (Eq, Show)
+
+data ContactRequestBinding = CRBRatchet ConnVerifyCodes | CRBRequest ByteString
+  deriving (Eq, Show)
+
 extraSMPServerHosts :: Map TransportHost TransportHost
 extraSMPServerHosts =
   M.fromList
@@ -1602,7 +1626,7 @@ newtype LinkKey = LinkKey ByteString -- sha3-256(fixed_data)
 instance ToField LinkKey where toField (LinkKey s) = toField $ Binary s
 
 -- | Parameters for creating a connection with a prepared link.
-data PreparedLinkParams = PreparedLinkParams
+data PreparedLinkParams k = PreparedLinkParams
   { -- | Correlation ID / determines sender ID
     plpNonce :: C.CbNonce,
     -- | Queue E2EE DH key pair
@@ -1617,8 +1641,8 @@ data PreparedLinkParams = PreparedLinkParams
     plpSrvWithAuth :: SMPServerWithAuth,
     -- | Initial PQ keys
     plpInitKeys :: InitialKeys,
-    -- | Contact address double ratchet keys
-    plpAddressKeys :: Maybe (RatchetKeyId, RcvE2EPrivRatchetParams 'C.X448)
+    -- | Contact address double ratchet keys or invitation X3DH keys
+    plpRatchetKeys :: k
   }
 
 instance ConnectionModeI c => ToField (ConnectionLink c) where toField = toField . Binary . strEncode

src/Simplex/Messaging/Agent/RetryInterval.hs

@@ -10,6 +10,7 @@ module Simplex.Messaging.Agent.RetryInterval
     withRetryInterval,
     withRetryIntervalCount,
     withRetryForeground,
+    withRetryInterval2,
     withRetryLock2,
     updateRetryInterval2,
     nextRetryDelay,
@@ -86,9 +87,24 @@ withRetryForeground ri isForeground isOnline action = callAction 0 $ initialInte
                 | otherwise = (elapsed + delay, nextRetryDelay elapsed' delay ri)
           callAction elapsed' delay'
 
+withRetryInterval2 :: forall m. MonadIO m => RetryInterval2 -> (RI2State -> (RetryIntervalMode -> m ()) -> m ()) -> m ()
+withRetryInterval2 = withRetryWait2 $ liftIO . threadDelay'
+
 -- This function allows action to toggle between slow and fast retry intervals.
 withRetryLock2 :: forall m. MonadIO m => RetryInterval2 -> TMVar () -> (RI2State -> (RetryIntervalMode -> m ()) -> m ()) -> m ()
-withRetryLock2 RetryInterval2 {riSlow, riFast} lock action =
+withRetryLock2 ri lock = withRetryWait2 wait ri
+  where
+    wait delay = do
+      waiting <- newTVarIO True
+      _ <- liftIO . forkIO $ do
+        threadDelay' delay
+        atomically $ whenM (readTVar waiting) $ void $ tryPutTMVar lock ()
+      atomically $ do
+        takeTMVar lock
+        writeTVar waiting False
+
+withRetryWait2 :: forall m. Monad m => (Int64 -> m ()) -> RetryInterval2 -> (RI2State -> (RetryIntervalMode -> m ()) -> m ()) -> m ()
+withRetryWait2 wait RetryInterval2 {riSlow, riFast} action =
   callAction (0, initialInterval riSlow) (0, initialInterval riFast)
   where
     callAction :: (Int64, Int64) -> (Int64, Int64) -> m ()
@@ -102,14 +118,6 @@ withRetryLock2 RetryInterval2 {riSlow, riFast} lock action =
           let elapsed' = elapsed + delay
               delay' = nextRetryDelay elapsed' delay ri
           call (elapsed', delay')
-        wait delay = do
-          waiting <- newTVarIO True
-          _ <- liftIO . forkIO $ do
-            threadDelay' delay
-            atomically $ whenM (readTVar waiting) $ void $ tryPutTMVar lock ()
-          atomically $ do
-            takeTMVar lock
-            writeTVar waiting False
 
 nextRetryDelay :: Int64 -> Int64 -> RetryInterval -> Int64
 nextRetryDelay elapsed delay RetryInterval {increaseAfter, maxInterval} =

src/Simplex/Messaging/Agent/Store.hs

@@ -219,6 +219,7 @@ canAbortRcvSwitch ConnData {connAgentVersion} = maybe False canAbort . rcvSwchSt
       -- if switch is in RSReceivedMessage status, aborting switch (deleting new queue)
       -- will break the connection because the sender would have original queue deleted
       RSReceivedMessage -> False
+      RSReceivedQEND -> False
 
 data ClientNtfCreds = ClientNtfCreds
   { -- | key pair to be used by the notification server to authorize transmissions
@@ -489,7 +490,8 @@ data PendingCommand = PendingCommand
     corrId :: ACorrId,
     userId :: UserId,
     connId :: ConnId,
-    command :: AgentCommand
+    command :: AgentCommand,
+    createdAt :: UTCTime
   }
 
 data AgentCmdType = ACClient | ACInternal

src/Simplex/Messaging/Agent/Store/AgentStore.hs

@@ -76,6 +76,7 @@ module Simplex.Messaging.Agent.Store.AgentStore
     getExpiredServiceConns,
     deleteExpiredServiceRequests,
     getDeletedWaitingDeliveryConnIds,
+    getConnRatchetSync,
     setConnRatchetSync,
     addProcessedRatchetKeyHash,
     checkRatchetKeyHashExists,
@@ -99,6 +100,7 @@ module Simplex.Messaging.Agent.Store.AgentStore
     setSndQueuePrimary,
     deleteConnRcvQueue,
     incRcvDeleteErrors,
+    getEndedRcvQueueConnIds,
     deleteConnSndQueue,
     getPrimaryRcvQueue,
     getRcvQueue,
@@ -165,6 +167,8 @@ module Simplex.Messaging.Agent.Store.AgentStore
     createRatchet,
     deleteRatchet,
     getRatchet,
+    getRatchetVerifyCodes,
+    ratchetVerifyCodes,
     getRatchetForUpdate,
     getSkippedMsgKeys,
     updateRatchet,
@@ -213,6 +217,7 @@ module Simplex.Messaging.Agent.Store.AgentStore
     -- Rcv files
     createRcvFile,
     createRcvFileRedirect,
+    startPreparedRcvFile,
     lockRcvFileForUpdate,
     getRcvFile,
     getRcvFileByEntityId,
@@ -234,6 +239,7 @@ module Simplex.Messaging.Agent.Store.AgentStore
     getRcvFilesExpired,
     -- Snd files
     createSndFile,
+    startPreparedSndFile,
     lockSndFileForUpdate,
     getSndFile,
     getSndFileByEntityId,
@@ -291,6 +297,7 @@ import Data.ByteString (ByteString)
 import qualified Data.ByteString.Base64.URL as U
 import qualified Data.ByteString.Char8 as B
 import qualified Data.ByteString.Lazy as LB
+import Data.Either (partitionEithers)
 import Data.Functor (($>))
 import Data.Int (Int64)
 import Data.List (foldl', sortBy)
@@ -757,6 +764,10 @@ incRcvDeleteErrors :: DB.Connection -> RcvQueue -> IO ()
 incRcvDeleteErrors db RcvQueue {connId, dbQueueId} =
   DB.execute db "UPDATE rcv_queues SET delete_errors = delete_errors + 1 WHERE conn_id = ? AND rcv_queue_id = ?" (connId, dbQueueId)
 
+getEndedRcvQueueConnIds :: DB.Connection -> IO [ConnId]
+getEndedRcvQueueConnIds db =
+  map fromOnly <$> DB.query db "SELECT conn_id FROM rcv_queues WHERE switch_status = ? AND deleted = 0" (Only RSReceivedQEND)
+
 deleteConnRcvQueue :: DB.Connection -> RcvQueue -> IO ()
 deleteConnRcvQueue db RcvQueue {connId, dbQueueId} =
   DB.execute db "DELETE FROM rcv_queues WHERE conn_id = ? AND rcv_queue_id = ?" (connId, dbQueueId)
@@ -1249,7 +1260,7 @@ getLastMsg db connId msgId =
         LEFT JOIN snd_messages s ON s.conn_id = r.conn_id AND s.rcpt_internal_id = r.internal_id
         WHERE r.conn_id = ? AND r.broker_id = ?
       |]
-      (connId, Binary msgId)
+      (connId, msgId)
 
 toRcvMsg :: (Int64, InternalTs, BrokerId, BrokerTs) :. (AgentMsgId, MsgIntegrity, MsgHash, AgentMessageType, MsgBody, PQEncryption, Maybe AgentMsgId, Maybe MsgReceiptStatus, BoolInt) -> RcvMsg
 toRcvMsg ((agentMsgId, internalTs, brokerId, brokerTs) :. (sndMsgId, integrity, internalHash, msgType, msgBody, pqEncryption, rcptInternalId_, rcptStatus_, BI userAck)) =
@@ -1281,7 +1292,7 @@ checkRcvMsgHashExists db connId hash =
 getRcvMsgBrokerTs :: DB.Connection -> ConnId -> SMP.MsgId -> IO (Either StoreError BrokerTs)
 getRcvMsgBrokerTs db connId msgId =
   firstRow fromOnly (SEMsgNotFound "getRcvMsgBrokerTs") $
-    DB.query db "SELECT broker_ts FROM rcv_messages WHERE conn_id = ? AND broker_id = ?" (connId, Binary msgId)
+    DB.query db "SELECT broker_ts FROM rcv_messages WHERE conn_id = ? AND broker_id = ?" (connId, msgId)
 
 deleteMsg :: DB.Connection -> ConnId -> InternalId -> IO ()
 deleteMsg db connId msgId =
@@ -1474,9 +1485,11 @@ createSndRatchet db connId ratchetState (CR.AE2ERatchetParams s (CR.E2ERatchetPa
     db
     [sql|
       INSERT INTO ratchets
-        (conn_id, ratchet_state, x3dh_pub_key_1, x3dh_pub_key_2, pq_pub_kem) VALUES (?, ?, ?, ?, ?)
+        (conn_id, ratchet_state, rc_verify_code_ad, rc_verify_code_pq, x3dh_pub_key_1, x3dh_pub_key_2, pq_pub_kem) VALUES (?, ?, ?, ?, ?, ?, ?)
       ON CONFLICT (conn_id) DO UPDATE SET
         ratchet_state = EXCLUDED.ratchet_state,
+        rc_verify_code_ad = EXCLUDED.rc_verify_code_ad,
+        rc_verify_code_pq = EXCLUDED.rc_verify_code_pq,
         x3dh_priv_key_1 = NULL,
         x3dh_priv_key_2 = NULL,
         x3dh_pub_key_1 = EXCLUDED.x3dh_pub_key_1,
@@ -1484,7 +1497,13 @@ createSndRatchet db connId ratchetState (CR.AE2ERatchetParams s (CR.E2ERatchetPa
         pq_priv_kem = NULL,
         pq_pub_kem = EXCLUDED.pq_pub_kem
     |]
-    (connId, ratchetState, x3dhPubKey1, x3dhPubKey2, CR.ARKP s <$> pqPubKem)
+    ((connId, ratchetState) :. verifyCodesRow (ratchetVerifyCodes ratchetState) :. (x3dhPubKey1, x3dhPubKey2, CR.ARKP s <$> pqPubKem))
+
+ratchetVerifyCodes :: RatchetX448 -> ConnVerifyCodes
+ratchetVerifyCodes CR.Ratchet {rcAD, rcVCPQ} = ConnVerifyCodes {codeAD = C.sha256Hash $ unStr rcAD, codePQ = unStr <$> rcVCPQ}
+
+verifyCodesRow :: ConnVerifyCodes -> (Binary ByteString, Maybe (Binary ByteString))
+verifyCodesRow ConnVerifyCodes {codeAD, codePQ} = (Binary codeAD, Binary <$> codePQ)
 
 getSndRatchet :: DB.Connection -> ConnId -> CR.VersionE2E -> IO (Either StoreError (RatchetX448, CR.AE2ERatchetParams 'C.X448))
 getSndRatchet db connId v =
@@ -1505,10 +1524,12 @@ createRatchet db connId rc =
   DB.execute
     db
     [sql|
-      INSERT INTO ratchets (conn_id, ratchet_state)
-      VALUES (?, ?)
+      INSERT INTO ratchets (conn_id, ratchet_state, rc_verify_code_ad, rc_verify_code_pq)
+      VALUES (?, ?, ?, ?)
       ON CONFLICT (conn_id) DO UPDATE SET
-        ratchet_state = ?,
+        ratchet_state = EXCLUDED.ratchet_state,
+        rc_verify_code_ad = EXCLUDED.rc_verify_code_ad,
+        rc_verify_code_pq = EXCLUDED.rc_verify_code_pq,
         x3dh_priv_key_1 = NULL,
         x3dh_priv_key_2 = NULL,
         x3dh_pub_key_1 = NULL,
@@ -1516,7 +1537,42 @@ createRatchet db connId rc =
         pq_priv_kem = NULL,
         pq_pub_kem = NULL
     |]
-    (connId, rc, rc)
+    ((connId, rc) :. verifyCodesRow (ratchetVerifyCodes rc))
+
+getRatchetVerifyCodes :: DB.Connection -> [ConnId] -> IO (Map ConnId ConnVerifyCodes)
+getRatchetVerifyCodes db connIds = do
+  (saved, computed) <- partitionEithers . mapMaybe rowCodes <$> selectCodes
+  stored <- saveCodes computed
+  pure $ M.fromList $ saved <> stored
+  where
+    rowCodes (connId, codeAD_, codePQ_, rc_) = case codeAD_ of
+      Just codeAD -> Just $ Left $ toCodes (connId, codeAD, codePQ_)
+      Nothing -> Right . (connId,) . ratchetVerifyCodes <$> rc_
+    toCodes (connId, Binary codeAD, codePQ_) = (connId, ConnVerifyCodes {codeAD, codePQ = fromBinary <$> codePQ_})
+    codesRow (connId, cs) = verifyCodesRow cs :. Only connId
+    codesQuery :: Query
+    codesQuery = "SELECT conn_id, rc_verify_code_ad, rc_verify_code_pq, CASE WHEN rc_verify_code_ad IS NULL THEN ratchet_state END FROM ratchets"
+    selectCodes :: IO [(ConnId, Maybe (Binary ByteString), Maybe (Binary ByteString), Maybe RatchetX448)]
+    saveCodes :: [(ConnId, ConnVerifyCodes)] -> IO [(ConnId, ConnVerifyCodes)]
+#if defined(dbPostgres)
+    selectCodes = DB.query db (codesQuery <> " WHERE conn_id IN ?") (Only (In connIds))
+    saveCodes cs =
+      map toCodes
+        <$> DB.returning
+          db
+          [sql|
+            UPDATE ratchets r
+            SET rc_verify_code_ad = COALESCE(r.rc_verify_code_ad, (upd.rc_verify_code_ad :: BYTEA)),
+              rc_verify_code_pq = CASE WHEN r.rc_verify_code_ad IS NULL THEN (upd.rc_verify_code_pq :: BYTEA) ELSE r.rc_verify_code_pq END
+            FROM (VALUES(?, ?, ?)) AS upd(rc_verify_code_ad, rc_verify_code_pq, conn_id)
+            WHERE r.conn_id = (upd.conn_id :: BYTEA)
+            RETURNING r.conn_id, r.rc_verify_code_ad, r.rc_verify_code_pq
+          |]
+          (map codesRow cs)
+#else
+    selectCodes = concat <$> mapM (DB.query db (codesQuery <> " WHERE conn_id = ?") . Only) connIds
+    saveCodes cs = cs <$ unless (null cs) (DB.executeMany db "UPDATE ratchets SET rc_verify_code_ad = ?, rc_verify_code_pq = ? WHERE conn_id = ?" $ map codesRow cs)
+#endif
 
 deleteRatchet :: DB.Connection -> ConnId -> IO ()
 deleteRatchet db connId =
@@ -1544,12 +1600,16 @@ getRatchet_ q db connId =
   where
     ratchet = maybe (Left SERatchetNotFound) Right . fromOnly
 
-getSkippedMsgKeys :: DB.Connection -> ConnId -> IO SkippedMsgKeys
-getSkippedMsgKeys db connId =
-  skipped <$> DB.query db "SELECT header_key, msg_n, msg_key FROM skipped_messages WHERE conn_id = ?" (Only connId)
+getSkippedMsgKeys :: DB.Connection -> ConnId -> Int -> IO SkippedMsgKeys
+getSkippedMsgKeys db connId maxKeys = do
+  (keys, oldKeys) <- splitAt maxKeys <$> DB.query db "SELECT skipped_message_id, header_key, msg_n, msg_key FROM skipped_messages WHERE conn_id = ? ORDER BY skipped_message_id DESC LIMIT ?" (connId, maxKeys + 1)
+  case oldKeys of
+    (skippedMsgId :: Int64, _, _, _) : _ -> DB.execute db "DELETE FROM skipped_messages WHERE conn_id = ? AND skipped_message_id <= ?" (connId, skippedMsgId)
+    [] -> pure ()
+  pure $ skipped keys
   where
     skipped = foldl' addSkippedKey M.empty
-    addSkippedKey smks (hk, msgN, mk) = M.alter (Just . addMsgKey) hk smks
+    addSkippedKey smks (_, hk, msgN, mk) = M.alter (Just . addMsgKey) hk smks
       where
         addMsgKey = maybe (M.singleton msgN mk) (M.insert msgN mk)
 
@@ -1668,7 +1728,7 @@ getPendingServerCommand db connId srv_ = getWorkItem "command" getCmdId getComma
         DB.query
           db
           [sql|
-            SELECT c.corr_id, cs.user_id, c.command
+            SELECT c.corr_id, cs.user_id, c.command, c.created_at
             FROM commands c
             JOIN connections cs USING (conn_id)
             WHERE c.command_id = ?
@@ -1676,7 +1736,7 @@ getPendingServerCommand db connId srv_ = getWorkItem "command" getCmdId getComma
           (Only cmdId)
       where
         err = SEInternal $ "command  " <> bshow cmdId <> " returned []"
-        pendingCommand (corrId, userId, command) = PendingCommand {cmdId, corrId, userId, connId, command}
+        pendingCommand (corrId, userId, command, createdAt) = PendingCommand {cmdId, corrId, userId, connId, command, createdAt}
     markCommandFailed cmdId = DB.execute db "UPDATE commands SET failed = 1 WHERE command_id = ?" (Only cmdId)
 
 updateCommandServer :: DB.Connection -> AsyncCmdId -> SMPServer -> IO (Either StoreError ())
@@ -2450,12 +2510,12 @@ getSubscriptionServers db onlyNeeded =
 
 -- TODO [certs rcv] check index for getting queues with service present
 getUserServerRcvQueueSubs :: DB.Connection -> UserId -> SMPServer -> Bool -> ServiceAssoc -> Int -> Maybe SMP.RecipientId -> IO [RcvQueueSub]
-getUserServerRcvQueueSubs db userId (SMPServer h p kh) onlyNeeded hasService limit cursor_ =
-  map toRcvQueueSub <$> case cursor_ of
+getUserServerRcvQueueSubs db userId srv@(SMPServer h p kh) onlyNeeded hasService limit cursor_ =
+  map (toServerRcvQueueSub srv) <$> case cursor_ of
     Nothing -> DB.query db (q <> orderLimit) (userId, h, p, kh, limit)
     Just cursor -> DB.query db (q <> " AND q.rcv_id > ? " <> orderLimit) (userId, h, p, kh, cursor, limit)
   where
-    q = rcvQueueSubQuery <> toSubscribe <> " c.deleted = 0 AND q.deleted = 0 AND c.user_id = ? AND q.host = ? AND q.port = ? AND COALESCE(q.server_key_hash, s.key_hash) = ?" <> serviceCond
+    q = serverRcvQueueSubQuery <> toSubscribe <> " c.deleted = 0 AND q.deleted = 0 AND c.user_id = ? AND q.host = ? AND q.port = ? AND COALESCE(q.server_key_hash, s.key_hash) = ?" <> serviceCond
     orderLimit = " ORDER BY q.rcv_id LIMIT ?"
     toSubscribe
       | onlyNeeded = " WHERE q.to_subscribe = 1 AND "
@@ -2468,7 +2528,7 @@ unassocUserServerRcvQueueSubs :: DB.Connection -> UserId -> SMPServer -> IO [Rcv
 unassocUserServerRcvQueueSubs db userId srv@(SMPServer h p kh) = do
   deleteClientService db userId srv
 #if defined(dbPostgres)
-  map toRcvQueueSub
+  map (toServerRcvQueueSub srv)
     <$> DB.query
       db
       (removeRcvAssocsQuery <> " " <> returningColumns)
@@ -2476,15 +2536,15 @@ unassocUserServerRcvQueueSubs db userId srv@(SMPServer h p kh) = do
   where
     returningColumns =
       [sql|
-        RETURNING c.user_id, rcv_queues.conn_id, rcv_queues.host, rcv_queues.port, COALESCE(rcv_queues.server_key_hash, s.key_hash),
+        RETURNING c.user_id, rcv_queues.conn_id,
           rcv_queues.rcv_id, rcv_queues.rcv_private_key, rcv_queues.status, c.enable_ntfs, rcv_queues.client_notice_id,
           rcv_queues.rcv_queue_id, rcv_queues.rcv_primary, rcv_queues.replace_rcv_queue_id
       |]
 #else
-  qs <- map toRcvQueueSub
+  qs <- map (toServerRcvQueueSub srv)
     <$> DB.query
       db
-      (rcvQueueSubQuery <> " WHERE c.user_id = ? AND q.host = ? AND q.port = ? AND COALESCE(q.server_key_hash, s.key_hash) = ? AND q.rcv_service_assoc = 1")
+      (serverRcvQueueSubQuery <> " WHERE c.user_id = ? AND q.host = ? AND q.port = ? AND COALESCE(q.server_key_hash, s.key_hash) = ? AND q.rcv_service_assoc = 1")
       (userId, h, p, kh)
   DB.execute db removeRcvAssocsQuery (h, p, userId, kh)
   pure qs
@@ -2743,6 +2803,11 @@ getDeletedWaitingDeliveryConnIds :: DB.Connection -> IO [ConnId]
 getDeletedWaitingDeliveryConnIds db =
   map fromOnly <$> DB.query_ db "SELECT conn_id FROM connections WHERE deleted_at_wait_delivery IS NOT NULL"
 
+getConnRatchetSync :: DB.Connection -> ConnId -> IO (Either StoreError RatchetSyncState)
+getConnRatchetSync db connId =
+  firstRow fromOnly SEConnNotFound $
+    DB.query db "SELECT ratchet_sync_state FROM connections WHERE conn_id = ?" (Only connId)
+
 setConnRatchetSync :: DB.Connection -> ConnId -> RatchetSyncState -> IO ()
 setConnRatchetSync db connId ratchetSyncState =
   DB.execute db "UPDATE connections SET ratchet_sync_state = ? WHERE conn_id = ?" (ratchetSyncState, connId)
@@ -2764,21 +2829,35 @@ checkRatchetKeyHashExists db connId hash =
       (connId, Binary hash)
 
 deleteRatchetKeyHashesExpired :: DB.Connection -> NominalDiffTime -> Int -> IO ()
-deleteRatchetKeyHashesExpired db ttl limit = do
+deleteRatchetKeyHashesExpired db ttl maxConnHashes = do
   cutoffTs <- addUTCTime (-ttl) <$> getCurrentTime
+#if defined(dbPostgres)
   DB.execute
     db
-    [sql|
-      DELETE FROM processed_ratchet_key_hashes
-      WHERE processed_ratchet_key_hash_id IN (
-        SELECT processed_ratchet_key_hash_id
-        FROM processed_ratchet_key_hashes
-        WHERE created_at < ?
-        ORDER BY created_at ASC
-        LIMIT ?
-      )
-    |]
-    (cutoffTs, limit)
+    ("DELETE FROM processed_ratchet_key_hashes h USING (" <> maxExcessIdsQuery <> ") e WHERE h.conn_id = e.conn_id AND h.processed_ratchet_key_hash_id <= e.max_excess_id AND h.created_at < ?")
+    (maxConnHashes, maxConnHashes, cutoffTs)
+#else
+  maxExcessIds <- DB.query db maxExcessIdsQuery (maxConnHashes, maxConnHashes)
+  DB.executeMany
+    db
+    "DELETE FROM processed_ratchet_key_hashes WHERE conn_id = ? AND processed_ratchet_key_hash_id <= ? AND created_at < ?"
+    (map (\(connId :: ConnId, maxExcessId :: Int64) -> (connId, maxExcessId, cutoffTs)) maxExcessIds)
+#endif
+  where
+    maxExcessIdsQuery :: Query
+    maxExcessIdsQuery =
+      [sql|
+        SELECT conn_id, (
+          SELECT processed_ratchet_key_hash_id
+          FROM processed_ratchet_key_hashes
+          WHERE conn_id = c.conn_id
+          ORDER BY processed_ratchet_key_hash_id DESC
+          LIMIT 1 OFFSET ?
+        ) AS max_excess_id
+        FROM processed_ratchet_key_hashes c
+        GROUP BY conn_id
+        HAVING COUNT(*) > ?
+      |]
 
 -- | returns all connection queues, the first queue is the primary one
 getRcvQueuesByConnId_ :: DB.Connection -> ConnId -> IO (Maybe (NonEmpty RcvQueue))
@@ -2835,17 +2914,32 @@ getRcvQueueSubsByConnId_ db connId =
 rcvQueueSubQuery :: Query
 rcvQueueSubQuery =
   [sql|
-    SELECT c.user_id, q.conn_id, q.host, q.port, COALESCE(q.server_key_hash, s.key_hash), q.rcv_id, q.rcv_private_key, q.status, c.enable_ntfs, q.client_notice_id,
+    SELECT q.host, q.port, COALESCE(q.server_key_hash, s.key_hash), c.user_id, q.conn_id, q.rcv_id, q.rcv_private_key, q.status, c.enable_ntfs, q.client_notice_id,
       q.rcv_queue_id, q.rcv_primary, q.replace_rcv_queue_id
     FROM rcv_queues q
     JOIN servers s ON q.host = s.host AND q.port = s.port
     JOIN connections c ON q.conn_id = c.conn_id
   |]
 
-toRcvQueueSub :: (UserId, ConnId, NonEmpty TransportHost, ServiceName, C.KeyHash, SMP.RecipientId, SMP.RcvPrivateAuthKey) :. (QueueStatus, Maybe BoolInt, Maybe NoticeId, Int64, BoolInt, Maybe Int64) -> RcvQueueSub
-toRcvQueueSub ((userId, connId, host, port, keyHash, rcvId, rcvPrivateKey) :. (status, enableNtfs_, clientNoticeId, dbQueueId, BI primary, dbReplaceQueueId)) =
+serverRcvQueueSubQuery :: Query
+serverRcvQueueSubQuery =
+  [sql|
+    SELECT c.user_id, q.conn_id, q.rcv_id, q.rcv_private_key, q.status, c.enable_ntfs, q.client_notice_id,
+      q.rcv_queue_id, q.rcv_primary, q.replace_rcv_queue_id
+    FROM rcv_queues q
+    JOIN servers s ON q.host = s.host AND q.port = s.port
+    JOIN connections c ON q.conn_id = c.conn_id
+  |]
+
+type RcvQueueSubRow = (UserId, ConnId, SMP.RecipientId, SMP.RcvPrivateAuthKey) :. (QueueStatus, Maybe BoolInt, Maybe NoticeId, Int64, BoolInt, Maybe Int64)
+
+toRcvQueueSub :: (NonEmpty TransportHost, ServiceName, C.KeyHash) :. RcvQueueSubRow -> RcvQueueSub
+toRcvQueueSub ((host, port, keyHash) :. r) = toServerRcvQueueSub (SMPServer host port keyHash) r
+
+toServerRcvQueueSub :: SMPServer -> RcvQueueSubRow -> RcvQueueSub
+toServerRcvQueueSub server ((userId, connId, rcvId, rcvPrivateKey) :. (status, enableNtfs_, clientNoticeId, dbQueueId, BI primary, dbReplaceQueueId)) =
   let enableNtfs = maybe True unBI enableNtfs_
-   in RcvQueueSub {userId, connId, server = SMPServer host port keyHash, rcvId, rcvPrivateKey, status, enableNtfs, clientNoticeId, dbQueueId, primary, dbReplaceQueueId}
+   in RcvQueueSub {userId, connId, server, rcvId, rcvPrivateKey, status, enableNtfs, clientNoticeId, dbQueueId, primary, dbReplaceQueueId}
 
 getRcvQueueById :: DB.Connection -> ConnId -> Int64 -> IO (Either StoreError RcvQueue)
 getRcvQueueById db connId dbRcvId =
@@ -2946,7 +3040,7 @@ insertRcvMsgDetails_ db connId RcvQueue {dbQueueId} RcvMsgData {msgMeta, interna
       VALUES
         (?,?,?,?,?,?,?,?,?,?)
     |]
-    (connId, dbQueueId, internalRcvId, fst recipient, sndMsgId, Binary (fst broker), snd broker, Binary internalHash, Binary externalPrevSndHash, integrity)
+    (connId, dbQueueId, internalRcvId, fst recipient, sndMsgId, fst broker, snd broker, Binary internalHash, Binary externalPrevSndHash, integrity)
   DB.execute db "INSERT INTO encrypted_rcv_message_hashes (conn_id, hash) VALUES (?,?)" (connId, Binary encryptedMsgHash)
 
 updateRcvMsgHash :: DB.Connection -> ConnId -> AgentMsgId -> InternalRcvId -> MsgHash -> IO ()
@@ -3094,7 +3188,8 @@ createRcvFile db gVar userId fd@FileDescription {chunks} prefixPath tmpPath file
 createRcvFileRedirect :: DB.Connection -> TVar ChaChaDRG -> UserId -> FileDescription 'FRecipient -> FilePath -> FilePath -> CryptoFile -> FilePath -> CryptoFile -> Bool -> IO (Either StoreError RcvFileId)
 createRcvFileRedirect _ _ _ FileDescription {redirect = Nothing} _ _ _ _ _ _ = pure $ Left $ SEInternal "createRcvFileRedirect called without redirect"
 createRcvFileRedirect db gVar userId redirectFd@FileDescription {chunks = redirectChunks, redirect = Just RedirectFileInfo {size, digest}} prefixPath redirectPath redirectFile dstPath dstFile approvedRelays = runExceptT $ do
-  (dstEntityId, dstId) <- ExceptT $ insertRcvFile db gVar userId dummyDst prefixPath dstPath dstFile Nothing Nothing approvedRelays
+  nonce <- atomically $ C.randomCbNonce gVar
+  (dstEntityId, dstId) <- ExceptT $ insertRcvFile db gVar userId (dummyDst nonce) prefixPath dstPath dstFile Nothing Nothing approvedRelays
   (_, redirectId) <- ExceptT $ insertRcvFile db gVar userId redirectFd prefixPath redirectPath redirectFile (Just dstId) (Just dstEntityId) approvedRelays
   liftIO $
     forM_ redirectChunks $ \fc@FileChunk {replicas} -> do
@@ -3102,7 +3197,7 @@ createRcvFileRedirect db gVar userId redirectFd@FileDescription {chunks = redire
       forM_ (zip [1 ..] replicas) $ \(rno, replica) -> insertRcvFileChunkReplica db rno replica chunkId
   pure dstEntityId
   where
-    dummyDst =
+    dummyDst nonce =
       FileDescription
         { party = SFRecipient,
           size,
@@ -3110,11 +3205,34 @@ createRcvFileRedirect db gVar userId redirectFd@FileDescription {chunks = redire
           redirect = Nothing,
           -- updated later with updateRcvFileRedirect
           key = C.unsafeSbKey $ B.replicate 32 '#',
-          nonce = C.cbNonce "",
+          nonce,
           chunkSize = FileSize 0,
           chunks = []
         }
 
+startPreparedRcvFile :: DB.Connection -> RcvFileId -> IO (Either StoreError [XFTPServer])
+startPreparedRcvFile db rcvFileEntityId = runExceptT $ do
+  rcvFileId <- ExceptT $ getRcvFileIdByEntityId_ db rcvFileEntityId
+  liftIO $ do
+    updatedAt <- getCurrentTime
+    DB.execute
+      db
+      "UPDATE rcv_files SET status = ?, updated_at = ? WHERE (rcv_file_id = ? OR redirect_id = ?) AND status = ?"
+      (RFSReceiving, updatedAt, rcvFileId, rcvFileId, RFSPrepared)
+    map toXFTPServer
+      <$> DB.query
+        db
+        [sql|
+          SELECT DISTINCT
+            s.xftp_host, s.xftp_port, s.xftp_key_hash
+          FROM rcv_file_chunk_replicas r
+          JOIN xftp_servers s ON s.xftp_server_id = r.xftp_server_id
+          JOIN rcv_file_chunks c ON c.rcv_file_chunk_id = r.rcv_file_chunk_id
+          JOIN rcv_files f ON f.rcv_file_id = c.rcv_file_id
+          WHERE (f.rcv_file_id = ? OR f.redirect_id = ?) AND r.replica_number = 1
+        |]
+        (rcvFileId, rcvFileId)
+
 insertRcvFile :: DB.Connection -> TVar ChaChaDRG -> UserId -> FileDescription 'FRecipient -> FilePath -> FilePath -> CryptoFile -> Maybe DBRcvFileId -> Maybe RcvFileId -> Bool -> IO (Either StoreError (RcvFileId, DBRcvFileId))
 insertRcvFile db gVar userId FileDescription {size, digest, key, nonce, chunkSize, redirect} prefixPath tmpPath (CryptoFile savePath cfArgs) redirectId_ redirectEntityId_ approvedRelays = runExceptT $ do
   let (redirectDigest_, redirectSize_) = case redirect of
@@ -3125,7 +3243,7 @@ insertRcvFile db gVar userId FileDescription {size, digest, key, nonce, chunkSiz
       DB.execute
         db
         "INSERT INTO rcv_files (rcv_file_entity_id, user_id, size, digest, key, nonce, chunk_size, prefix_path, tmp_path, save_path, save_file_key, save_file_nonce, status, redirect_id, redirect_entity_id, redirect_digest, redirect_size, approved_relays) VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?,?)"
-        ((Binary rcvFileEntityId, userId, size, digest, key, nonce, chunkSize, prefixPath, tmpPath) :. (savePath, fileKey <$> cfArgs, fileNonce <$> cfArgs, RFSReceiving, redirectId_, Binary <$> redirectEntityId_, redirectDigest_, redirectSize_, BI approvedRelays))
+        ((Binary rcvFileEntityId, userId, size, digest, key, nonce, chunkSize, prefixPath, tmpPath) :. (savePath, fileKey <$> cfArgs, fileNonce <$> cfArgs, RFSPrepared, redirectId_, Binary <$> redirectEntityId_, redirectDigest_, redirectSize_, BI approvedRelays))
   rcvFileId <- liftIO $ insertedRowId db
   pure (rcvFileEntityId, rcvFileId)
 
@@ -3430,13 +3548,23 @@ createSndFile db gVar userId (CryptoFile path cfArgs) numRecipients prefixPath k
     DB.execute
       db
       "INSERT INTO snd_files (snd_file_entity_id, user_id, path, src_file_key, src_file_nonce, num_recipients, prefix_path, key, nonce, status, redirect_size, redirect_digest, storage_time) VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?)"
-      ((Binary sndFileEntityId, userId, path, fileKey <$> cfArgs, fileNonce <$> cfArgs, numRecipients) :. (prefixPath, key, nonce, SFSNew, redirectSize_, redirectDigest_, storageHours))
+      ((Binary sndFileEntityId, userId, path, fileKey <$> cfArgs, fileNonce <$> cfArgs, numRecipients) :. (prefixPath, key, nonce, SFSPrepared, redirectSize_, redirectDigest_, storageHours))
   where
     (redirectSize_, redirectDigest_) =
       case redirect_ of
         Nothing -> (Nothing, Nothing)
         Just RedirectFileInfo {size, digest} -> (Just size, Just digest)
 
+startPreparedSndFile :: DB.Connection -> SndFileId -> IO (Either StoreError ())
+startPreparedSndFile db sndFileEntityId = runExceptT $ do
+  sndFileId <- ExceptT $ getSndFileIdByEntityId_ db sndFileEntityId
+  liftIO $ do
+    updatedAt <- getCurrentTime
+    DB.execute
+      db
+      "UPDATE snd_files SET status = ?, updated_at = ? WHERE snd_file_id = ? AND status = ?"
+      (SFSNew, updatedAt, sndFileId, SFSPrepared)
+
 getSndFileByEntityId :: DB.Connection -> SndFileId -> IO (Either StoreError SndFile)
 getSndFileByEntityId db sndFileEntityId = runExceptT $ do
   sndFileId <- ExceptT $ getSndFileIdByEntityId_ db sndFileEntityId

src/Simplex/Messaging/Agent/Store/Postgres/DB.hs

@@ -13,6 +13,7 @@ module Simplex.Messaging.Agent.Store.Postgres.DB
     execute,
     execute_,
     executeMany,
+    returning,
     query,
     query_,
     blobFieldDecoder,
@@ -60,6 +61,10 @@ executeMany :: ToRow q => PSQL.Connection -> Query -> [q] -> IO ()
 executeMany db q qs = void $ PSQL.executeMany db q qs `E.catch` addSql q
 {-# INLINE executeMany #-}
 
+returning :: (ToRow q, FromRow r) => PSQL.Connection -> Query -> [q] -> IO [r]
+returning db q qs = PSQL.returning db q qs `E.catch` addSql q
+{-# INLINE returning #-}
+
 query :: (ToRow q, FromRow r) => PSQL.Connection -> Query -> q -> IO [r]
 query db q qs = PSQL.query db q qs `E.catch` addSql q
 {-# INLINE query #-}

src/Simplex/Messaging/Agent/Store/Postgres/Migrations/App.hs

@@ -15,6 +15,9 @@ import Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260410_receive_attem
 import Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260411_service_certs
 import Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260712_address_dr_rpc
 import Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260823_snd_files_entitlement
+import Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260919_ratchet_verify_codes
+import Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260929_ratchet_indexes
+import Simplex.Messaging.Agent.Store.Postgres.Migrations.M20261008_rcv_switch_status
 import Simplex.Messaging.Agent.Store.Shared (Migration (..))
 
 schemaMigrations :: [(String, Text, Maybe Text)]
@@ -29,7 +32,10 @@ schemaMigrations =
     ("20260410_receive_attempts", m20260410_receive_attempts, Just down_m20260410_receive_attempts),
     ("20260411_service_certs", m20260411_service_certs, Just down_m20260411_service_certs),
     ("20260712_address_dr_rpc", m20260712_address_dr_rpc, Just down_m20260712_address_dr_rpc),
-    ("20260823_snd_files_entitlement", m20260823_snd_files_entitlement, Just down_m20260823_snd_files_entitlement)
+    ("20260823_snd_files_entitlement", m20260823_snd_files_entitlement, Just down_m20260823_snd_files_entitlement),
+    ("20260919_ratchet_verify_codes", m20260919_ratchet_verify_codes, Just down_m20260919_ratchet_verify_codes),
+    ("20260929_ratchet_indexes", m20260929_ratchet_indexes, Just down_m20260929_ratchet_indexes),
+    ("20261008_rcv_switch_status", m20261008_rcv_switch_status, Just down_m20261008_rcv_switch_status)
   ]
 
 -- | The list of migrations in ascending order by date

src/Simplex/Messaging/Agent/Store/Postgres/Migrations/M20260919_ratchet_verify_codes.hs

@@ -0,0 +1,21 @@
+{-# LANGUAGE OverloadedStrings #-}
+{-# LANGUAGE QuasiQuotes #-}
+
+module Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260919_ratchet_verify_codes where
+
+import Data.Text (Text)
+import Text.RawString.QQ (r)
+
+m20260919_ratchet_verify_codes :: Text
+m20260919_ratchet_verify_codes =
+  [r|
+ALTER TABLE ratchets ADD COLUMN rc_verify_code_ad BYTEA;
+ALTER TABLE ratchets ADD COLUMN rc_verify_code_pq BYTEA;
+|]
+
+down_m20260919_ratchet_verify_codes :: Text
+down_m20260919_ratchet_verify_codes =
+  [r|
+ALTER TABLE ratchets DROP COLUMN rc_verify_code_pq;
+ALTER TABLE ratchets DROP COLUMN rc_verify_code_ad;
+|]

src/Simplex/Messaging/Agent/Store/Postgres/Migrations/M20260929_ratchet_indexes.hs

@@ -0,0 +1,34 @@
+{-# LANGUAGE OverloadedStrings #-}
+{-# LANGUAGE QuasiQuotes #-}
+
+module Simplex.Messaging.Agent.Store.Postgres.Migrations.M20260929_ratchet_indexes where
+
+import Data.Text (Text)
+import Text.RawString.QQ (r)
+
+m20260929_ratchet_indexes :: Text
+m20260929_ratchet_indexes =
+  [r|
+DROP INDEX idx_skipped_messages_conn_id;
+CREATE INDEX idx_skipped_messages_conn_id ON skipped_messages(conn_id, skipped_message_id);
+
+DELETE FROM processed_ratchet_key_hashes AS h
+WHERE EXISTS (
+  SELECT 1 FROM processed_ratchet_key_hashes d
+  WHERE d.conn_id = h.conn_id AND d.hash = h.hash AND d.processed_ratchet_key_hash_id < h.processed_ratchet_key_hash_id
+);
+
+DROP INDEX idx_processed_ratchet_key_hashes_hash;
+CREATE UNIQUE INDEX idx_processed_ratchet_key_hashes_hash ON processed_ratchet_key_hashes(conn_id, hash);
+CREATE INDEX idx_processed_ratchet_key_hashes_conn_id ON processed_ratchet_key_hashes(conn_id, processed_ratchet_key_hash_id);
+|]
+
+down_m20260929_ratchet_indexes :: Text
+down_m20260929_ratchet_indexes =
+  [r|
+DROP INDEX idx_processed_ratchet_key_hashes_conn_id;
+DROP INDEX idx_processed_ratchet_key_hashes_hash;
+CREATE INDEX idx_processed_ratchet_key_hashes_hash ON processed_ratchet_key_hashes(conn_id, hash);
+DROP INDEX idx_skipped_messages_conn_id;
+CREATE INDEX idx_skipped_messages_conn_id ON skipped_messages(conn_id);
+|]

src/Simplex/Messaging/Agent/Store/Postgres/Migrations/M20261008_rcv_switch_status.hs

@@ -0,0 +1,20 @@
+{-# LANGUAGE OverloadedStrings #-}
+{-# LANGUAGE QuasiQuotes #-}
+
+module Simplex.Messaging.Agent.Store.Postgres.Migrations.M20261008_rcv_switch_status where
+
+import Data.Text (Text)
+import Text.RawString.QQ (r)
+
+m20261008_rcv_switch_status :: Text
+m20261008_rcv_switch_status =
+  [r|
+CREATE INDEX idx_rcv_queues_switch_status ON rcv_queues(switch_status);
+|]
+
+down_m20261008_rcv_switch_status :: Text
+down_m20261008_rcv_switch_status =
+  [r|
+UPDATE rcv_queues SET switch_status = NULL WHERE switch_status = 'received_qend';
+DROP INDEX idx_rcv_queues_switch_status;
+|]

src/Simplex/Messaging/Agent/Store/Postgres/Migrations/agent_postgres_schema.sql

@@ -448,7 +448,9 @@ CREATE TABLE smp_agent_test_protocol_schema.ratchets (
     x3dh_pub_key_1 bytea,
     x3dh_pub_key_2 bytea,
     pq_priv_kem bytea,
-    pq_pub_kem bytea
+    pq_pub_kem bytea,
+    rc_verify_code_ad bytea,
+    rc_verify_code_pq bytea
 );
 
 
@@ -1167,11 +1169,15 @@ CREATE INDEX idx_ntf_tokens_ntf_host_ntf_port ON smp_agent_test_protocol_schema.
 
 
 
+CREATE INDEX idx_processed_ratchet_key_hashes_conn_id ON smp_agent_test_protocol_schema.processed_ratchet_key_hashes USING btree (conn_id, processed_ratchet_key_hash_id);
+
+
+
 CREATE INDEX idx_processed_ratchet_key_hashes_created_at ON smp_agent_test_protocol_schema.processed_ratchet_key_hashes USING btree (created_at);
 
 
 
-CREATE INDEX idx_processed_ratchet_key_hashes_hash ON smp_agent_test_protocol_schema.processed_ratchet_key_hashes USING btree (conn_id, hash);
+CREATE UNIQUE INDEX idx_processed_ratchet_key_hashes_hash ON smp_agent_test_protocol_schema.processed_ratchet_key_hashes USING btree (conn_id, hash);
 
 
 
@@ -1227,6 +1233,10 @@ CREATE UNIQUE INDEX idx_rcv_queues_ntf ON smp_agent_test_protocol_schema.rcv_que
 
 
 
+CREATE INDEX idx_rcv_queues_switch_status ON smp_agent_test_protocol_schema.rcv_queues USING btree (switch_status);
+
+
+
 CREATE INDEX idx_rcv_queues_to_subscribe ON smp_agent_test_protocol_schema.rcv_queues USING btree (to_subscribe);
 
 
@@ -1239,7 +1249,7 @@ CREATE UNIQUE INDEX idx_server_certs_user_id_host_port ON smp_agent_test_protoco
 
 
 
-CREATE INDEX idx_skipped_messages_conn_id ON smp_agent_test_protocol_schema.skipped_messages USING btree (conn_id);
+CREATE INDEX idx_skipped_messages_conn_id ON smp_agent_test_protocol_schema.skipped_messages USING btree (conn_id, skipped_message_id);
 
 
 

src/Simplex/Messaging/Agent/Store/SQLite/Common.hs

@@ -100,7 +100,8 @@ withTransaction st = withTransactionPriority st False
 {-# INLINE withTransaction #-}
 
 withTransactionPriority :: DBStore -> Bool -> (DB.Connection -> IO a) -> IO a
-withTransactionPriority st priority action = withConnectionPriority st priority $ dbBusyLoop . transaction
+withTransactionPriority st@DBStore {dbClosed} priority action = withConnectionPriority st priority $ \db ->
+  ifM (readTVarIO dbClosed) (fail "withTransaction: database closed") (dbBusyLoop $ transaction db)
   where
     transaction db@DB.Connection {conn} = SQL.withImmediateTransaction conn $ action db
 

src/Simplex/Messaging/Agent/Store/SQLite/Migrations/App.hs

@@ -51,6 +51,10 @@ import Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260410_receive_attempt
 import Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260411_service_certs
 import Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260712_address_dr_rpc
 import Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260823_snd_files_entitlement
+import Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260919_ratchet_verify_codes
+import Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260929_ratchet_indexes
+import Simplex.Messaging.Agent.Store.SQLite.Migrations.M20261005_commands_autoincrement
+import Simplex.Messaging.Agent.Store.SQLite.Migrations.M20261008_rcv_switch_status
 import Simplex.Messaging.Agent.Store.Shared (Migration (..))
 
 schemaMigrations :: [(String, Query, Maybe Query)]
@@ -101,7 +105,11 @@ schemaMigrations =
     ("m20260410_receive_attempts", m20260410_receive_attempts, Just down_m20260410_receive_attempts),
     ("m20260411_service_certs", m20260411_service_certs, Just down_m20260411_service_certs),
     ("m20260712_address_dr_rpc", m20260712_address_dr_rpc, Just down_m20260712_address_dr_rpc),
-    ("m20260823_snd_files_entitlement", m20260823_snd_files_entitlement, Just down_m20260823_snd_files_entitlement)
+    ("m20260823_snd_files_entitlement", m20260823_snd_files_entitlement, Just down_m20260823_snd_files_entitlement),
+    ("m20260919_ratchet_verify_codes", m20260919_ratchet_verify_codes, Just down_m20260919_ratchet_verify_codes),
+    ("m20260929_ratchet_indexes", m20260929_ratchet_indexes, Just down_m20260929_ratchet_indexes),
+    ("m20261005_commands_autoincrement", m20261005_commands_autoincrement, Just down_m20261005_commands_autoincrement),
+    ("m20261008_rcv_switch_status", m20261008_rcv_switch_status, Just down_m20261008_rcv_switch_status)
   ]
 
 -- | The list of migrations in ascending order by date

src/Simplex/Messaging/Agent/Store/SQLite/Migrations/M20260919_ratchet_verify_codes.hs

@@ -0,0 +1,20 @@
+{-# LANGUAGE QuasiQuotes #-}
+
+module Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260919_ratchet_verify_codes where
+
+import Database.SQLite.Simple (Query)
+import Database.SQLite.Simple.QQ (sql)
+
+m20260919_ratchet_verify_codes :: Query
+m20260919_ratchet_verify_codes =
+  [sql|
+ALTER TABLE ratchets ADD COLUMN rc_verify_code_ad BLOB;
+ALTER TABLE ratchets ADD COLUMN rc_verify_code_pq BLOB;
+  |]
+
+down_m20260919_ratchet_verify_codes :: Query
+down_m20260919_ratchet_verify_codes =
+  [sql|
+ALTER TABLE ratchets DROP COLUMN rc_verify_code_pq;
+ALTER TABLE ratchets DROP COLUMN rc_verify_code_ad;
+  |]

src/Simplex/Messaging/Agent/Store/SQLite/Migrations/M20260929_ratchet_indexes.hs

@@ -0,0 +1,29 @@
+{-# LANGUAGE QuasiQuotes #-}
+
+module Simplex.Messaging.Agent.Store.SQLite.Migrations.M20260929_ratchet_indexes where
+
+import Database.SQLite.Simple (Query)
+import Database.SQLite.Simple.QQ (sql)
+
+-- idx_skipped_messages_conn_id is not changed: SQLite index entries include rowid (skipped_message_id)
+m20260929_ratchet_indexes :: Query
+m20260929_ratchet_indexes =
+  [sql|
+DELETE FROM processed_ratchet_key_hashes AS h
+WHERE EXISTS (
+  SELECT 1 FROM processed_ratchet_key_hashes d
+  WHERE d.conn_id = h.conn_id AND d.hash = h.hash AND d.processed_ratchet_key_hash_id < h.processed_ratchet_key_hash_id
+);
+
+DROP INDEX idx_processed_ratchet_key_hashes_hash;
+CREATE UNIQUE INDEX idx_processed_ratchet_key_hashes_hash ON processed_ratchet_key_hashes(conn_id, hash);
+CREATE INDEX idx_processed_ratchet_key_hashes_conn_id ON processed_ratchet_key_hashes(conn_id, processed_ratchet_key_hash_id);
+  |]
+
+down_m20260929_ratchet_indexes :: Query
+down_m20260929_ratchet_indexes =
+  [sql|
+DROP INDEX idx_processed_ratchet_key_hashes_conn_id;
+DROP INDEX idx_processed_ratchet_key_hashes_hash;
+CREATE INDEX idx_processed_ratchet_key_hashes_hash ON processed_ratchet_key_hashes(conn_id, hash);
+  |]

src/Simplex/Messaging/Agent/Store/SQLite/Migrations/M20261005_commands_autoincrement.hs

@@ -0,0 +1,33 @@
+{-# LANGUAGE QuasiQuotes #-}
+
+module Simplex.Messaging.Agent.Store.SQLite.Migrations.M20261005_commands_autoincrement where
+
+import Database.SQLite.Simple (Query)
+import Database.SQLite.Simple.QQ (sql)
+
+m20261005_commands_autoincrement :: Query
+m20261005_commands_autoincrement =
+  [sql|
+INSERT INTO sqlite_sequence (name, seq)
+SELECT 'commands', MAX(ROWID) FROM commands;
+
+PRAGMA writable_schema=1;
+
+UPDATE sqlite_master SET sql = replace(sql, 'command_id INTEGER PRIMARY KEY,', 'command_id INTEGER PRIMARY KEY AUTOINCREMENT,')
+WHERE name = 'commands' AND type = 'table';
+
+PRAGMA writable_schema=RESET;
+  |]
+
+down_m20261005_commands_autoincrement :: Query
+down_m20261005_commands_autoincrement =
+  [sql|
+DELETE FROM sqlite_sequence WHERE name = 'commands';
+
+PRAGMA writable_schema=1;
+
+UPDATE sqlite_master SET sql = replace(sql, 'command_id INTEGER PRIMARY KEY AUTOINCREMENT,', 'command_id INTEGER PRIMARY KEY,')
+WHERE name = 'commands' AND type = 'table';
+
+PRAGMA writable_schema=RESET;
+  |]

src/Simplex/Messaging/Agent/Store/SQLite/Migrations/M20261008_rcv_switch_status.hs

@@ -0,0 +1,19 @@
+{-# LANGUAGE QuasiQuotes #-}
+
+module Simplex.Messaging.Agent.Store.SQLite.Migrations.M20261008_rcv_switch_status where
+
+import Database.SQLite.Simple (Query)
+import Database.SQLite.Simple.QQ (sql)
+
+m20261008_rcv_switch_status :: Query
+m20261008_rcv_switch_status =
+  [sql|
+CREATE INDEX idx_rcv_queues_switch_status ON rcv_queues(switch_status);
+|]
+
+down_m20261008_rcv_switch_status :: Query
+down_m20261008_rcv_switch_status =
+  [sql|
+UPDATE rcv_queues SET switch_status = NULL WHERE switch_status = 'received_qend';
+DROP INDEX idx_rcv_queues_switch_status;
+|]

src/Simplex/Messaging/Agent/Store/SQLite/Migrations/agent_schema.sql

@@ -182,7 +182,9 @@ CREATE TABLE ratchets(
   x3dh_pub_key_1 BLOB,
   x3dh_pub_key_2 BLOB,
   pq_priv_kem BLOB,
-  pq_pub_kem BLOB
+  pq_pub_kem BLOB,
+  rc_verify_code_ad BLOB,
+  rc_verify_code_pq BLOB
 ) WITHOUT ROWID, STRICT;
 CREATE TABLE skipped_messages(
   skipped_message_id INTEGER PRIMARY KEY,
@@ -245,7 +247,7 @@ CREATE TABLE ntf_subscriptions(
   ON DELETE RESTRICT ON UPDATE CASCADE
 ) WITHOUT ROWID, STRICT;
 CREATE TABLE commands(
-  command_id INTEGER PRIMARY KEY,
+  command_id INTEGER PRIMARY KEY AUTOINCREMENT,
   conn_id BLOB NOT NULL REFERENCES connections ON DELETE CASCADE,
   host TEXT,
   port TEXT,
@@ -259,6 +261,7 @@ CREATE TABLE commands(
   FOREIGN KEY(host, port) REFERENCES servers
   ON DELETE RESTRICT ON UPDATE CASCADE
 ) STRICT;
+CREATE TABLE sqlite_sequence(name,seq);
 CREATE TABLE snd_message_deliveries(
   snd_message_delivery_id INTEGER PRIMARY KEY AUTOINCREMENT,
   conn_id BLOB NOT NULL REFERENCES connections ON DELETE CASCADE,
@@ -267,7 +270,6 @@ CREATE TABLE snd_message_deliveries(
   failed INTEGER DEFAULT 0,
   FOREIGN KEY(conn_id, internal_id) REFERENCES messages ON DELETE CASCADE DEFERRABLE INITIALLY DEFERRED
 ) STRICT;
-CREATE TABLE sqlite_sequence(name,seq);
 CREATE TABLE users(
   user_id INTEGER PRIMARY KEY AUTOINCREMENT
   ,
@@ -571,10 +573,6 @@ CREATE INDEX idx_encrypted_rcv_message_hashes_hash ON encrypted_rcv_message_hash
   conn_id,
   hash
 );
-CREATE INDEX idx_processed_ratchet_key_hashes_hash ON processed_ratchet_key_hashes(
-  conn_id,
-  hash
-);
 CREATE INDEX idx_snd_messages_rcpt_internal_id ON snd_messages(
   conn_id,
   rcpt_internal_id
@@ -639,6 +637,15 @@ CREATE INDEX idx_connections_deleted ON connections(deleted);
 CREATE INDEX idx_connections_service_request_expires_at ON connections(
   service_request_expires_at
 );
+CREATE UNIQUE INDEX idx_processed_ratchet_key_hashes_hash ON processed_ratchet_key_hashes(
+  conn_id,
+  hash
+);
+CREATE INDEX idx_processed_ratchet_key_hashes_conn_id ON processed_ratchet_key_hashes(
+  conn_id,
+  processed_ratchet_key_hash_id
+);
+CREATE INDEX idx_rcv_queues_switch_status ON rcv_queues(switch_status);
 CREATE TRIGGER tr_rcv_queue_insert
 AFTER INSERT ON rcv_queues
 FOR EACH ROW

src/Simplex/Messaging/Client.hs

@@ -67,6 +67,7 @@ module Simplex.Messaging.Client
     enableSMPQueuesNtfs,
     disableSMPQueuesNtfs,
     sendSMPMessage,
+    QueueDrained,
     ackSMPMessage,
     suspendSMPQueue,
     deleteSMPQueue,
@@ -166,7 +167,7 @@ import Simplex.Messaging.Parsers (defaultJSON, dropPrefix, enumJSON, sumTypeJSON
 import Simplex.Messaging.Protocol
 import Simplex.Messaging.Protocol.Types
 import Simplex.Messaging.Server.QueueStore.QueueInfo
-import Simplex.Messaging.SimplexName (SimplexDomain)
+import Simplex.Messaging.SimplexName (SimplexDomain, fullDomainName)
 import Simplex.Messaging.TMap (TMap)
 import qualified Simplex.Messaging.TMap as TM
 import Simplex.Messaging.Transport
@@ -195,6 +196,7 @@ data PClient v err msg = PClient
     transportHost :: TransportHost,
     tcpConnectTimeout :: NetworkTimeout,
     tcpTimeout :: NetworkTimeout,
+    proxiedRelayVRange :: VersionRange v,
     sendPings :: TVar Bool,
     lastReceived :: TVar UTCTime,
     timeoutErrorCount :: TVar Int,
@@ -240,6 +242,7 @@ smpClientStub g sessionId thVersion thAuth = do
               transportHost = "localhost",
               tcpConnectTimeout,
               tcpTimeout,
+              proxiedRelayVRange = supportedClientSMPRelayVRange,
               sendPings,
               lastReceived,
               timeoutErrorCount,
@@ -477,6 +480,7 @@ data ProtocolClientConfig v = ProtocolClientConfig
     serviceCredentials :: Maybe ServiceCredentials,
     -- | client-server protocol version range
     serverVRange :: VersionRange v,
+    proxiedRelayVRange :: VersionRange v,
     -- | agree shared session secret (used in SMP proxy for additional encryption layer)
     agreeSecret :: Bool,
     -- | Whether connecting client is a proxy server. See comment in ClientHandshake
@@ -495,6 +499,7 @@ defaultClientConfig clientALPN useSNI serverVRange =
       clientALPN,
       serviceCredentials = Nothing,
       serverVRange,
+      proxiedRelayVRange = serverVRange,
       agreeSecret = False,
       proxyServer = False,
       useSNI
@@ -505,6 +510,7 @@ defaultSMPClientConfig :: ProtocolClientConfig SMPVersion
 defaultSMPClientConfig =
   (defaultClientConfig (Just alpnSupportedSMPHandshakes) False supportedClientSMPRelayVRange)
     { defaultTransport = (show defaultSMPPort, transport @TLS),
+      proxiedRelayVRange = supportedClientSMPRelayVRange,
       agreeSecret = True
     }
 {-# INLINE defaultSMPClientConfig #-}
@@ -568,7 +574,7 @@ type SMPTransportSession = TransportSession BrokerMsg
 -- A single queue can be used for multiple 'SMPClient' instances,
 -- as 'SMPServerTransmission' includes server information.
 getProtocolClient :: forall v err msg. Protocol v err msg => TVar ChaChaDRG -> NetworkRequestMode -> TransportSession msg -> ProtocolClientConfig v -> [HostName] -> Maybe (TBQueue (ServerTransmissionBatch v err msg)) -> UTCTime -> (ProtocolClient v err msg -> IO ()) -> IO (Either (ProtocolClientError err) (ProtocolClient v err msg))
-getProtocolClient g nm transportSession@(_, srv, _) cfg@ProtocolClientConfig {qSize, networkConfig, clientALPN, serviceCredentials, serverVRange, agreeSecret, proxyServer, useSNI} presetDomains msgQ proxySessTs disconnected = do
+getProtocolClient g nm transportSession@(_, srv, _) cfg@ProtocolClientConfig {qSize, networkConfig, clientALPN, serviceCredentials, serverVRange, proxiedRelayVRange, agreeSecret, proxyServer, useSNI} presetDomains msgQ proxySessTs disconnected = do
   case chooseTransportHost networkConfig (host srv) of
     Right useHost ->
       (getCurrentTime >>= mkProtocolClient useHost >>= runClient useTransport useHost)
@@ -593,6 +599,7 @@ getProtocolClient g nm transportSession@(_, srv, _) cfg@ProtocolClientConfig {qS
             transportHost,
             tcpConnectTimeout,
             tcpTimeout,
+            proxiedRelayVRange,
             sendPings,
             lastReceived,
             timeoutErrorCount,
@@ -1054,11 +1061,11 @@ proxySMPMessage c nm proxiedRelay spKey sId flags msg = proxyOKSMPCommand c nm p
 -- through `proxySMPCommand` and pattern-matches the expected RNAME response.
 -- Version-gated on the destination relay (mirrors `connectSMPProxiedRelay`):
 -- the client never sends RSLV to a relay that predates names support.
-proxyResolveName :: SMPClient -> NetworkRequestMode -> ProxiedRelay -> SimplexDomain -> ExceptT SMPClientError IO (Either ProxyClientError NameRecord)
+proxyResolveName :: SMPClient -> NetworkRequestMode -> ProxiedRelay -> SimplexDomain -> ExceptT SMPClientError IO (Either ProxyClientError NameResponse)
 proxyResolveName c nm proxiedRelay name
   | prVersion proxiedRelay >= namesSMPVersion =
-      proxySMPCommand c nm proxiedRelay Nothing NoEntity (RSLV name) >>= \case
-        Right (RNAME nr) -> pure $ Right nr
+      proxySMPCommand c nm proxiedRelay Nothing NoEntity (RSLV (NQDomain name)) >>= \case
+        Right (RNAME reg) | resolvedNameOrNotFound name reg -> pure $ Right reg
         Right r -> throwE $ unexpectedResponse r
         Left e -> pure $ Left e
   | otherwise = throwE $ PCETransportError TEVersion
@@ -1066,25 +1073,32 @@ proxyResolveName c nm proxiedRelay name
 -- | Direct (non-PFWD) name resolution. Exposes the client IP to the resolver;
 -- callers that want anonymity should use `proxyResolveName` via the standard
 -- proxy fallback in the agent. RSLV requires no entity ID or authorization
--- (see `noAuthCmd` in Protocol.hs). Version-gated on the session here, not the
--- encoder, so an old server never receives RSLV.
-directResolveName :: SMPClient -> NetworkRequestMode -> SimplexDomain -> ExceptT SMPClientError IO NameRecord
+-- (see `noAuthCmd` in Protocol.hs). Gated on the session version, below which
+-- the server has no RSLV at all; the encoder gates the query format separately.
+directResolveName :: SMPClient -> NetworkRequestMode -> SimplexDomain -> ExceptT SMPClientError IO NameResponse
 directResolveName c nm name
   | thVersion (thParams c) >= namesSMPVersion =
-      sendProtocolCommand c nm Nothing NoEntity (Cmd SResolver (RSLV name)) >>= \case
-        RNAME nr -> pure nr
+      sendProtocolCommand c nm Nothing NoEntity (Cmd SResolver (RSLV (NQDomain name))) >>= \case
+        RNAME reg | resolvedNameOrNotFound name reg -> pure reg
         r -> throwE $ unexpectedResponse r
   | otherwise = throwE $ PCETransportError TEVersion
 
+resolvedNameOrNotFound :: SimplexDomain -> NameResponse -> Bool
+resolvedNameOrNotFound d NameResponse {registration} = case registration of
+  NRRegistered {nameRecord} -> T.toLower (nrName nameRecord) == fullDomainName d
+  _ -> True
+
+type QueueDrained = Bool
+
 -- | Acknowledge message delivery (server deletes the message).
 --
 -- https://github.com/simplex-chat/simplexmq/blob/master/protocol/simplex-messaging.md#acknowledge-message-delivery
 -- This command is always sent in background request mode
-ackSMPMessage :: SMPClient -> RcvPrivateAuthKey -> QueueId -> MsgId -> ExceptT SMPClientError IO ()
+ackSMPMessage :: SMPClient -> RcvPrivateAuthKey -> QueueId -> MsgId -> ExceptT SMPClientError IO QueueDrained
 ackSMPMessage c rpKey rId msgId =
   sendSMPCommand c NRMBackground (Just rpKey) rId (ACK msgId) >>= \case
-    OK -> return ()
-    cmd@MSG {} -> liftIO $ writeSMPMessage c rId cmd
+    OK -> pure True
+    cmd@MSG {} -> liftIO (writeSMPMessage c rId cmd) $> False
     r -> throwE $ unexpectedResponse r
 
 -- | Irreversibly suspend SMP queue.
@@ -1110,10 +1124,10 @@ deleteSMPQueues = okSMPCommands DEL
 -- send PRXY :: SMPServer -> Maybe BasicAuth -> Command Sender
 -- receives PKEY :: SessionId -> X.CertificateChain -> X.SignedExact X.PubKey -> BrokerMsg
 connectSMPProxiedRelay :: SMPClient -> NetworkRequestMode -> SMPServer -> Maybe BasicAuth -> ExceptT SMPClientError IO ProxiedRelay
-connectSMPProxiedRelay c@ProtocolClient {client_ = PClient {tcpConnectTimeout, tcpTimeout}} nm relayServ@ProtocolServer {port = relayPort, keyHash = C.KeyHash kh} proxyAuth =
+connectSMPProxiedRelay c@ProtocolClient {client_ = PClient {tcpConnectTimeout, tcpTimeout, proxiedRelayVRange}} nm relayServ@ProtocolServer {port = relayPort, keyHash = C.KeyHash kh} proxyAuth =
   sendProtocolCommand_ c nm Nothing tOut Nothing NoEntity (Cmd SProxiedClient (PRXY relayServ proxyAuth)) >>= \case
     PKEY sId vr (CertChainPubKey chain key) ->
-      case supportedClientSMPRelayVRange `compatibleVersion` vr of
+      case proxiedRelayVRange `compatibleVersion` vr of
         Nothing -> throwE $ transportErr TEVersion
         Just (Compatible v) -> do
           relayKey <- liftEitherWith (const $ transportErr $ TEHandshake IDENTITY) =<< liftIO (runExceptT $ validateRelay chain key)
@@ -1164,7 +1178,7 @@ instance StrEncoding ProxyClientError where
 -- consider how to process slow responses - is it handled somehow locally or delegated to the caller
 -- this method is used in the client
 -- sends PFWD :: C.PublicKeyX25519 -> EncTransmission -> Command Sender
--- receives PRES :: EncResponse -> BrokerMsg -- proxy to client
+-- receives PRES :: Maybe C.CbNonce -> EncResponse -> BrokerMsg -- proxy to client
 
 -- When client sends message via proxy, there may be one successful scenario and 9 error scenarios
 -- as shown below (WTF stands for unexpected response, ??? for response that failed to parse).
@@ -1217,20 +1231,20 @@ proxySMPCommand c@ProtocolClient {thParams = proxyThParams, client_ = PClient {c
   -- encode
   let TransmissionForAuth {tForAuth, tToSend} = encodeTransmissionForAuth serverThParams (CorrId corrId, sId, Cmd (sParty @p) command)
   -- serviceAuth is False here – proxied commands are not used with service certificates
-  auth <- liftEitherWith PCETransportError $ authTransmission serverThAuth False spKey nonce tForAuth
+  auth <- liftEitherWith PCETransportError $ authTransmission serverThAuth False spKey (Just nonce) tForAuth
   b <- case batchTransmissions serverThParams [Right (auth, tToSend)] of
     [] -> throwE $ PCETransportError TELargeMsg
     TBError e _ : _ -> throwE $ PCETransportError e
     TBTransmission s _ : _ -> pure s
     TBTransmissions s _ _ : _ -> pure s
-  et <- liftEitherWith PCECryptoError $ EncTransmission <$> C.cbEncrypt cmdSecret nonce b paddedProxiedTLength
+  et <- liftEitherWith PCECryptoError $ EncTransmission <$> C.cbEncrypt cmdSecret (encTransmissionNonce v nonce) b paddedProxiedTLength
   -- proxy interaction errors are wrapped
   let tOut = Just $ 2 * netTimeoutInt tcpTimeout nm
   tryE (sendProtocolCommand_ c nm (Just nonce) tOut Nothing (EntityId sessionId) (Cmd SProxiedClient (PFWD v cmdPubKey et))) >>= \case
     Right r -> case r of
-      PRES (EncResponse er) -> do
+      PRES nonce_ (EncResponse er) -> do
         -- server interaction errors are thrown directly
-        t' <- liftEitherWith PCECryptoError $ C.cbDecrypt cmdSecret (C.reverseNonce nonce) er
+        t' <- liftEitherWith PCECryptoError $ C.cbDecrypt cmdSecret (fromMaybe (C.reverseNonce nonce) nonce_) er
         case tParse serverThParams t' of
           t'' :| [] -> case tDecodeClient serverThParams t'' of
             (_, _, cmd) -> case cmd of
@@ -1248,10 +1262,10 @@ proxySMPCommand c@ProtocolClient {thParams = proxyThParams, client_ = PClient {c
 
 -- this method is used in the proxy
 -- sends RFWD :: EncFwdTransmission -> Command Sender
--- receives RRES :: EncFwdResponse -> BrokerMsg
+-- receives RRES :: Maybe C.CbNonce -> EncFwdResponse -> BrokerMsg
 -- proxy should send PRES to the client with EncResponse
 -- Always uses background timeout mode
-forwardSMPTransmission :: SMPClient -> CorrId -> VersionSMP -> C.PublicKeyX25519 -> EncTransmission -> ExceptT SMPClientError IO EncResponse
+forwardSMPTransmission :: SMPClient -> C.CorrCbNonce -> VersionSMP -> C.PublicKeyX25519 -> EncTransmission -> ExceptT SMPClientError IO (Maybe C.CbNonce, EncResponse)
 forwardSMPTransmission c@ProtocolClient {thParams, client_ = PClient {clientCorrId = g}} fwdCorrId fwdVersion fwdKey fwdTransmission = do
   -- prepare params
   sessSecret <- case thAuth thParams of
@@ -1263,11 +1277,11 @@ forwardSMPTransmission c@ProtocolClient {thParams, client_ = PClient {clientCorr
       eft = EncFwdTransmission $ C.cbEncryptNoPad sessSecret nonce (smpEncode fwdT)
   -- send
   sendProtocolCommand_ c NRMBackground (Just nonce) Nothing Nothing NoEntity (Cmd SProxyService (RFWD eft)) >>= \case
-    RRES (EncFwdResponse efr) -> do
+    RRES nonce_ (EncFwdResponse efr) -> do
       -- unwrap
       r' <- liftEitherWith PCECryptoError $ C.cbDecryptNoPad sessSecret (C.reverseNonce nonce) efr
       FwdResponse {fwdCorrId = _, fwdResponse} <- liftEitherWith (const $ PCEResponseError BLOCK) $ smpDecode r'
-      pure fwdResponse
+      pure (nonce_, fwdResponse)
     r -> throwE $ unexpectedResponse r
 
 -- get queue information - always sent interactively
@@ -1389,7 +1403,7 @@ mkTransmission_ :: forall v err msg. Protocol v err msg => ProtocolClient v err
 mkTransmission_ ProtocolClient {thParams, client_ = PClient {clientCorrId, sentCommands}} nonce_ (entityId, pKey_, command) = do
   nonce@(C.CbNonce corrId) <- maybe (atomically $ C.randomCbNonce clientCorrId) pure nonce_
   let TransmissionForAuth {tForAuth, tToSend} = encodeTransmissionForAuth thParams (CorrId corrId, entityId, command)
-      auth = authTransmission (thAuth thParams) (useServiceAuth command) pKey_ nonce tForAuth
+      auth = authTransmission (thAuth thParams) (useServiceAuth command) pKey_ (Just nonce) tForAuth
   r <- mkRequest (CorrId corrId)
   pure ((,tToSend) <$> auth, r)
   where
@@ -1408,14 +1422,14 @@ mkTransmission_ ProtocolClient {thParams, client_ = PClient {clientCorrId, sentC
       atomically $ TM.insert corrId r sentCommands
       pure r
 
-authTransmission :: Maybe (THandleAuth 'TClient) -> Bool -> Maybe C.APrivateAuthKey -> C.CbNonce -> ByteString -> Either TransportError (Maybe TAuthorizations)
-authTransmission thAuth serviceAuth pKey_ nonce t = traverse authenticate pKey_
+authTransmission :: Maybe (THandleAuth 'TClient) -> Bool -> Maybe C.APrivateAuthKey -> Maybe C.CbNonce -> ByteString -> Either TransportError (Maybe TAuthorizations)
+authTransmission thAuth serviceAuth pKey_ nonce_ t = traverse authenticate pKey_
   where
     authenticate :: C.APrivateAuthKey -> Either TransportError TAuthorizations
     authenticate (C.APrivateAuthKey a pk) = (,serviceSig) <$> case a of
-      C.SX25519 -> case thAuth of
-        Just THAuthClient {peerServerPubKey = k} -> Right $ TAAuthenticator $ C.cbAuthenticate k pk nonce t'
-        Nothing -> Left TENoServerAuth
+      C.SX25519 -> case (thAuth, nonce_) of
+        (Just THAuthClient {peerServerPubKey = k}, Just nonce) -> Right $ TAAuthenticator $ C.cbAuthenticate k pk nonce t'
+        _ -> Left TENoServerAuth
       C.SEd25519 -> sign pk
       C.SEd448 -> sign pk
     -- When command is signed by both entity key and service key,

src/Simplex/Messaging/Crypto.hs

@@ -140,8 +140,10 @@ module Simplex.Messaging.Crypto
     gcmIV,
 
     -- * NaCl crypto_box
-    CbNonce (unCbNonce),
+    CbNonceI (..),
+    CbNonce,
     pattern CbNonce,
+    CorrCbNonce,
     cbEncrypt,
     cbEncryptNoPad,
     cbEncryptMaxLenBS,
@@ -152,8 +154,12 @@ module Simplex.Messaging.Crypto
     sbEncryptNoPad,
     sbDecryptNoPad,
     cbNonce,
+    unsafeCbNonce,
+    corrCbNonce,
     randomCbNonce,
     reverseNonce,
+    xorNonce,
+    toCbNonce,
 
     -- * NaCl crypto_secretbox
     SbKey (unSbKey),
@@ -232,6 +238,7 @@ import Data.ASN1.Types
 import Data.Aeson (FromJSON (..), ToJSON (..))
 import qualified Data.Attoparsec.ByteString.Char8 as A
 import Data.Bifunctor (bimap, first)
+import Data.Bits (xor)
 import Data.ByteArray (ByteArrayAccess)
 import qualified Data.ByteArray as BA
 import Data.ByteString.Base64 (decode)
@@ -254,7 +261,7 @@ import Simplex.Messaging.Agent.Store.DB (Binary (..), FromField (..), ToField (.
 import Simplex.Messaging.Encoding
 import Simplex.Messaging.Encoding.String
 import Simplex.Messaging.Parsers (parseAll, parseString)
-import Simplex.Messaging.Util ((<$?>))
+import Simplex.Messaging.Util (packZipWith, (<$?>))
 
 -- | Cryptographic algorithms.
 data Algorithm = Ed25519 | Ed448 | X25519 | X448
@@ -931,6 +938,8 @@ data CryptoError
     CERatchetEarlierMessage Word32
   | -- | duplicate message number
     CERatchetDuplicateMessage
+  | -- | KEM key generation failed, indicating a broken RNG
+    CryptoKEMKeyGenError
   deriving (Eq, Show, Exception)
 
 aesKeySize :: Int
@@ -1045,9 +1054,7 @@ md5Hash = BA.convert . (hash :: ByteString -> Digest MD5)
 
 -- | AEAD-GCM encryption with associated data.
 --
--- Used as part of double ratchet encryption.
--- This function requires 16 bytes IV, it transforms IV in cryptonite_aes_gcm_init here:
--- https://github.com/haskell-crypto/cryptonite/blob/master/cbits/cryptonite_aes.c
+-- Used as part of double ratchet encryption, with a 16-byte IV (see @initAEAD@).
 encryptAEAD :: Key -> IV -> Int -> ByteString -> ByteString -> ExceptT CryptoError IO (AuthTag, ByteString)
 encryptAEAD aesKey ivBytes paddedLen ad msg = do
   aead <- initAEAD @AES256 aesKey ivBytes
@@ -1067,10 +1074,7 @@ encryptAEADNoPad aesKey ivBytes ad msg = do
 
 -- | AEAD-GCM decryption with associated data.
 --
--- Used as part of double ratchet encryption.
--- This function requires 16 bytes IV, it transforms IV in cryptonite_aes_gcm_init here:
--- https://github.com/haskell-crypto/cryptonite/blob/master/cbits/cryptonite_aes.c
--- To make it compatible with WebCrypto we will need to start using initAEADGCM.
+-- Used as part of double ratchet encryption, with a 16-byte IV (see @initAEAD@).
 decryptAEAD :: Key -> IV -> ByteString -> ByteString -> AuthTag -> ExceptT CryptoError IO ByteString
 decryptAEAD aesKey ivBytes ad msg (AuthTag authTag) = do
   aead <- initAEAD @AES256 aesKey ivBytes
@@ -1148,9 +1152,9 @@ maxLength :: forall i. KnownNat i => Int
 maxLength = fromIntegral (natVal $ Proxy @i)
 {-# INLINE maxLength #-}
 
--- this function requires 16 bytes IV, it transforms IV in cryptonite_aes_gcm_init here:
--- https://github.com/haskell-crypto/cryptonite/blob/master/cbits/cryptonite_aes.c
--- This is used for double ratchet encryption, so to make it compatible with WebCrypto we will need to deprecate it and start using initAEADGCM
+-- The 16-byte double ratchet IV is intentionally not the 96-bit IV recommended by NIST SP 800-38D, so GCM derives J0 = GHASH(IV || 0^64 || [128]_64),
+-- as in crypton_aes_gcm_init: https://hackage.haskell.org/package/crypton-0.34/src/cbits/crypton_aes.c
+-- WebCrypto and other SP 800-38D implementations interoperate only when given all 16 IV bytes.
 initAEAD :: forall c. AES.BlockCipher c => Key -> IV -> ExceptT CryptoError IO (AES.AEAD c)
 initAEAD (Key aesKey) (IV ivBytes) = do
   iv <- makeIV @c ivBytes
@@ -1327,8 +1331,8 @@ sbEncryptNoPad (SbKey key) (CbNonce nonce) = cryptoBox key nonce
 {-# INLINE sbEncryptNoPad #-}
 
 -- | NaCl @crypto_box@ encrypt with a shared DH secret and 192-bit nonce.
-cbEncryptMaxLenBS :: KnownNat i => DhSecret X25519 -> CbNonce -> MaxLenBS i -> ByteString
-cbEncryptMaxLenBS (DhSecretX25519 secret) (CbNonce nonce) = cryptoBox secret nonce . unMaxLenBS . padMaxLenBS
+cbEncryptMaxLenBS :: (KnownNat i, CbNonceI n) => DhSecret X25519 -> n -> MaxLenBS i -> ByteString
+cbEncryptMaxLenBS (DhSecretX25519 secret) nonce = cryptoBox secret (unCbNonce nonce) . unMaxLenBS . padMaxLenBS
 {-# INLINE cbEncryptMaxLenBS #-}
 
 cryptoBox :: ByteArrayAccess key => key -> ByteString -> ByteString -> ByteString
@@ -1338,8 +1342,8 @@ cryptoBox secret nonce s = BA.convert tag <> c
     tag = Poly1305.auth rs c
 
 -- | NaCl @crypto_box@ decrypt with a shared DH secret and 192-bit nonce.
-cbDecrypt :: DhSecret X25519 -> CbNonce -> ByteString -> Either CryptoError ByteString
-cbDecrypt (DhSecretX25519 secret) = sbDecrypt_ secret
+cbDecrypt :: CbNonceI n => DhSecret X25519 -> n -> ByteString -> Either CryptoError ByteString
+cbDecrypt (DhSecretX25519 secret) = sbDecrypt_ secret . toCbNonce
 {-# INLINE cbDecrypt #-}
 
 -- | NaCl @crypto_box@ decrypt with a shared DH secret and 192-bit nonce (without unpadding).
@@ -1386,9 +1390,16 @@ cbAuthenticate k pk nonce msg = CbAuthenticator $ cbEncryptNoPad (dh' k pk) nonc
 cbVerify :: PublicKeyX25519 -> PrivateKeyX25519 -> CbNonce -> CbAuthenticator -> ByteString -> Bool
 cbVerify k pk nonce (CbAuthenticator s) authorized = cbDecryptNoPad (dh' k pk) nonce s == Right (sha512Hash authorized)
 
-newtype CbNonce = CryptoBoxNonce {unCbNonce :: ByteString}
+class CbNonceI n where
+  unCbNonce :: n -> ByteString
+
+newtype CbNonce = CryptoBoxNonce ByteString
   deriving (Eq, Show)
-  deriving newtype (FromField)
+
+instance CbNonceI CbNonce where
+  unCbNonce (CryptoBoxNonce s) = s
+
+instance FromField CbNonce where fromField = blobFieldDecoder cbNonce
 
 instance ToField CbNonce where toField (CryptoBoxNonce s) = toField $ Binary s
 
@@ -1399,7 +1410,7 @@ pattern CbNonce s <- CryptoBoxNonce s
 
 instance StrEncoding CbNonce where
   strEncode (CbNonce s) = strEncode s
-  strP = cbNonce <$> strP
+  strP = cbNonce <$?> strP
 
 instance ToJSON CbNonce where
   toJSON = strToJSON
@@ -1408,13 +1419,16 @@ instance ToJSON CbNonce where
 instance FromJSON CbNonce where
   parseJSON = strParseJSON "CbNonce"
 
-cbNonce :: ByteString -> CbNonce
+cbNonce :: ByteString -> Either String CbNonce
 cbNonce s
-  | len == 24 = CryptoBoxNonce s
-  | len > 24 = CryptoBoxNonce . fst $ B.splitAt 24 s
-  | otherwise = CryptoBoxNonce $ s <> B.replicate (24 - len) (toEnum 0)
-  where
-    len = B.length s
+  | B.length s == 24 = Right $ CryptoBoxNonce s
+  | otherwise = Left "CbNonce: invalid length"
+
+unsafeCbNonce :: ByteString -> CbNonce
+unsafeCbNonce s = either error id $ cbNonce s
+
+corrCbNonce :: ByteString -> Either String CorrCbNonce
+corrCbNonce s = CorrCbNonce . unCbNonce <$> cbNonce s
 
 randomCbNonce :: TVar ChaChaDRG -> STM CbNonce
 randomCbNonce = fmap CryptoBoxNonce . randomBytes 24
@@ -1422,13 +1436,30 @@ randomCbNonce = fmap CryptoBoxNonce . randomBytes 24
 randomBytes :: Int -> TVar ChaChaDRG -> STM ByteString
 randomBytes n gVar = stateTVar gVar $ randomBytesGenerate n
 
-reverseNonce :: CbNonce -> CbNonce
-reverseNonce (CryptoBoxNonce s) = CryptoBoxNonce (B.reverse s)
+reverseNonce :: CbNonceI n => n -> CbNonce
+reverseNonce = CryptoBoxNonce . B.reverse . unCbNonce
+
+xorNonce :: CbNonceI n => ByteString -> n -> CbNonce
+xorNonce s nonce = CryptoBoxNonce $ packZipWith xor s n <> B.drop (B.length s) n
+  where
+    n = unCbNonce nonce
+
+toCbNonce :: CbNonceI n => n -> CbNonce
+toCbNonce = CryptoBoxNonce . unCbNonce
 
 instance Encoding CbNonce where
   smpEncode = unCbNonce
   smpP = CryptoBoxNonce <$> A.take 24
 
+newtype CorrCbNonce = CorrCbNonce ByteString
+
+instance CbNonceI CorrCbNonce where
+  unCbNonce (CorrCbNonce s) = s
+
+instance Encoding CorrCbNonce where
+  smpEncode = smpEncode . unCbNonce
+  smpP = corrCbNonce <$?> smpP
+
 newtype SbKey = SecretBoxKey {unSbKey :: ByteString}
   deriving (Eq, Show)
   deriving newtype (FromField)

src/Simplex/Messaging/Crypto/Ratchet.hs

@@ -72,6 +72,7 @@ module Simplex.Messaging.Crypto.Ratchet
     rcEncryptHeader,
     rcEncryptMsg,
     rcDecrypt,
+    maxSkippedMsgKeys,
     -- used in tests
     MsgHeader (..),
     RatchetInitParams (..),
@@ -119,7 +120,7 @@ import Simplex.Messaging.Crypto.SNTRUP761.Bindings
 import Simplex.Messaging.Encoding
 import Simplex.Messaging.Encoding.String
 import Simplex.Messaging.Parsers (defaultJSON, parseE, parseE')
-import Simplex.Messaging.Util (($>>=), (<$?>))
+import Simplex.Messaging.Util ((<$?>))
 import Simplex.Messaging.Version
 import Simplex.Messaging.Version.Internal
 import UnliftIO.STM
@@ -327,9 +328,13 @@ instance StrEncoding AnyE2ERatchetParamsUri where
             Nothing -> pure $ AnyE2ERatchetParamsUri SRKSProposed a $ E2ERatchetParamsUri vr k1 k2 Nothing
       _ -> fail "bad e2e params"
     where
-      kemP query =
-        queryParam_ "kem_key" query
-          $>>= \k -> Just . kemParams k <$> queryParam_ "kem_ct" query
+      kemP query = do
+        k_ <- queryParam_ "kem_key" query
+        ct_ <- queryParam_ "kem_ct" query
+        case (k_, ct_) of
+          (Just k, _) -> pure $ Just $ kemParams k ct_
+          (Nothing, Nothing) -> pure Nothing
+          (Nothing, Just _) -> fail "bad e2e params: kem_ct without kem_key"
       kemParams k = \case
         Nothing -> ARKP SRKSProposed $ RKParamsProposed k
         Just ct -> ARKP SRKSAccepted $ RKParamsAccepted ct k
@@ -451,6 +456,7 @@ generateSndE2EParams g v = \case
 
 data RatchetInitParams = RatchetInitParams
   { assocData :: Str,
+    rcVerifyCodePQ :: Str,
     ratchetKey :: RatchetKey,
     sndHK :: HeaderKey,
     rcvNextHK :: HeaderKey,
@@ -493,14 +499,14 @@ pqX3dhRcv (rpk1, rpk2, rpKem_) (E2ERatchetParams _ sk1 sk2 sKem_) = do
 
 pqX3dh :: DhAlgorithm a => (PublicKey a, PublicKey a) -> DhSecret a -> DhSecret a -> DhSecret a -> Maybe RatchetKEMAccepted -> RatchetInitParams
 pqX3dh (sk1, rk1) dh1 dh2 dh3 kemAccepted =
-  RatchetInitParams {assocData, ratchetKey = RatchetKey sk, sndHK = Key hk, rcvNextHK = Key nhk, kemAccepted}
+  RatchetInitParams {assocData, rcVerifyCodePQ = Str vcPQ, ratchetKey = RatchetKey sk, sndHK = Key hk, rcvNextHK = Key nhk, kemAccepted}
   where
     assocData = Str $ pubKeyBytes sk1 <> pubKeyBytes rk1
     dhs = dhBytes' dh1 <> dhBytes' dh2 <> dhBytes' dh3 <> pq
     pq = maybe "" (\RatchetKEMAccepted {rcPQRss = KEMSharedKey ss} -> BA.convert ss) kemAccepted
-    (hk, nhk, sk) =
-      let salt = B.replicate 64 '\0'
-       in hkdf3 salt dhs "SimpleXX3DH"
+    salt = B.replicate 64 '\0'
+    (hk, nhk, sk) = hkdf3 salt dhs "SimpleXX3DH"
+    vcPQ = hkdf salt dhs "SimpleXVerifyCode" 32
 
 type RatchetX448 = Ratchet 'X448
 
@@ -509,6 +515,8 @@ data Ratchet a = Ratchet
     rcVersion :: RatchetVersions,
     -- associated data - must be the same in both parties ratchets
     rcAD :: Str,
+    -- verification code covering all handshake keys, absent in the ratchets created before it was added
+    rcVCPQ :: Maybe Str,
     rcDHRs :: PrivateKey a,
     rcKEM :: Maybe RatchetKEM,
     rcSupportKEM :: PQSupport, -- defines header size, can only be enabled once
@@ -637,13 +645,14 @@ instance FromField MessageKey where fromField = blobFieldDecoder smpDecode
 -- @
 initSndRatchet ::
   forall a. (AlgorithmI a, DhAlgorithm a) => RatchetVersions -> PublicKey a -> PrivateKey a -> (RatchetInitParams, Maybe KEMKeyPair) -> Ratchet a
-initSndRatchet rcVersion rcDHRr rcDHRs (RatchetInitParams {assocData, ratchetKey, sndHK, rcvNextHK, kemAccepted}, rcPQRs_) = do
+initSndRatchet rcVersion rcDHRr rcDHRs (RatchetInitParams {assocData, rcVerifyCodePQ, ratchetKey, sndHK, rcvNextHK, kemAccepted}, rcPQRs_) = do
   -- state.RK, state.CKs, state.NHKs = KDF_RK_HE(SK, DH(state.DHRs, state.DHRr) || state.PQRss)
   let (rcRK, rcCKs, rcNHKs) = rootKdf ratchetKey rcDHRr rcDHRs (rcPQRss <$> kemAccepted)
       pqOn = isJust rcPQRs_
    in Ratchet
         { rcVersion,
           rcAD = assocData,
+          rcVCPQ = Just rcVerifyCodePQ,
           rcDHRs,
           rcKEM = (`RatchetKEM` kemAccepted) <$> rcPQRs_,
           rcSupportKEM = PQSupport pqOn,
@@ -668,10 +677,11 @@ initSndRatchet rcVersion rcDHRr rcDHRs (RatchetInitParams {assocData, ratchetKey
 -- as part of the connection request and random salt was received from the sender.
 initRcvRatchet ::
   forall a. (AlgorithmI a, DhAlgorithm a) => RatchetVersions -> PrivateKey a -> (RatchetInitParams, Maybe KEMKeyPair) -> PQSupport -> Ratchet a
-initRcvRatchet rcVersion rcDHRs (RatchetInitParams {assocData, ratchetKey, sndHK, rcvNextHK, kemAccepted}, rcPQRs_) pqSupport =
+initRcvRatchet rcVersion rcDHRs (RatchetInitParams {assocData, rcVerifyCodePQ, ratchetKey, sndHK, rcvNextHK, kemAccepted}, rcPQRs_) pqSupport =
   Ratchet
     { rcVersion,
       rcAD = assocData,
+      rcVCPQ = Just rcVerifyCodePQ,
       rcDHRs,
       -- rcKEM:
       -- state.PQRs = bob_pq_kem_key_pair
@@ -953,6 +963,9 @@ type DecryptResult a = (Either CryptoError ByteString, Ratchet a, SkippedMsgDiff
 maxSkip :: Word32
 maxSkip = 512
 
+maxSkippedMsgKeys :: Int
+maxSkippedMsgKeys = 2000
+
 rcDecrypt ::
   forall a.
   (AlgorithmI a, DhAlgorithm a) =>

src/Simplex/Messaging/Crypto/SNTRUP761/Bindings.hs

@@ -16,6 +16,8 @@ module Simplex.Messaging.Crypto.SNTRUP761.Bindings
   ) where
 
 import Control.Concurrent.STM
+import Control.Exception (throwIO)
+import Control.Monad (when)
 import Crypto.Random (ChaChaDRG)
 import Data.Aeson (FromJSON (..), ToJSON (..))
 import Data.Bifunctor (bimap)
@@ -23,6 +25,7 @@ import Data.ByteArray (ScrubbedBytes)
 import qualified Data.ByteArray as BA
 import Data.ByteString (ByteString)
 import Simplex.Messaging.Agent.Store.DB (FromField (..), ToField (..))
+import qualified Simplex.Messaging.Crypto as C
 import Simplex.Messaging.Crypto.SNTRUP761.Bindings.Defines
 import Simplex.Messaging.Crypto.SNTRUP761.Bindings.FFI
 import Simplex.Messaging.Crypto.SNTRUP761.Bindings.RNG (rngFuncPtr, withDRG)
@@ -55,14 +58,13 @@ pattern KEMSharedKey s <- KEMSharedKey_ s
 type KEMKeyPair = (KEMPublicKey, KEMSecretKey)
 
 sntrup761Keypair :: TVar ChaChaDRG -> IO KEMKeyPair
-sntrup761Keypair drg =
-  bimap KEMPublicKey_ KEMSecretKey
-    <$> BA.allocRet
-      c_SNTRUP761_SECRETKEY_SIZE
-      ( \skPtr ->
-          BA.alloc c_SNTRUP761_PUBLICKEY_SIZE $ \pkPtr ->
-            withDRG drg $ \cxtPtr -> c_sntrup761_keypair pkPtr skPtr cxtPtr rngFuncPtr
-      )
+sntrup761Keypair drg = do
+  ((r, pk), sk) <-
+    BA.allocRet c_SNTRUP761_SECRETKEY_SIZE $ \skPtr ->
+      BA.allocRet c_SNTRUP761_PUBLICKEY_SIZE $ \pkPtr ->
+        withDRG drg $ \cxtPtr -> c_sntrup761_keypair pkPtr skPtr cxtPtr rngFuncPtr
+  when (r /= 0) $ throwIO C.CryptoKEMKeyGenError
+  pure (KEMPublicKey_ pk, KEMSecretKey sk)
 
 sntrup761Enc :: TVar ChaChaDRG -> KEMPublicKey -> IO (KEMCiphertext, KEMSharedKey)
 sntrup761Enc drg (KEMPublicKey pk) =

src/Simplex/Messaging/Crypto/SNTRUP761/Bindings/FFI.hs

@@ -10,9 +10,9 @@ import Foreign
 import Foreign.C
 import Simplex.Messaging.Crypto.SNTRUP761.Bindings.RNG (RNGContext, RNGFunc)
 
--- void sntrup761_keypair (uint8_t *pk, uint8_t *sk, void *random_ctx, sntrup761_random_func *random);
+-- int sntrup761_keypair (uint8_t *pk, uint8_t *sk, void *random_ctx, sntrup761_random_func *random);
 foreign import ccall "sntrup761_keypair"
-  c_sntrup761_keypair :: Ptr Word8 -> Ptr Word8 -> Ptr RNGContext -> FunPtr RNGFunc -> IO ()
+  c_sntrup761_keypair :: Ptr Word8 -> Ptr Word8 -> Ptr RNGContext -> FunPtr RNGFunc -> IO CInt
 
 -- void sntrup761_enc (uint8_t *c, uint8_t *k, const uint8_t *pk, void *random_ctx, sntrup761_random_func *random);
 foreign import ccall "sntrup761_enc"

src/Simplex/Messaging/Crypto/ShortLink.hs

@@ -17,6 +17,7 @@ module Simplex.Messaging.Crypto.ShortLink
     encodeSignUserData,
     newOwnerAuth,
     encryptLinkData,
+    encryptFixedData,
     encryptUserData,
     decryptLinkData,
   )
@@ -84,9 +85,10 @@ newOwnerAuth g ownerId signingKey = do
   pure (ownerPrivKey, OwnerAuth {ownerId, ownerKey, authOwnerSig})
 
 encryptLinkData :: TVar ChaChaDRG -> C.SbKey -> (ByteString, ByteString) -> ExceptT AgentErrorType IO QueueLinkData
-encryptLinkData g k = bimapM (encrypt fixedDataPaddedLength) (encrypt userDataPaddedLength)
-  where
-    encrypt len = encryptData g k len
+encryptLinkData g k = bimapM (encryptFixedData g k) (encryptUserData g k)
+
+encryptFixedData :: TVar ChaChaDRG -> C.SbKey -> ByteString -> ExceptT AgentErrorType IO EncDataBytes
+encryptFixedData g k s = encryptData g k fixedDataPaddedLength s
 
 encryptUserData :: TVar ChaChaDRG -> C.SbKey -> ByteString -> ExceptT AgentErrorType IO EncDataBytes
 encryptUserData g k s = encryptData g k userDataPaddedLength s

src/Simplex/Messaging/Names/Record.hs

@@ -1,3 +1,6 @@
+{-# LANGUAGE DerivingStrategies #-}
+{-# LANGUAGE GeneralizedNewtypeDeriving #-}
+{-# LANGUAGE LambdaCase #-}
 {-# LANGUAGE NamedFieldPuns #-}
 {-# LANGUAGE OverloadedStrings #-}
 {-# LANGUAGE StrictData #-}
@@ -5,13 +8,24 @@
 
 module Simplex.Messaging.Names.Record
   ( NameRecord (..),
+    NameResponse (..),
+    NameRegistration (..),
+    NamePricing (..),
+    USDCents (..),
+    NameReservedReason (..),
   )
 where
 
+import Data.Aeson (FromJSON (..), ToJSON (..))
 import qualified Data.Aeson as J
 import qualified Data.Aeson.TH as JQ
+import Data.Int (Int64)
+import Data.Map.Strict (Map)
 import Data.Text (Text)
-import Simplex.Messaging.Parsers (defaultJSON, dropPrefix)
+import qualified Data.Text as T
+import Simplex.Messaging.Encoding.String
+import Simplex.Messaging.Parsers (defaultJSON, dropPrefix, taggedObjectJSON)
+import Simplex.Messaging.SystemTime (SystemSeconds)
 
 -- | Resolved name record returned by the names role. JSON keys match the
 --   resolver REST output; both FromJSON (resolver -> server) and ToJSON
@@ -44,3 +58,87 @@ $( JQ.deriveJSON
     defaultJSON {J.omitNothingFields = False, J.fieldLabelModifier = dropPrefix "nr"}
     ''NameRecord
  )
+
+-- | US cents.
+newtype USDCents = USDCents Int64
+  deriving (Eq, Ord, Show)
+  deriving newtype (ToJSON, FromJSON)
+
+-- | What the registry holds for a name, and the block it was read at.
+data NameResponse = NameResponse
+  { -- | absent only from a v20/v21 router, which sent the record alone
+    lastBlockTs :: Maybe SystemSeconds,
+    registration :: NameRegistration
+  }
+  deriving (Eq, Show)
+
+-- | What the registry holds for a name.
+data NameRegistration
+  = -- | Held by someone. Always carries a record, empty where none was set.
+    NRRegistered
+      { -- | absent only from a v20/v21 router, which sent the record alone
+        expires :: Maybe SystemSeconds,
+        -- | unix seconds, > expires: until here only the owner may renew
+        graceUntil :: Maybe SystemSeconds,
+        -- | held back as well, which is why it will not free up at expiry
+        reservedReason_ :: Maybe NameReservedReason,
+        nameRecord :: NameRecord
+      }
+  | -- | Held by nobody, and registrable now.
+    NRAvailable {pricing :: NamePricing}
+  | -- | Held back by the registry, and not for sale at its price.
+    NRReserved {reservedReason :: NameReservedReason}
+  deriving (Eq, Show)
+
+-- | Enough to price the name locally, which the router cannot do behind a hash.
+data NamePricing = NamePricing
+  { -- | US cents per year, for the lengths the registry prices specially
+    registrationPrices :: Map Int USDCents,
+    -- | US cents per year for every other length
+    basePrice :: USDCents,
+    -- | characters; the registry refuses shorter labels
+    minLabelLength :: Int
+  }
+  deriving (Eq, Show)
+
+-- | Why the registry holds a name back.
+data NameReservedReason
+  = -- | held for SimpleX
+    NRRInternal
+  | NRRTrademark
+  | NRRCommunity
+  | -- | added to the registry after this version, and still reserved
+    NRRUnknown Text
+  deriving (Eq, Show)
+
+instance TextEncoding NameReservedReason where
+  textEncode = \case
+    NRRInternal -> "internal"
+    NRRTrademark -> "trademark"
+    NRRCommunity -> "community"
+    NRRUnknown t -> t
+  textDecode = Just . reservedReasonOf
+
+-- | An unknown reason is kept as text, capped: it reaches a client as a word.
+-- Capping precedes the match, so what is kept encodes back to what it decoded.
+reservedReasonOf :: Text -> NameReservedReason
+reservedReasonOf t = case T.take 32 $ T.takeWhile (\c -> c > ' ' && c < '\DEL') t of
+  "internal" -> NRRInternal
+  "trademark" -> NRRTrademark
+  "community" -> NRRCommunity
+  r -> NRRUnknown r
+
+instance ToJSON NameReservedReason where
+  toJSON = textToJSON
+  toEncoding = textToEncoding
+
+instance FromJSON NameReservedReason where
+  parseJSON = textParseJSON "NameReservedReason"
+
+$(JQ.deriveJSON defaultJSON ''NamePricing)
+
+-- taggedObjectJSON, not sumTypeJSON: this JSON is the RNAME payload and the
+-- resolver contract, so it must not vary with the swift build flag.
+$(JQ.deriveJSON (taggedObjectJSON $ dropPrefix "NR") ''NameRegistration)
+
+$(JQ.deriveJSON defaultJSON ''NameResponse)

src/Simplex/Messaging/Notifications/Server/Store/ntf_server_schema.sql

@@ -1,5 +1,6 @@
 
 
+
 SET statement_timeout = 0;
 SET lock_timeout = 0;
 SET idle_in_transaction_session_timeout = 0;
@@ -312,3 +313,4 @@ ALTER TABLE ONLY ntf_server.subscriptions
 
 
 
+

src/Simplex/Messaging/Protocol.hs

@@ -80,6 +80,12 @@ module Simplex.Messaging.Protocol
     ErrorType (..),
     CommandError (..),
     ProxyError (..),
+    NameQuery (..),
+    NameResponse (..),
+    NameRegistration (..),
+    NamePricing (..),
+    USDCents (..),
+    NameReservedReason (..),
     NameErrorType (..),
     BrokerErrorType (..),
     NetworkError (..),
@@ -139,7 +145,10 @@ module Simplex.Messaging.Protocol
     RcvNtfDhSecret,
     Message (..),
     RcvMessage (..),
-    MsgId,
+    MsgId (unMsgId),
+    mkMsgId,
+    unsafeMsgId,
+    randomMsgId,
     MsgBody,
     IdsHash (..),
     ServiceSub (..),
@@ -162,6 +171,7 @@ module Simplex.Messaging.Protocol
     EncFwdTransmission (..),
     EncResponse (..),
     EncTransmission (..),
+    encTransmissionNonce,
     FwdResponse (..),
     FwdTransmission (..),
     NameRecord (..),
@@ -226,8 +236,10 @@ module Simplex.Messaging.Protocol
 where
 
 import Control.Applicative (optional, (<|>))
+import Control.Concurrent.STM (STM, TVar)
 import Control.Exception (Exception, SomeException, displayException, fromException)
 import Control.Monad.Except
+import Crypto.Random (ChaChaDRG)
 import Data.Aeson (FromJSON (..), ToJSON (..))
 import qualified Data.Aeson as J
 import qualified Data.Aeson.TH as J
@@ -235,6 +247,7 @@ import Data.Attoparsec.ByteString.Char8 (Parser, (<?>))
 import qualified Data.Attoparsec.ByteString.Char8 as A
 import Data.Bifunctor (bimap, first)
 import Data.Bits (xor)
+import qualified Data.ByteArray as BA
 import qualified Data.ByteString as BS
 import qualified Data.ByteString.Base64 as B64
 import Data.ByteString.Char8 (ByteString)
@@ -261,16 +274,16 @@ import qualified GHC.TypeLits as TE
 import qualified GHC.TypeLits as Type
 import Network.Socket (ServiceName)
 import qualified Network.TLS as TLS
-import Simplex.Messaging.Agent.Store.DB (Binary (..), FromField (..), ToField (..))
+import Simplex.Messaging.Agent.Store.DB (Binary (..), FromField (..), ToField (..), blobFieldDecoder)
 import qualified Simplex.Messaging.Crypto as C
 import Simplex.Messaging.Encoding
 import Simplex.Messaging.Encoding.String
-import Simplex.Messaging.Names.Record (NameRecord (..))
+import Simplex.Messaging.Names.Record
 import Simplex.Messaging.Parsers
 import Simplex.Messaging.Protocol.Types
 import Simplex.Messaging.Server.QueueStore.QueueInfo
 import Simplex.Messaging.ServiceScheme
-import Simplex.Messaging.SimplexName (SimplexDomain)
+import Simplex.Messaging.SimplexName (LabelHash, SimplexDomain (..), SimplexTLD (..), fullDomainName, labelHash)
 import Simplex.Messaging.Transport
 import Simplex.Messaging.Transport.Client (TransportHost, TransportHosts (..))
 import Simplex.Messaging.Util (bshow, eitherToMaybe, safeDecodeUtf8, (<$?>))
@@ -603,7 +616,7 @@ data Command (p :: Party) where
   -- - corrId: unique correlation ID between proxy and relay, also used as a nonce to encrypt forwarded transmission
   RFWD :: EncFwdTransmission -> Command ProxyService -- use CorrId as CbNonce, proxy to relay
   -- Resolve SimpleX name.
-  RSLV :: SimplexDomain -> Command Resolver
+  RSLV :: NameQuery -> Command Resolver
 
 deriving instance Show (Command p)
 
@@ -695,19 +708,24 @@ instance Encoding NewNtfCreds where
 newtype EncTransmission = EncTransmission ByteString
   deriving (Show)
 
+encTransmissionNonce :: C.CbNonceI n => VersionSMP -> n -> C.CbNonce
+encTransmissionNonce v nonce
+  | v >= fwdNoncesSMPVersion = C.xorNonce (smpEncode v) nonce
+  | otherwise = C.toCbNonce nonce
+
 data FwdTransmission = FwdTransmission
-  { fwdCorrId :: CorrId,
+  { fwdCorrId :: C.CorrCbNonce,
     fwdVersion :: VersionSMP,
     fwdKey :: C.PublicKeyX25519,
     fwdTransmission :: EncTransmission
   }
 
 instance Encoding FwdTransmission where
-  smpEncode FwdTransmission {fwdCorrId = CorrId corrId, fwdVersion, fwdKey, fwdTransmission = EncTransmission t} =
-    smpEncode (corrId, fwdVersion, fwdKey, Tail t)
+  smpEncode FwdTransmission {fwdCorrId, fwdVersion, fwdKey, fwdTransmission = EncTransmission t} =
+    smpEncode (fwdCorrId, fwdVersion, fwdKey, Tail t)
   smpP = do
-    (corrId, fwdVersion, fwdKey, Tail t) <- smpP
-    pure FwdTransmission {fwdCorrId = CorrId corrId, fwdVersion, fwdKey, fwdTransmission = EncTransmission t}
+    (fwdCorrId, fwdVersion, fwdKey, Tail t) <- smpP
+    pure FwdTransmission {fwdCorrId, fwdVersion, fwdKey, fwdTransmission = EncTransmission t}
 
 newtype EncFwdTransmission = EncFwdTransmission ByteString
   deriving (Show)
@@ -730,8 +748,8 @@ data BrokerMsg where
   NMSG :: C.CbNonce -> EncNMsgMeta -> BrokerMsg
   -- Should include certificate chain
   PKEY :: SessionId -> VersionRangeSMP -> CertChainPubKey -> BrokerMsg -- TLS-signed server key for proxy shared secret and initial sender key
-  RRES :: EncFwdResponse -> BrokerMsg -- relay to proxy
-  PRES :: EncResponse -> BrokerMsg -- proxy to client
+  RRES :: Maybe C.CbNonce -> EncFwdResponse -> BrokerMsg -- relay to proxy
+  PRES :: Maybe C.CbNonce -> EncResponse -> BrokerMsg -- proxy to client
   END :: BrokerMsg
   ENDS :: Int64 -> IdsHash -> BrokerMsg
   DELD :: BrokerMsg
@@ -739,8 +757,8 @@ data BrokerMsg where
   OK :: BrokerMsg
   ERR :: ErrorType -> BrokerMsg
   PONG :: BrokerMsg
-  -- Resolved SimpleX name.
-  RNAME :: NameRecord -> BrokerMsg
+  -- What the router knows about a SimpleX name.
+  RNAME :: NameResponse -> BrokerMsg
   deriving (Eq, Show)
 
 data RcvMessage = RcvMessage
@@ -753,16 +771,16 @@ newtype EncFwdResponse = EncFwdResponse ByteString
   deriving (Eq, Show)
 
 data FwdResponse = FwdResponse
-  { fwdCorrId :: CorrId,
+  { fwdCorrId :: C.CorrCbNonce,
     fwdResponse :: EncResponse
   }
 
 instance Encoding FwdResponse where
-  smpEncode FwdResponse {fwdCorrId = CorrId corrId, fwdResponse = EncResponse t} =
-    smpEncode (corrId, Tail t)
+  smpEncode FwdResponse {fwdCorrId, fwdResponse = EncResponse t} =
+    smpEncode (fwdCorrId, Tail t)
   smpP = do
-    (corrId, Tail t) <- smpP
-    pure FwdResponse {fwdCorrId = CorrId corrId, fwdResponse = EncResponse t}
+    (fwdCorrId, Tail t) <- smpP
+    pure FwdResponse {fwdCorrId, fwdResponse = EncResponse t}
 
 newtype EncResponse = EncResponse ByteString
   deriving (Eq, Show)
@@ -785,7 +803,7 @@ toMsgInfo = \case
   Message {msgId, msgTs} -> msgInfo msgId msgTs MTMessage
   MessageQuota {msgId, msgTs} -> msgInfo msgId msgTs MTQuota
   where
-    msgInfo msgId msgTs msgType = MsgInfo {msgId = decodeLatin1 $ B64.encode msgId, msgTs = systemToUTCTime msgTs, msgType}
+    msgInfo msgId msgTs msgType = MsgInfo {msgId = decodeLatin1 $ B64.encode $ unMsgId msgId, msgTs = systemToUTCTime msgTs, msgType}
 
 messageId :: Message -> MsgId
 messageId = \case
@@ -1317,7 +1335,10 @@ instance ProtocolTypeI p => FromJSON (ProtocolServer p) where
   parseJSON = strParseJSON "ProtocolServer"
 
 newtype BasicAuth = BasicAuth {unBasicAuth :: ByteString}
-  deriving (Eq, Ord, Show)
+  deriving (Ord, Show)
+
+instance Eq BasicAuth where
+  BasicAuth s == BasicAuth s' = BA.constEq s s'
 
 instance IsString BasicAuth where fromString = BasicAuth . B.pack
 
@@ -1483,7 +1504,34 @@ type RcvNtfPublicDhKey = C.PublicKeyX25519
 type RcvNtfDhSecret = C.DhSecretX25519
 
 -- | SMP message server ID.
-type MsgId = ByteString
+newtype MsgId = MsgId {unMsgId :: ByteString}
+  deriving (Eq, Show)
+
+instance C.CbNonceI MsgId where
+  unCbNonce (MsgId s) = B.take 24 s
+
+instance Encoding MsgId where
+  smpEncode (MsgId s) = smpEncode s
+  smpP = mkMsgId <$?> smpP
+
+instance StrEncoding MsgId where
+  strEncode (MsgId s) = strEncode s
+  strP = mkMsgId <$?> strP
+
+instance FromField MsgId where fromField = blobFieldDecoder mkMsgId
+
+instance ToField MsgId where toField (MsgId s) = toField $ Binary s
+
+mkMsgId :: ByteString -> Either String MsgId
+mkMsgId s
+  | B.length s >= 24 = Right $ MsgId s
+  | otherwise = Left "MsgId: invalid length"
+
+unsafeMsgId :: ByteString -> MsgId
+unsafeMsgId s = either error id $ mkMsgId s
+
+randomMsgId :: Int -> TVar ChaChaDRG -> STM MsgId
+randomMsgId n = fmap MsgId . C.randomBytes (max 24 n)
 
 -- | SMP message body.
 type MsgBody = ByteString
@@ -1589,11 +1637,28 @@ data ErrorType
     DUPLICATE_ -- not part of SMP protocol, used internally
   deriving (Eq, Show)
 
+-- | What RSLV asks about: a name, or the hash of a second-level label.
+data NameQuery = NQDomain SimplexDomain | NQHash LabelHash SimplexTLD
+  deriving (Eq, Show)
+
+instance Encoding NameQuery where
+  smpEncode = \case
+    NQDomain d -> encodeUtf8 $ fullDomainName d
+    NQHash h tld -> strEncode h <> strEncode tld
+  smpP = NQHash <$> strP <*> strP <|> NQDomain <$> strP
+
+-- | Hashed from v22, except a name with subnames or a web TLD.
+hashedQuery :: NameQuery -> NameQuery
+hashedQuery q = case q of
+  NQDomain SimplexDomain {nameTLD, domain, subDomain}
+    | null subDomain && nameTLD /= TLDWeb -> NQHash (labelHash domain) nameTLD
+  _ -> q
+
 -- | Name resolution error
 data NameErrorType
   = -- | the names role / resolver is not configured on this server
     NO_RESOLVER
-  | -- | the name is not registered (resolver returned not-found)
+  | -- | the name does not resolve; sent only to a session below v22
     NOT_FOUND
   | -- | backing resolver/RPC failure - contains the diagnostic detail
     RESOLVER {resolverErr :: Text}
@@ -1791,8 +1856,7 @@ instance PartyI p => ProtocolEncoding SMPVersion ErrorType (Command p) where
   encodeProtocol v = \case
     NEW NewQueueReq {rcvAuthKey = rKey, rcvDhKey = dhKey, auth_, subMode, queueReqData, ntfCreds}
       | v >= newNtfCredsSMPVersion -> new <> e (subMode, queueReqData, ntfCreds)
-      | v >= shortLinksSMPVersion -> new <> e (subMode, queueReqData)
-      | otherwise -> new <> e (subMode, senderCanSecure (queueReqMode <$> queueReqData))
+      | otherwise -> new <> e (subMode, queueReqData)
       where
         new = e (NEW_, ' ', rKey, dhKey, auth_)
     SUB -> e SUB_
@@ -1822,7 +1886,7 @@ instance PartyI p => ProtocolEncoding SMPVersion ErrorType (Command p) where
     PRXY host auth_ -> e (PRXY_, ' ', host, auth_)
     PFWD fwdV pubKey (EncTransmission s) -> e (PFWD_, ' ', fwdV, pubKey, Tail s)
     RFWD (EncFwdTransmission s) -> e (RFWD_, ' ', Tail s)
-    RSLV d -> e (RSLV_, ' ', d)
+    RSLV q -> e (RSLV_, ' ', if v >= nameAvailSMPVersion then hashedQuery q else q)
     where
       e :: Encoding a => a -> ByteString
       e = smpEncode
@@ -1879,20 +1943,18 @@ instance ProtocolEncoding SMPVersion ErrorType Cmd where
     CT SCreator NEW_ -> Cmd SCreator <$> newCmd
       where
         newCmd
-          | v >= newNtfCredsSMPVersion = new smpP smpP
-          | v >= shortLinksSMPVersion = new smpP nothing
-          | otherwise = new (qReq <$> smpP) nothing
+          | v >= newNtfCredsSMPVersion = new smpP
+          | otherwise = new nothing
           where
             nothing = pure Nothing
-            new p2 p3 = NEW <$> do
+            new p3 = NEW <$> do
               rcvAuthKey <- _smpP
               rcvDhKey <- smpP
               auth_ <- smpP
               subMode <- smpP
-              queueReqData <- p2
+              queueReqData <- smpP
               ntfCreds <- p3
               pure NewQueueReq {rcvAuthKey, rcvDhKey, auth_, subMode, queueReqData, ntfCreds}
-            qReq sndSecure = Just $ if sndSecure then QRMessaging Nothing else QRContact Nothing
     CT SRecipient tag ->
       Cmd SRecipient <$> case tag of
         SUB_ -> pure SUB
@@ -1929,7 +1991,7 @@ instance ProtocolEncoding SMPVersion ErrorType Cmd where
     CT SNotifierService NSUBS_
       | v >= rcvServiceSMPVersion -> Cmd SNotifierService <$> (NSUBS <$> _smpP <*> smpP)
       | otherwise -> pure $ Cmd SNotifierService $ NSUBS (-1) mempty
-    CT SResolver RSLV_ -> Cmd SResolver . RSLV <$> _smpP <* A.takeByteString
+    CT SResolver RSLV_ -> Cmd SResolver . RSLV <$> _smpP
 
   fromProtocolError = fromProtocolError @SMPVersion @ErrorType @BrokerMsg
   {-# INLINE fromProtocolError #-}
@@ -1943,8 +2005,7 @@ instance ProtocolEncoding SMPVersion ErrorType BrokerMsg where
     IDS QIK {rcvId, sndId, rcvPublicDhKey = srvDh, queueMode, linkId, serviceId, serverNtfCreds}
       | v >= newNtfCredsSMPVersion -> ids <> e (queueMode, linkId, serviceId, serverNtfCreds)
       | v >= serviceCertsSMPVersion -> ids <> e (queueMode, linkId, serviceId)
-      | v >= shortLinksSMPVersion -> ids <> e (queueMode, linkId)
-      | otherwise -> ids <> e (senderCanSecure queueMode)
+      | otherwise -> ids <> e (queueMode, linkId)
       where
         ids = e (IDS_, ' ', rcvId, sndId, srvDh)
     LNK sId d -> e (LNK_, ' ', sId, d)
@@ -1958,8 +2019,8 @@ instance ProtocolEncoding SMPVersion ErrorType BrokerMsg where
     NID nId srvNtfDh -> e (NID_, ' ', nId, srvNtfDh)
     NMSG nmsgNonce encNMsgMeta -> e (NMSG_, ' ', nmsgNonce, encNMsgMeta)
     PKEY sid vr certKey -> e (PKEY_, ' ', sid, vr, certKey)
-    RRES (EncFwdResponse encBlock) -> e (RRES_, ' ', Tail encBlock)
-    PRES (EncResponse encBlock) -> e (PRES_, ' ', Tail encBlock)
+    RRES nonce_ (EncFwdResponse encBlock) -> fwdResp RRES_ nonce_ encBlock
+    PRES nonce_ (EncResponse encBlock) -> fwdResp PRES_ nonce_ encBlock
     END -> e END_
     ENDS n idsHash -> serviceResp ENDS_ n idsHash
     DELD -> e DELD_
@@ -1972,13 +2033,20 @@ instance ProtocolEncoding SMPVersion ErrorType BrokerMsg where
             | v < clientNoticesSMPVersion -> BLOCKED info {notice = Nothing}
           _ -> err
     PONG -> e PONG_
-    RNAME rec -> e (RNAME_, ' ', Tail $ LB.toStrict $ J.encode rec)
+    RNAME res
+      | v >= nameAvailSMPVersion -> e (RNAME_, ' ', Tail $ LB.toStrict $ J.encode res)
+      | otherwise -> case registration res of
+          NRRegistered {nameRecord} -> e (RNAME_, ' ', Tail $ LB.toStrict $ J.encode nameRecord)
+          _ -> e (ERR_, ' ', NAME NOT_FOUND)
     where
       e :: Encoding a => a -> ByteString
       e = smpEncode
       serviceResp tag n idsHash
         | v >= rcvServiceSMPVersion = e (tag, ' ', n, idsHash)
         | otherwise = e (tag, ' ', n)
+      fwdResp tag nonce_ encBlock
+        | v >= fwdNoncesSMPVersion = e (tag, ' ', nonce_, Tail encBlock)
+        | otherwise = e (tag, ' ', Tail encBlock)
 
   protocolP v = \case
     MSG_ -> do
@@ -1988,19 +2056,17 @@ instance ProtocolEncoding SMPVersion ErrorType BrokerMsg where
         bodyP = EncRcvMsgBody . unTail <$> smpP
     ALLS_ -> pure ALLS
     IDS_
-      | v >= newNtfCredsSMPVersion -> ids smpP smpP smpP smpP
-      | v >= serviceCertsSMPVersion -> ids smpP smpP smpP nothing
-      | v >= shortLinksSMPVersion -> ids smpP smpP nothing nothing
-      | otherwise -> ids (qm <$> smpP) nothing nothing nothing
+      | v >= newNtfCredsSMPVersion -> ids smpP smpP
+      | v >= serviceCertsSMPVersion -> ids smpP nothing
+      | otherwise -> ids nothing nothing
       where
-        qm sndSecure = Just $ if sndSecure then QMMessaging else QMContact
         nothing = pure Nothing
-        ids p1 p2 p3 p4 = do
+        ids p3 p4 = do
           rcvId <- _smpP
           sndId <- smpP
           rcvPublicDhKey <- smpP
-          queueMode <- p1
-          linkId <- p2
+          queueMode <- smpP
+          linkId <- smpP
           serviceId <- p3
           serverNtfCreds <- p4
           pure $ IDS QIK {rcvId, sndId, rcvPublicDhKey, queueMode, linkId, serviceId, serverNtfCreds}
@@ -2010,8 +2076,8 @@ instance ProtocolEncoding SMPVersion ErrorType BrokerMsg where
     NID_ -> NID <$> _smpP <*> smpP
     NMSG_ -> NMSG <$> _smpP <*> smpP
     PKEY_ -> PKEY <$> _smpP <*> smpP <*> smpP
-    RRES_ -> RRES <$> (EncFwdResponse . unTail <$> _smpP)
-    PRES_ -> PRES <$> (EncResponse . unTail <$> _smpP)
+    RRES_ -> fwdRespP RRES EncFwdResponse
+    PRES_ -> fwdRespP PRES EncResponse
     END_ -> pure END
     ENDS_ -> serviceRespP ENDS
     DELD_ -> pure DELD
@@ -2019,11 +2085,18 @@ instance ProtocolEncoding SMPVersion ErrorType BrokerMsg where
     OK_ -> pure OK
     ERR_ -> ERR <$> _smpP
     PONG_ -> pure PONG
-    RNAME_ -> fmap RNAME . J.eitherDecodeStrict . unTail <$?> _smpP
+    RNAME_
+      | v >= nameAvailSMPVersion -> fmap RNAME . J.eitherDecodeStrict . unTail <$?> _smpP
+      | otherwise -> fmap (RNAME . oldResponse) . J.eitherDecodeStrict . unTail <$?> _smpP
     where
+      oldResponse nameRecord =
+        NameResponse {lastBlockTs = Nothing, registration = NRRegistered {expires = Nothing, graceUntil = Nothing, reservedReason_ = Nothing, nameRecord}}
       serviceRespP resp
         | v >= rcvServiceSMPVersion = resp <$> _smpP <*> smpP
         | otherwise = resp <$> _smpP <*> pure mempty
+      fwdRespP :: (Maybe C.CbNonce -> a -> BrokerMsg) -> (ByteString -> a) -> Parser BrokerMsg
+      fwdRespP resp enc = resp <$> (A.space *> nonceP) <*> (enc <$> A.takeByteString)
+      nonceP = if v >= fwdNoncesSMPVersion then smpP else pure Nothing
 
   fromProtocolError = \case
     PECmdSyntax -> CMD SYNTAX
@@ -2040,9 +2113,9 @@ instance ProtocolEncoding SMPVersion ErrorType BrokerMsg where
     -- PONG response must not have queue ID
     PONG -> noEntityMsg
     PKEY {} -> noEntityMsg
-    RRES _ -> noEntityMsg
+    RRES {} -> noEntityMsg
     ALLS -> noEntityMsg
-    RNAME _ -> noEntityMsg
+    RNAME {} -> noEntityMsg
     -- other broker responses must have queue ID
     _
       | B.null entId -> Left $ CMD NO_ENTITY
@@ -2376,7 +2449,9 @@ tDecodeServer THandleParams {sessionId, thVersion = v, implySessId} = \case
           where
             cmdOrErr = parseProtocol @v @err @cmd v command >>= checkCredentials tAuth entityId
             t :: a -> (CorrId, EntityId, a)
-            t = (corrId,entityId,)
+            -- IDs are slices of the ~16 KB received block and are kept as subscription keys,
+            -- so without a copy one live key retains the whole block
+            t = (CorrId $ B.copy $ bs corrId,EntityId $ B.copy $ unEntityId entityId,)
         Left _ -> tError corrId PEBlock
     | otherwise -> tError corrId PESession
   Left _ -> tError "" PEBlock
@@ -2412,3 +2487,4 @@ $(J.deriveJSON defaultJSON ''BlockingInfo)
 
 -- run deriveJSON in one TH splice to allow mutual instance
 $(concat <$> mapM @[] (J.deriveJSON (sumTypeJSON id)) [''ProxyError, ''NameErrorType, ''ErrorType])
+

src/Simplex/Messaging/Server.hs

@@ -98,13 +98,12 @@ import Network.Socket (ServiceName, Socket, socketToHandle)
 import qualified Network.TLS as TLS
 import Numeric.Natural (Natural)
 import Simplex.Messaging.Agent.Lock
-import Simplex.Messaging.Client (ProtocolClient (thParams), ProtocolClientError (..), SMPClient, SMPClientError, clientHandlers, forwardSMPTransmission, smpProxyError, temporaryClientError, transportHost')
+import Simplex.Messaging.Client (ProtocolClient (thParams), ProtocolClientError (..), SMPClient, SMPClientError, clientHandlers, forwardSMPTransmission, smpProxyError, temporaryClientError)
 import Simplex.Messaging.Client.Agent (OwnServer, SMPClientAgent (..), SMPClientAgentEvent (..), closeSMPClientAgent, getSMPServerClient'', isOwnServer, lookupSMPServerClient, getConnectedSMPServerClient)
 import qualified Simplex.Messaging.Crypto as C
 import Simplex.Messaging.Encoding
 import Simplex.Messaging.Encoding.String
 import Simplex.Messaging.Protocol
-import Simplex.Messaging.SimplexName (SimplexDomain)
 import Simplex.Messaging.Server.Control
 import Simplex.Messaging.Server.Env.STM as Env
 import Simplex.Messaging.Server.Expiration
@@ -1093,6 +1092,7 @@ controlPortAuth h user admin role auth = do
   readTVarIO role >>= \case
     CPRNone -> do
       atomically $ writeTVar role $! newRole
+      when (newRole == CPRNone) $ logWarn "ControlPort: failed auth"
       hPutStrLn h $ currentRole newRole
     r -> hPutStrLn h $ currentRole r <> if r == newRole then "" else ", start new session to change."
   where
@@ -1145,14 +1145,16 @@ cancelSub s = case subThread s of
 type VerifiedTransmissionOrError s = Either (Transmission BrokerMsg) (VerifiedTransmission s)
 
 receive :: forall c s. (Transport c, MsgStoreClass s) => THandleSMP c 'TServer -> s -> Client s -> M s ()
-receive h@THandle {params = THandleParams {thAuth, sessionId}} ms Client {rcvQ, sndQ, rcvActiveAt} = do
+receive h@THandle {params = THandleParams {thAuth, sessionId}} ms c@Client {rcvQ, rcvActiveAt} = do
   labelMyThread . B.unpack $ "client $" <> encode sessionId <> " receive"
   sa <- asks serverActive
   stats <- asks serverStats
+  authDelay <- asks $ authErrorDelay . config
   liftIO $ forever $ do
     ts <- tGetServer h
     unlessM (readTVarIO sa) $ throwIO $ userError "server stopped"
-    atomically . (writeTVar rcvActiveAt $!) =<< getSystemTime
+    receivedAt <- getSystemTime
+    atomically $ writeTVar rcvActiveAt $! receivedAt
     let (es, ts') = partitionEithers $ L.toList ts
         errs = map (second ERR) es
     errs' <- case ts' of
@@ -1168,7 +1170,7 @@ receive h@THandle {params = THandleParams {thAuth, sessionId}} ms Client {rcvQ,
         mapM_ (atomically . writeTBQueue rcvQ) $ L.nonEmpty cmds
         pure $ errs ++ errs'
       [] -> pure errs
-    mapM_ (atomically . writeTBQueue sndQ . (,[])) $ L.nonEmpty errs'
+    mapM_ (sendResponses c stats authDelay receivedAt . (,[])) $ L.nonEmpty errs'
   where
     sameParty :: SParty p -> SignedTransmission Cmd -> Bool
     sameParty p (_, _, (_, _, Cmd p' _)) = isJust $ testEquality p p'
@@ -1191,6 +1193,27 @@ receive h@THandle {params = THandleParams {thAuth, sessionId}} ms Client {rcvQ,
             GET -> incStat $ msgGetAuth stats
             _ -> pure ()
 
+isAuthError :: Transmission BrokerMsg -> Bool
+isAuthError (_, _, r) = case r of
+  ERR AUTH -> True
+  _ -> False
+
+sendResponses :: Client s -> ServerStats -> Int64 -> SystemTime -> (NonEmpty (Transmission BrokerMsg), [Transmission BrokerMsg]) -> IO ()
+sendResponses c@Client {sndQ} stats authDelay since rs@(ts, _)
+  | any isAuthError ts = sendAfterDelay c stats authDelay since rs
+  | otherwise = atomically $ writeTBQueue sndQ rs
+
+sendAfterDelay :: Client s -> ServerStats -> Int64 -> SystemTime -> (NonEmpty (Transmission BrokerMsg), [Transmission BrokerMsg]) -> IO ()
+sendAfterDelay c@Client {sndQ, clientTHParams = THandleParams {sessionId}} stats delay since rs
+  | delay > 0 = do
+      elapsed <- subtract (microseconds since) . microseconds <$> getSystemTime
+      when (elapsed >= delay) $ incStat $ authDelayExceeded stats
+      forkClient c (B.unpack $ "client $" <> encode sessionId <> " auth") $ threadDelay' (delay - elapsed `mod` delay) >> write
+  | otherwise = write
+  where
+    write = atomically $ writeTBQueue sndQ rs
+    microseconds (MkSystemTime s ns) = s * 1000000 + fromIntegral (ns `div` 1000)
+
 send :: Transport c => MVar (THandleSMP c 'TServer) -> Client s -> IO ()
 send th c@Client {sndQ, msgQ, clientTHParams = THandleParams {sessionId}} = do
   labelMyThread . B.unpack $ "client $" <> encode sessionId <> " send"
@@ -1319,11 +1342,6 @@ isContactQueue QueueRec {queueMode, senderKey} = case queueMode of
   Just QMContact -> True
   Nothing -> isNothing senderKey -- for backward compatibility with pre-SKEY contact addresses
 
-isSecuredMsgQueue :: QueueRec -> Bool
-isSecuredMsgQueue QueueRec {queueMode, senderKey} = case queueMode of
-  Just QMContact -> False
-  _ -> isJust senderKey
-
 -- Random correlation ID is used as a nonce in case crypto_box authenticator is used to authorize transmission
 verifyCmdAuthorization :: Maybe (THandleAuth 'TServer) -> Maybe TAuthorizations -> ByteString -> CorrId -> C.APublicAuthKey -> Bool
 verifyCmdAuthorization thAuth tAuth authorized corrId key = maybe False (verify key) tAuth
@@ -1339,7 +1357,7 @@ verifyCmdAuthorization thAuth tAuth authorized corrId key = maybe False (verify
 
 verifyCmdAuth :: Maybe (THandleAuth 'TServer) -> C.PublicKeyX25519 -> C.CbAuthenticator -> ByteString -> CorrId -> Bool
 verifyCmdAuth thAuth k authenticator authorized (CorrId corrId) = case thAuth of
-  Just THAuthServer {serverPrivKey = pk} -> C.cbVerify k pk (C.cbNonce corrId) authenticator authorized
+  Just THAuthServer {serverPrivKey = pk} -> either (const False) (\nonce -> C.cbVerify k pk nonce authenticator authorized) $ C.cbNonce corrId
   Nothing -> False
 
 dummyVerifyCmd :: Maybe (THandleAuth 'TServer) -> Maybe TAuthorizations -> ByteString -> CorrId -> Bool
@@ -1379,15 +1397,18 @@ client
   ms
   clnt@Client {clientId, rcvQ, sndQ, msgQ, clientTHParams = thParams'@THandleParams {sessionId}, procThreads} = do
     labelMyThread . B.unpack $ "client $" <> encode sessionId <> " commands"
+    stats <- asks serverStats
+    authDelay <- asks $ authErrorDelay . config
     let clntServiceId = (\THClientService {serviceId} -> serviceId) <$> (peerClientService =<< thAuth thParams')
         process batchSubs t acc@(rs, msgs) =
           (maybe acc (\(!r, !msg_) -> (r : rs, maybe msgs (: msgs) msg_)))
             <$> processCommand clntServiceId batchSubs t
     forever $ do
       batch <- atomically (readTBQueue rcvQ)
+      processedAt <- liftIO getSystemTime
       batchSubs <- prepareBatchSubs clntServiceId batch
       foldrM (process batchSubs) ([], []) batch
-        >>= \(rs_, msgs) -> mapM_ (atomically . writeTBQueue sndQ . (,msgs)) (L.nonEmpty rs_)
+        >>= \(rs_, msgs) -> mapM_ (liftIO . sendResponses clnt stats authDelay processedAt . (,msgs)) (L.nonEmpty rs_)
   where
     prepareBatchSubs ::
       Maybe ServiceId ->
@@ -1458,16 +1479,15 @@ client
         ServerStats {pMsgFwds, pMsgFwdsOwn} <- asks serverStats
         let inc = mkIncProxyStats pMsgFwds pMsgFwdsOwn
         liftIO (lookupSMPServerClient a sessId) >>= \case
-          Just (own, smp) -> do
+          Just (own, smp) | Right fwdCorrId <- C.corrCbNonce (bs corrId) -> do
             inc own pRequests
             forkProxiedCmd $ do
-              liftIO (runExceptT (forwardSMPTransmission smp corrId fwdV pubKey encBlock) `E.catches` clientHandlers)  >>= \case
-                Right r -> PRES r <$ inc own pSuccesses
+              liftIO (runExceptT (forwardSMPTransmission smp fwdCorrId fwdV pubKey encBlock) `E.catches` clientHandlers)  >>= \case
+                Right (nonce_, r) -> PRES nonce_ r <$ inc own pSuccesses
                 Left e -> ERR (smpProxyError e) <$ case e of
                   PCEProtocolError {} -> inc own pSuccesses
-                  _ -> do
-                    logWarn $ "Error forwarding to relay: " <> decodeLatin1 (strEncode $ transportHost' smp) <> " own=" <> tshow own <> " " <> tshow e
-                    inc own pErrorsOther
+                  _ -> inc own pErrorsOther
+          Just _ -> pure $ Just $ ERR $ CMD SYNTAX
           Nothing -> inc False pRequests >> inc False pErrorsConnect $> Just (ERR $ PROXY NO_SESSION)
       where
         forkProxiedCmd :: M s BrokerMsg -> M s (Maybe BrokerMsg)
@@ -1496,15 +1516,19 @@ client
         Just nenv -> pure (Just nenv)
     -- Runs on a forked thread so RSLV does not block other commands;
     -- concurrency is limited by serverResolverConcurrency in forkCmd.
-    resolveNameMsg :: NamesEnv -> SimplexDomain -> M s BrokerMsg
-    resolveNameMsg nenv d = do
+    resolveNameMsg :: VersionSMP -> NamesEnv -> NameQuery -> M s BrokerMsg
+    resolveNameMsg v nenv q = do
       st <- asks (rslvStats . serverStats)
       (selector, msg) <-
-        liftIO (resolveName nenv d) <&> \case
-          Right rec -> (rslvSucc, RNAME rec)
-          Left e@NOT_FOUND -> (rslvNotFound, ERR $ NAME e)
+        liftIO (resolveName nenv q) <&> \case
+          Right res -> (if answered (registration res) then rslvSucc else rslvNotFound, RNAME res)
           Left e -> (rslvResolverErrs, ERR $ NAME e)
       incStat (selector st) $> msg
+      where
+        -- below v22 the encoder answers anything but a record as NAME NOT_FOUND
+        answered = \case
+          NRRegistered {} -> True
+          _ -> v >= nameAvailSMPVersion
     transportErr :: TransportError -> ErrorType
     transportErr = PROXY . BROKER . TRANSPORT
     mkIncProxyStats :: MonadIO m => ProxyStats -> ProxyStats -> OwnServer -> (ProxyStats -> IORef Int) -> m ()
@@ -1521,7 +1545,7 @@ client
       Cmd SProxyService (RFWD encBlock) -> (response . (corrId, NoEntity,) =<<) <$> processForwardedCommand encBlock
       Cmd SResolver (RSLV d) -> rslvNamesEnv >>= \case
         Nothing -> pure $ response (corrId, NoEntity, ERR (NAME NO_RESOLVER))
-        Just nenv -> forkCmd serverResolverConcurrency corrId NoEntity (resolveNameMsg nenv d)
+        Just nenv -> forkCmd serverResolverConcurrency corrId NoEntity (resolveNameMsg (thVersion thParams') nenv d)
       Cmd SSenderLink command -> case command of
         LKEY k -> withQueue $ \q qr -> checkMode QMMessaging qr $ secureQueue_ q k $>> getQueueLink_ q qr
         LGET -> withQueue $ \q qr -> checkContact qr $ getQueueLink_ q qr
@@ -1953,7 +1977,7 @@ client
             getDelivered :: Sub -> STM (Maybe (ServerSub, SystemSeconds))
             getDelivered Sub {delivered, subThread} = do
               readTVar delivered $>>= \(msgId', ts) ->
-                if msgId == msgId' || B.null msgId
+                if msgId == msgId'
                   then writeTVar delivered Nothing $> Just (subThread, ts)
                   else pure Nothing
             updateStats :: ServerStats -> Bool -> SystemSeconds -> Message -> IO ()
@@ -1999,7 +2023,7 @@ client
                       when (isJust (queueData qr) && isSecuredMsgQueue qr) $ void $ liftIO $
                         deleteQueueLinkData (queueStore ms) q
                       ServerConfig {messageExpiration, expireMessagesOnSend, msgIdBytes} <- asks config
-                      msgId <- randomId' msgIdBytes
+                      msgId <- atomically . randomMsgId msgIdBytes =<< asks random
                       msg_ <- liftIO $ runExceptT $ do
                         when expireMessagesOnSend $ mapM_ (expireMessages stats) messageExpiration
                         msg <- liftIO $ mkMessage msgId body
@@ -2112,7 +2136,7 @@ client
               liftIO $ storeNtf ns nId ntf
               incStat . ntfCount =<< asks serverStats
 
-            mkMessageNotification :: ByteString -> SystemTime -> RcvNtfDhSecret -> M s MsgNtf
+            mkMessageNotification :: MsgId -> SystemTime -> RcvNtfDhSecret -> M s MsgNtf
             mkMessageNotification msgId msgTs rcvNtfDhSecret = do
               ntfNonce <- atomically . C.randomCbNonce =<< asks random
               let msgMeta = NMsgMeta {msgId, msgTs}
@@ -2121,14 +2145,16 @@ client
 
         processForwardedCommand :: EncFwdTransmission -> M s (Maybe BrokerMsg)
         processForwardedCommand (EncFwdTransmission s) = fmap (either (Just . ERR) id) . runExceptT $ do
+          startedAt <- liftIO getSystemTime
           THAuthServer {serverPrivKey, sessSecret'} <- maybe (throwE $ transportErr TENoServerAuth) pure (thAuth thParams')
           sessSecret <- maybe (throwE $ transportErr TENoServerAuth) pure sessSecret'
-          let proxyNonce = C.cbNonce $ bs corrId
+          proxyNonce <- liftEitherWith (const CRYPTO) $ C.cbNonce $ bs corrId
           s' <- liftEitherWith (const CRYPTO) $ C.cbDecryptNoPad sessSecret proxyNonce s
           FwdTransmission {fwdCorrId, fwdVersion, fwdKey, fwdTransmission = EncTransmission et} <- liftEitherWith (const $ CMD SYNTAX) $ smpDecode s'
+          unless (fwdVersion `isCompatible` thServerVRange thParams') $ throwE $ transportErr TEVersion
           let clientSecret = C.dh' fwdKey serverPrivKey
-              clientNonce = C.cbNonce $ bs fwdCorrId
-          b <- liftEitherWith (const CRYPTO) $ C.cbDecrypt clientSecret clientNonce et
+              clientNonce = fwdCorrId
+          b <- liftEitherWith (const CRYPTO) $ C.cbDecrypt clientSecret (encTransmissionNonce fwdVersion clientNonce) et
           let clntTHParams = smpTHParamsSetVersion fwdVersion thParams'
           -- only allowing single forwarded transactions
           t' <- case tParse clntTHParams b of
@@ -2141,9 +2167,13 @@ client
                   TBError _ _ : _ -> throwE BLOCK
                   TBTransmission b' _ : _ -> pure b'
                   TBTransmissions b' _ _ : _ -> pure b'
-                r2 <- liftEitherWith (const BLOCK) $ EncResponse <$> C.cbEncrypt clientSecret (C.reverseNonce clientNonce) r' paddedProxiedTLength
+                nonce_ <-
+                  if fwdVersion >= fwdNoncesSMPVersion
+                    then Just <$> (atomically . C.randomCbNonce =<< asks random)
+                    else pure Nothing
+                r2 <- liftEitherWith (const BLOCK) $ EncResponse <$> C.cbEncrypt clientSecret (fromMaybe (C.reverseNonce clientNonce) nonce_) r' paddedProxiedTLength
                 let fr = FwdResponse {fwdCorrId, fwdResponse = r2}
-                pure $ RRES $ EncFwdResponse $ C.cbEncryptNoPad sessSecret (C.reverseNonce proxyNonce) (smpEncode fr)
+                pure $ RRES nonce_ $ EncFwdResponse $ C.cbEncryptNoPad sessSecret (C.reverseNonce proxyNonce) (smpEncode fr)
           -- the inner response, or Nothing if forked (RSLV).
           r_ <- lift (rejectOrVerify clntThAuth t') >>= \case
             -- rejectOrVerify filters allowed commands, no need to repeat it here.
@@ -2152,14 +2182,18 @@ client
               Cmd SResolver (RSLV d) -> lift $ rslvNamesEnv >>= \case
                 Nothing -> pure $ Just (corrId', entId', ERR (NAME NO_RESOLVER))
                 Just nenv -> forkCmd serverResolverConcurrency corrId NoEntity $ do
-                  msg <- resolveNameMsg nenv d
+                  msg <- resolveNameMsg (thVersion clntTHParams) nenv d
                   either ERR id <$> runExceptT (encodeResp (corrId', entId', msg))
               -- INTERNAL because processCommand never returns Nothing for sender commands;
               -- `fst` drops the empty message only returned for SUB.
               _ -> Just . maybe (corrId', entId', ERR INTERNAL) fst <$> lift (processCommand Nothing (Right (M.empty, M.empty, M.empty)) t'')
           stats <- asks serverStats
           incStat $ pMsgFwdsRecv stats
-          traverse encodeResp r_
+          traverse encodeResp r_ >>= \case
+            Just rres | any isAuthError r_ -> do
+              authDelay <- asks $ authErrorDelay . config
+              Nothing <$ liftIO (sendAfterDelay clnt stats authDelay startedAt ([(corrId, NoEntity, rres)], []))
+            rres_ -> pure rres_
           where
             rejectOrVerify :: Maybe (THandleAuth 'TServer) -> SignedTransmissionOrError ErrorType Cmd -> M s (VerifiedTransmissionOrError s)
             rejectOrVerify clntThAuth = \case
@@ -2185,7 +2219,7 @@ client
           MessageQuota {} -> RcvMsgQuota msgTs'
           where
             encrypt :: KnownNat i => C.MaxLenBS i -> RcvMessage
-            encrypt body = RcvMessage msgId' . EncRcvMsgBody $ C.cbEncryptMaxLenBS (rcvDhSecret qr) (C.cbNonce msgId') body
+            encrypt body = RcvMessage msgId' . EncRcvMsgBody $ C.cbEncryptMaxLenBS (rcvDhSecret qr) msgId' body
             msgId' = messageId msg
             msgTs' = messageTs msg
 
@@ -2234,7 +2268,7 @@ client
                     SubPending -> QSubPending
                     SubThread _ -> QSubThread
                 ProhibitSub -> pure QProhibitSub
-              qDelivered <- decodeLatin1 . encode . fst <$$> readTVarIO delivered
+              qDelivered <- decodeLatin1 . encode . unMsgId . fst <$$> readTVarIO delivered
               pure QSub {qSubThread, qDelivered}
 
         ok :: Transmission BrokerMsg

src/Simplex/Messaging/Server/Env/STM.hs

@@ -68,6 +68,7 @@ module Simplex.Messaging.Server.Env.STM
     defaultInactiveClientExpiration,
     defaultProxyClientConcurrency,
     defaultNameResolverConcurrency,
+    defaultAuthErrorDelayMs,
     defaultMaxJournalMsgCount,
     defaultMaxJournalStateLines,
     defaultIdleQueueInterval,
@@ -187,6 +188,7 @@ data ServerConfig s = ServerConfig
     ntfDeliveryInterval :: Int,
     -- | interval between sending pending END events to unsubscribed clients, seconds
     pendingENDInterval :: Int,
+    authErrorDelay :: Int64,
     smpCredentials :: ServerCredentials,
     httpCredentials :: Maybe ServerCredentials,
     -- | SMP client-server protocol version range
@@ -255,6 +257,9 @@ defaultProxyClientConcurrency = 32
 defaultNameResolverConcurrency :: Int
 defaultNameResolverConcurrency = 1000
 
+defaultAuthErrorDelayMs :: Int64
+defaultAuthErrorDelayMs = 50
+
 journalMsgStoreDepth :: Int
 journalMsgStoreDepth = 5
 
@@ -571,7 +576,10 @@ newProhibitedSub = do
   return Sub {subThread = ProhibitSub, delivered}
 
 newEnv :: ServerConfig s -> IO (Env s)
-newEnv config@ServerConfig {smpCredentials, httpCredentials, serverStoreCfg, smpAgentCfg, information, messageExpiration, idleQueueInterval, msgQueueQuota, maxJournalMsgCount, maxJournalStateLines, namesConfig} = do
+newEnv config@ServerConfig {smpCredentials, httpCredentials, serverStoreCfg, smpAgentCfg, information, messageExpiration, idleQueueInterval, msgQueueQuota, maxJournalMsgCount, maxJournalStateLines, namesConfig, msgIdBytes} = do
+  when (msgIdBytes < 24) $ do
+    logError "Error: msgIdBytes must be at least 24"
+    exitFailure
   serverActive <- newTVarIO True
   server <- newServer
   msgStore_ <- case serverStoreCfg of

src/Simplex/Messaging/Server/Main.hs

@@ -580,6 +580,7 @@ smpServerCLI_ generateSite serveStaticFiles attachStaticFiles cfgPath logPath =
               prometheusMetricsFile = combine logPath "smp-server-metrics.txt",
               pendingENDInterval = 15000000, -- 15 seconds
               ntfDeliveryInterval = 1500000, -- 1.5 second
+              authErrorDelay = 1000 * readIniDefault defaultAuthErrorDelayMs "TRANSPORT" "auth_error_delay_ms" ini,
               smpServerVRange = supportedServerSMPRelayVRange,
               transportConfig =
                 mkTransportServerConfig

src/Simplex/Messaging/Server/Main/Init.hs

@@ -130,7 +130,9 @@ iniFileContent cfgPath logPath opts host basicAuth controlPortPwds =
     <> ("host = " <> T.pack host <> "\n")
     <> ("port = " <> defaultServerPorts <> "\n")
     <> "log_tls_errors = off\n\n\
-        \# Use `websockets = 443` to run websockets server in addition to plain TLS.\n\
+        \# Minimum time to respond with AUTH error, milliseconds.\n"
+    <> ("# auth_error_delay_ms = " <> tshow defaultAuthErrorDelayMs <> "\n\n")
+    <> "# Use `websockets = 443` to run websockets server in addition to plain TLS.\n\
         \# This option is deprecated and should be used for testing only.\n\
         \# , port 443 should be specified in port above\n\
         \websockets = off\n"

src/Simplex/Messaging/Server/MsgStore/Postgres.hs

@@ -110,12 +110,13 @@ instance MsgStoreClass PostgresMsgStore where
   expireOldMessages :: Bool -> PostgresMsgStore -> Int64 -> Int64 -> IO MessageStats
   expireOldMessages _tty ms now ttl =
     maybeFirstRow' newMessageStats toMessageStats $ withConnection st $ \db ->
-      DB.query db "CALL expire_old_messages(?,?,?,0,0,0)" (oldQueue, oldMsg, batchSize)
+      DB.query db "CALL expire_old_messages(?,?,0,0,0)" (oldMsg, batchSize)
     where
       st = dbStore $ queueStore_ ms
-      oldQueue = 0 :: Int64 -- expire all queues
       oldMsg = now - ttl
-      batchSize = 10000 :: Int
+      -- expired messages read per page in expire_old_messages, and the page is one
+      -- transaction: queues in it stay row-locked against SEND and ACK until it commits.
+      batchSize = 100 :: Int
       toMessageStats (expiredMsgsCount, storedMsgsCount, storedQueues) =
         MessageStats {expiredMsgsCount, storedMsgsCount, storedQueues}
 
@@ -201,7 +202,7 @@ instance MsgStoreClass PostgresMsgStore where
           DB.query
             db
             "SELECT quota_written, was_empty FROM write_message(?,?,?,?,?,?,?)"
-            (recipientId' q, Binary (messageId msg), systemSeconds (messageTs msg), msgQuota, ntf, Binary body, quota)
+            (recipientId' q, messageId msg, systemSeconds (messageTs msg), msgQuota, ntf, Binary body, quota)
     where
       toResult = \case
         ((msgQuota, wasEmpty) : _) -> if msgQuota then Nothing else Just (msg, wasEmpty)
@@ -270,14 +271,14 @@ instance MsgStoreClass PostgresMsgStore where
     uninterruptibleMask_ $
       withDB' "tryDelMsg" (queueStore_ ms) $ \db ->
         maybeFirstRow toMessage $
-          DB.query db "SELECT r_msg_id, r_msg_ts, r_msg_quota, r_msg_ntf_flag, r_msg_body FROM try_del_msg(?, ?)" (recipientId' q, Binary msgId)
+          DB.query db "SELECT r_msg_id, r_msg_ts, r_msg_quota, r_msg_ntf_flag, r_msg_body FROM try_del_msg(?, ?)" (recipientId' q, msgId)
 
   tryDelPeekMsg :: PostgresMsgStore -> PostgresQueue -> MsgId -> ExceptT ErrorType IO (Maybe Message, Maybe Message)
   tryDelPeekMsg ms q msgId =
     uninterruptibleMask_ $
       withDB' "tryDelPeekMsg" (queueStore_ ms) $ \db ->
         toResult . map toMessage
-          <$> DB.query db "SELECT r_msg_id, r_msg_ts, r_msg_quota, r_msg_ntf_flag, r_msg_body FROM try_del_peek_msg(?, ?)" (recipientId' q, Binary msgId)
+          <$> DB.query db "SELECT r_msg_id, r_msg_ts, r_msg_quota, r_msg_ntf_flag, r_msg_body FROM try_del_peek_msg(?, ?)" (recipientId' q, msgId)
     where
       toResult = \case
         [] -> (Nothing, Nothing)
@@ -296,13 +297,13 @@ uninterruptibleMask_ :: ExceptT ErrorType IO a -> ExceptT ErrorType IO a
 uninterruptibleMask_ = ExceptT . E.uninterruptibleMask_ . runExceptT
 {-# INLINE uninterruptibleMask_ #-}
 
-toMaybeMessage :: (Maybe (Binary MsgId), Maybe Int64, Maybe Bool, Maybe Bool, Maybe (Binary MsgBody)) -> Maybe Message
+toMaybeMessage :: (Maybe MsgId, Maybe Int64, Maybe Bool, Maybe Bool, Maybe (Binary MsgBody)) -> Maybe Message
 toMaybeMessage = \case
   (Just msgId, Just ts, Just msgQuota, Just ntf, Just body) -> Just $ toMessage (msgId, ts, msgQuota, ntf, body)
   _ -> Nothing
 
-toMessage :: (Binary MsgId, Int64, Bool, Bool, Binary MsgBody) -> Message
-toMessage (Binary msgId, ts, msgQuota, ntf, Binary body)
+toMessage :: (MsgId, Int64, Bool, Bool, Binary MsgBody) -> Message
+toMessage (msgId, ts, msgQuota, ntf, Binary body)
   | msgQuota = MessageQuota {msgId, msgTs}
   | otherwise = Message {msgId, msgTs, msgFlags = MsgFlags ntf, msgBody = C.unsafeMaxLenBS body} -- TODO [messages] unsafeMaxLenBS?
   where
@@ -364,8 +365,8 @@ deleteAllMessages ms =
       db
       [sql|
         UPDATE msg_queues
-        SET msg_queue_size = 0, msg_can_write = TRUE, msg_queue_expire = FALSE
-        WHERE msg_queue_size != 0 OR msg_can_write = FALSE OR msg_queue_expire = TRUE
+        SET msg_queue_size = 0, msg_can_write = TRUE
+        WHERE msg_queue_size != 0 OR msg_can_write = FALSE
       |]
 
 updateQueueCounts :: PostgresMsgStore -> IO ()
@@ -385,16 +386,15 @@ updateQueueCounts ms =
       db
       [sql|
         UPDATE msg_queues
-        SET msg_queue_size = 0, msg_can_write = TRUE, msg_queue_expire = FALSE
-        WHERE msg_queue_size != 0 OR msg_can_write = FALSE OR msg_queue_expire = TRUE
+        SET msg_queue_size = 0, msg_can_write = TRUE
+        WHERE msg_queue_size != 0 OR msg_can_write = FALSE
       |]
     void $ DB.execute_
       db
       [sql|
         UPDATE msg_queues q
         SET msg_queue_size = s.size,
-            msg_can_write = s.quota_count = 0,
-            msg_queue_expire = s.size > s.quota_count
+            msg_can_write = s.quota_count = 0
         FROM queue_stats s
         WHERE q.recipient_id = s.recipient_id
       |]
@@ -429,7 +429,7 @@ messageRecToText rId msg =
     tabFields = BB.char7 ',' `intersperse` fields
     fields =
       [ renderField (toField rId),
-        renderField (toField $ Binary (messageId msg)),
+        renderField (toField $ messageId msg),
         renderField (toField $ systemSeconds (messageTs msg)),
         renderField (toField msgQuota),
         renderField (toField ntf),

src/Simplex/Messaging/Server/Names.hs

@@ -20,7 +20,9 @@ import Control.Logger.Simple (logError)
 import Data.Bifunctor (first)
 import Data.Maybe (fromMaybe)
 import qualified Data.Text as T
-import Simplex.Messaging.Protocol (NameErrorType (..), NameRecord)
+import Data.Text.Encoding (decodeLatin1)
+import Simplex.Messaging.Encoding
+import Simplex.Messaging.Protocol (NameErrorType (..), NameQuery, NameResponse)
 import Simplex.Messaging.Server.Names.HttpResolver
   ( ResolverEnv,
     ResolverError (..),
@@ -30,7 +32,6 @@ import Simplex.Messaging.Server.Names.HttpResolver
     newResolverEnv,
     resolveHttp,
   )
-import Simplex.Messaging.SimplexName (SimplexDomain, fullDomainName)
 import System.Timeout (timeout)
 
 data NamesConfig = NamesConfig
@@ -58,9 +59,9 @@ pingEndpoint :: NamesEnv -> IO (Either ResolverError ())
 pingEndpoint NamesEnv {resolverEnv, config} =
   fromMaybe (Left ResolverTimeout) <$> timeout (resolverTimeoutMs config * 1000) (healthHttp resolverEnv)
 
-resolveName :: NamesEnv -> SimplexDomain -> IO (Either NameErrorType NameRecord)
-resolveName env d = do
-  r <- E.try (timeout (resolverTimeoutMs (config env) * 1000) (fetch env d))
+resolveName :: NamesEnv -> NameQuery -> IO (Either NameErrorType NameResponse)
+resolveName env q = do
+  r <- E.try (timeout (resolverTimeoutMs (config env) * 1000) (fetch env q))
   case r of
     Right result -> pure (fromMaybe (Left (RESOLVER "timeout")) result)
     Left e
@@ -69,14 +70,12 @@ resolveName env d = do
           logError $ "[NAMES] resolver fetch raised " <> T.pack (E.displayException e)
           pure (Left (RESOLVER "resolver error"))
 
-fetch :: NamesEnv -> SimplexDomain -> IO (Either NameErrorType NameRecord)
-fetch NamesEnv {resolverEnv} d =
-  first mapResolverError <$> resolveHttp resolverEnv (fullDomainName d)
+fetch :: NamesEnv -> NameQuery -> IO (Either NameErrorType NameResponse)
+fetch NamesEnv {resolverEnv} q =
+  first mapResolverError <$> resolveHttp resolverEnv (decodeLatin1 $ smpEncode q)
 
 mapResolverError :: ResolverError -> NameErrorType
 mapResolverError = \case
-  HttpStatusErr 404 -> NOT_FOUND
-  HttpStatusErr 400 -> NOT_FOUND
   HttpStatusErr code -> RESOLVER ("HTTP " <> T.pack (show code))
   HttpFailure _ -> RESOLVER "transport failure"
   BodyTooLarge -> RESOLVER "response too large"

src/Simplex/Messaging/Server/Names/HttpResolver.hs

@@ -8,10 +8,12 @@
 --
 -- The Python REST resolver (see scripts/resolver/snrc-resolve.py) exposes
 --
---   GET /resolve/<name>   -> 200 with a NameRecord JSON document
---                            404 / 400 for unknown names / TLDs
---                            502 for upstream RPC failures
---   GET /health           -> 200 when the resolver process is ready
+--   GET /v2/resolve/<query> -> 200 with a NameRegistration JSON document, for
+--                              all three registration shapes; 400 for unknown
+--                              TLDs, 502 for upstream RPC failures
+--   GET /v1/resolve/<name>  -> 200 with a NameRecord, what relays before SMP
+--                              v22 call as /resolve
+--   GET /health             -> 200 when the resolver process is ready
 --
 -- Boundary properties:
 --   * Response body read with `brReadSome maxResponseBytes` — adversarial
@@ -57,7 +59,7 @@ import qualified Network.HTTP.Client as HC
 import Network.HTTP.Client.TLS (tlsManagerSettings)
 import qualified Network.HTTP.Types as HT
 import Network.HTTP.Types.URI (urlEncode)
-import Simplex.Messaging.Names.Record (NameRecord)
+import Simplex.Messaging.Names.Record (NameResponse)
 
 data RpcAuth = AuthBearer Text | AuthBasic Text Text
 
@@ -108,14 +110,12 @@ authHeader = \case
     let encoded = BAE.convertToBase BAE.Base64 (encodeUtf8 u <> ":" <> encodeUtf8 p) :: ByteString
      in ("Authorization", "Basic " <> encoded)
 
--- | GET <baseUrl>/resolve/<percent-encoded name>, decoding the 200 body
--- directly into a NameRecord in one pass (no intermediate Aeson Value). The
--- name is percent-encoded (every non-unreserved byte per RFC 3986): the
--- resolver expects raw labels, so slashes/punctuation must not alter the path.
-resolveHttp :: ResolverEnv -> Text -> IO (Either ResolverError NameRecord)
-resolveHttp env name =
+-- | The query is a name or a bracketed label hash, percent-encoded (every
+-- non-unreserved byte per RFC 3986) so it cannot alter the path.
+resolveHttp :: ResolverEnv -> Text -> IO (Either ResolverError NameResponse)
+resolveHttp env q =
   (>>= first InvalidJson . J.eitherDecodeStrict . BL.toStrict)
-    <$> httpGet env ("/resolve/" <> B.unpack (urlEncode True (encodeUtf8 name)))
+    <$> httpGet env ("/v2/resolve/" <> B.unpack (urlEncode True (encodeUtf8 q)))
 
 -- | GET <baseUrl>/health; success = reachable with status < 400. The body is
 -- size-capped but NOT decoded — the probe only checks reachability.

src/Simplex/Messaging/Server/Prometheus.hs

@@ -126,6 +126,7 @@ prometheusMetrics sm rtm ts =
         _ntfServices,
         _rcvServicesSubMsg,
         _rcvServicesSubDuplicate,
+        _authDelayExceeded,
         _qCount,
         _msgCount,
         _ntfCount,
@@ -192,6 +193,10 @@ prometheusMetrics sm rtm ts =
       \# HELP simplex_smp_queues_notify_monthly Monthly active queues with notifications.\n\
       \# TYPE simplex_smp_queues_notify_monthly gauge\n\
       \simplex_smp_queues_notify_monthly " <> mstr (monthCount psNtf) <> "\n# monthCountNtf\n\
+      \\n\
+      \# HELP simplex_smp_auth_delay_exceeded Responses with AUTH errors sent without delay, as processing took longer than the delay.\n\
+      \# TYPE simplex_smp_auth_delay_exceeded counter\n\
+      \simplex_smp_auth_delay_exceeded " <> mshow _authDelayExceeded <> "\n# authDelayExceeded\n\
       \\n"
     subscriptions =
       "# Subscriptions\n\
@@ -469,11 +474,11 @@ prometheusMetrics sm rtm ts =
           \# TYPE simplex_smp_names_reqs counter\n\
           \simplex_smp_names_reqs " <> mshow _rslvReqs <> "\n# rslvReqs\n\
           \\n\
-          \# HELP simplex_smp_names_success NameRecord successfully resolved and returned.\n\
+          \# HELP simplex_smp_names_success NameRecord resolved, or availability answered.\n\
           \# TYPE simplex_smp_names_success counter\n\
           \simplex_smp_names_success " <> mshow _rslvSucc <> "\n# rslvSucc\n\
           \\n\
-          \# HELP simplex_smp_names_not_found Name not registered (resolver returned 404 / 400).\n\
+          \# HELP simplex_smp_names_not_found Answers a client below v22 reads as NOT_FOUND.\n\
           \# TYPE simplex_smp_names_not_found counter\n\
           \simplex_smp_names_not_found " <> mshow _rslvNotFound <> "\n# rslvNotFound\n\
           \\n\

src/Simplex/Messaging/Server/QueueStore.hs

@@ -13,12 +13,14 @@ module Simplex.Messaging.Server.QueueStore
     ServiceRec (..),
     CertFingerprint,
     ServerEntityStatus (..),
+    isSecuredMsgQueue,
   ) where
 
 import Control.Applicative (optional, (<|>))
 import qualified Data.ByteString.Char8 as B
 import Data.Functor (($>))
 import Data.List.NonEmpty (NonEmpty)
+import Data.Maybe (isJust)
 import qualified Data.X509 as X
 import qualified Data.X509.Validation as XV
 import Simplex.Messaging.Encoding
@@ -48,6 +50,11 @@ data QueueRec = QueueRec
   }
   deriving (Show)
 
+isSecuredMsgQueue :: QueueRec -> Bool
+isSecuredMsgQueue QueueRec {queueMode, senderKey} = case queueMode of
+  Just QMContact -> False
+  _ -> isJust senderKey
+
 data NtfCreds = NtfCreds
   { notifierId :: NotifierId,
     notifierKey :: NtfPublicAuthKey,

src/Simplex/Messaging/Server/QueueStore/Postgres.hs

@@ -159,7 +159,8 @@ instance StoreQueueClass q => QueueStoreClass q (PostgresQueueStore q) where
           db
           [sql|
             SELECT
-              (SELECT COUNT(1) FROM msg_queues WHERE deleted_at IS NULL) AS queue_count,
+              -- estimate via reltuples to avoid a full heap scan on every scrape
+              (SELECT GREATEST(reltuples, 0)::bigint FROM pg_class WHERE oid = 'msg_queues'::regclass) AS queue_count,
               (SELECT COUNT(1) FROM msg_queues WHERE deleted_at IS NULL AND notifier_id IS NOT NULL) AS notifier_count,
               (SELECT COUNT(1) FROM services WHERE service_role = ?) AS rcv_service_count,
               (SELECT COUNT(1) FROM services WHERE service_role = ?) AS ntf_service_count,
@@ -313,9 +314,9 @@ instance StoreQueueClass q => QueueStoreClass q (PostgresQueueStore q) where
   addQueueLinkData st sq lnkId d =
     withQueueRec sq "addQueueLinkData" $ \q -> case queueData q of
       Nothing ->
-        addLink q $ \db -> DB.execute db qry (d :. (lnkId, rId))
+        addLink q $ \db -> DB.execute db qry (d :. (lnkId, rId, QMContact))
       Just (lnkId', _) | lnkId' == lnkId ->
-        addLink q $ \db -> DB.execute db (qry <> " AND (fixed_data IS NULL OR fixed_data = ?)") (d :. (lnkId, rId, fst d))
+        addLink q $ \db -> DB.execute db (qry <> " AND (fixed_data IS NULL OR fixed_data = ?)") (d :. (lnkId, rId, QMContact, fst d))
       _ -> throwE AUTH
     where
       rId = recipientId sq
@@ -323,7 +324,8 @@ instance StoreQueueClass q => QueueStoreClass q (PostgresQueueStore q) where
         assertUpdated $ withDB' "addQueueLinkData" st update
         atomically $ writeTVar (queueRec sq) $ Just q {queueData = Just (lnkId, d)}
         withLog "addQueueLinkData" st $ \s -> logCreateLink s rId lnkId d
-      qry = "UPDATE msg_queues SET fixed_data = ?, user_data = ?, link_id = ? WHERE recipient_id = ? AND deleted_at IS NULL"
+      -- the sender key condition is checked in SQL because without cache each command reads its own copy of the queue record
+      qry = "UPDATE msg_queues SET fixed_data = ?, user_data = ?, link_id = ? WHERE recipient_id = ? AND deleted_at IS NULL AND (sender_key IS NULL OR queue_mode = ?)"
 
   deleteQueueLinkData :: PostgresQueueStore q -> q -> IO (Either ErrorType ())
   deleteQueueLinkData st sq =
@@ -823,6 +825,10 @@ instance ToField EncDataBytes where toField (EncDataBytes s) = toField (Binary s
 
 deriving newtype instance FromField EncDataBytes
 
+instance ToField MsgId where toField = toField . Binary . unMsgId
+
+instance FromField MsgId where fromField = blobFieldDecoder mkMsgId
+
 deriving newtype instance ToField (RoundedSystemTime t)
 
 deriving newtype instance FromField (RoundedSystemTime t)

src/Simplex/Messaging/Server/QueueStore/Postgres/Migrations.hs

@@ -20,7 +20,10 @@ serverSchemaMigrations =
     ("20250320_short_links", m20250320_short_links, Just down_m20250320_short_links),
     ("20250514_service_certs", m20250514_service_certs, Just down_m20250514_service_certs),
     ("20250903_store_messages", m20250903_store_messages, Just down_m20250903_store_messages),
-    ("20250915_queue_ids_hash", m20250915_queue_ids_hash, Just down_m20250915_queue_ids_hash)
+    ("20250915_queue_ids_hash", m20250915_queue_ids_hash, Just down_m20250915_queue_ids_hash),
+    ("20260916_prometheus_indexes", m20260916_prometheus_indexes, Just down_m20260916_prometheus_indexes),
+    ("20260917_msg_queues_hot", m20260917_msg_queues_hot, Just down_m20260917_msg_queues_hot),
+    ("20260918_expire_messages", m20260918_expire_messages, Just down_m20260918_expire_messages)
   ]
 
 -- | The list of migrations in ascending order by date
@@ -588,3 +591,648 @@ ALTER TABLE services
   DROP COLUMN queue_ids_hash;
     |]
     <> dropXorHashFuncs
+
+m20260916_prometheus_indexes :: Text
+m20260916_prometheus_indexes =
+  [r|
+CREATE INDEX idx_msg_queues_expire ON msg_queues (recipient_id) WHERE deleted_at IS NULL AND msg_queue_expire;
+CREATE INDEX idx_msg_queues_notifier_active ON msg_queues (notifier_id) WHERE deleted_at IS NULL AND notifier_id IS NOT NULL;
+    |]
+
+down_m20260916_prometheus_indexes :: Text
+down_m20260916_prometheus_indexes =
+  [r|
+DROP INDEX idx_msg_queues_expire;
+DROP INDEX idx_msg_queues_notifier_active;
+    |]
+
+m20260917_msg_queues_hot :: Text
+m20260917_msg_queues_hot =
+  [r|
+DROP INDEX idx_msg_queues_updated_at_recipient_id;
+
+ALTER TABLE msg_queues SET (
+  fillfactor = 80,
+  autovacuum_vacuum_scale_factor = 0.02,
+  autovacuum_analyze_scale_factor = 0.01,
+  autovacuum_vacuum_cost_limit = 1000
+);
+
+ALTER TABLE messages SET (
+  autovacuum_vacuum_scale_factor = 0.02,
+  autovacuum_analyze_scale_factor = 0.01,
+  toast.autovacuum_vacuum_scale_factor = 0.02
+);
+
+ALTER TABLE services SET (
+  fillfactor = 70,
+  autovacuum_vacuum_threshold = 1000,
+  autovacuum_vacuum_scale_factor = 0
+);
+
+DROP PROCEDURE expire_old_messages(bigint, bigint, integer);
+
+CREATE PROCEDURE expire_old_messages(IN p_old_ts bigint, IN batch_size integer, OUT r_expired_msgs_count bigint, OUT r_stored_msgs_count bigint, OUT r_stored_queues bigint)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  rids BYTEA[];
+  rid BYTEA;
+  last_rid BYTEA := '\x';
+  del_count BIGINT;
+  total_deleted BIGINT := 0;
+BEGIN
+  LOOP
+    SELECT array_agg(recipient_id)
+    INTO rids
+    FROM (
+      SELECT recipient_id
+      FROM msg_queues
+      WHERE deleted_at IS NULL
+        AND msg_queue_expire = TRUE
+        AND recipient_id > last_rid
+      ORDER BY recipient_id ASC
+      LIMIT batch_size
+    ) qs;
+
+    EXIT WHEN rids IS NULL OR cardinality(rids) = 0;
+
+    FOREACH rid IN ARRAY rids
+    LOOP
+      BEGIN
+        del_count := delete_expired_msgs(rid, p_old_ts);
+        total_deleted := total_deleted + del_count;
+      EXCEPTION WHEN OTHERS THEN
+        RAISE WARNING 'STORE, expire_old_messages, error expiring queue %: %', encode(rid, 'base64'), SQLERRM;
+        CONTINUE;
+      END;
+      COMMIT;
+    END LOOP;
+    last_rid := rids[cardinality(rids)];
+  END LOOP;
+
+  r_expired_msgs_count := total_deleted;
+  r_stored_msgs_count := (SELECT COUNT(1) FROM messages);
+  r_stored_queues := (SELECT COUNT(1) FROM msg_queues WHERE deleted_at IS NULL);
+END;
+$$;
+    |]
+
+down_m20260917_msg_queues_hot :: Text
+down_m20260917_msg_queues_hot =
+  [r|
+DROP PROCEDURE expire_old_messages(bigint, integer);
+
+CREATE PROCEDURE expire_old_messages(IN p_old_queue bigint, IN p_old_ts bigint, IN batch_size integer, OUT r_expired_msgs_count bigint, OUT r_stored_msgs_count bigint, OUT r_stored_queues bigint)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  rids BYTEA[];
+  rid BYTEA;
+  last_rid BYTEA := '\x';
+  del_count BIGINT;
+  total_deleted BIGINT := 0;
+BEGIN
+  LOOP
+    SELECT array_agg(recipient_id)
+    INTO rids
+    FROM (
+      SELECT recipient_id
+      FROM msg_queues
+      WHERE deleted_at IS NULL
+        AND updated_at > p_old_queue
+        AND msg_queue_expire = TRUE
+        AND recipient_id > last_rid
+      ORDER BY recipient_id ASC
+      LIMIT batch_size
+    ) qs;
+
+    EXIT WHEN rids IS NULL OR cardinality(rids) = 0;
+
+    FOREACH rid IN ARRAY rids
+    LOOP
+      BEGIN
+        del_count := delete_expired_msgs(rid, p_old_ts);
+        total_deleted := total_deleted + del_count;
+      EXCEPTION WHEN OTHERS THEN
+        RAISE WARNING 'STORE, expire_old_messages, error expiring queue %: %', encode(rid, 'base64'), SQLERRM;
+        CONTINUE;
+      END;
+      COMMIT;
+    END LOOP;
+    last_rid := rids[cardinality(rids)];
+  END LOOP;
+
+  r_expired_msgs_count := total_deleted;
+  r_stored_msgs_count := (SELECT COUNT(1) FROM messages);
+  r_stored_queues := (SELECT COUNT(1) FROM msg_queues WHERE deleted_at IS NULL);
+END;
+$$;
+
+CREATE INDEX idx_msg_queues_updated_at_recipient_id ON msg_queues (deleted_at, updated_at, msg_queue_expire, recipient_id);
+
+ALTER TABLE msg_queues RESET (fillfactor, autovacuum_vacuum_scale_factor, autovacuum_analyze_scale_factor, autovacuum_vacuum_cost_limit);
+ALTER TABLE messages RESET (autovacuum_vacuum_scale_factor, autovacuum_analyze_scale_factor, toast.autovacuum_vacuum_scale_factor);
+ALTER TABLE services RESET (fillfactor, autovacuum_vacuum_threshold, autovacuum_vacuum_scale_factor);
+    |]
+
+m20260918_expire_messages :: Text
+m20260918_expire_messages =
+  [r|
+CREATE INDEX idx_messages_expire ON messages (msg_ts, recipient_id) WHERE NOT msg_quota;
+
+DROP INDEX idx_messages_recipient_id_msg_ts;
+
+DROP INDEX idx_msg_queues_expire;
+
+DROP PROCEDURE expire_old_messages(bigint, integer);
+
+CREATE PROCEDURE expire_old_messages(IN p_old_ts bigint, IN batch_size integer, OUT r_expired_msgs_count bigint, OUT r_stored_msgs_count bigint, OUT r_stored_queues bigint)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  rids BYTEA[];
+  rid BYTEA;
+  last_ts BIGINT := -1;
+  last_rid BYTEA := '\x';
+  next_ts BIGINT;
+  next_rid BYTEA;
+  del_count BIGINT;
+  total_deleted BIGINT := 0;
+BEGIN
+  LOOP
+    -- The page is scanned in (msg_ts, recipient_id) order, so its last row is the next
+    -- keyset cursor. Advancing it past every row read, including queues left unexpired
+    -- because delete_expired_msgs skipped a locked row or raised, is what terminates the
+    -- loop; re-reading from the start instead would repeat those queues forever.
+    SELECT array_agg(DISTINCT recipient_id),
+           (array_agg(msg_ts ORDER BY msg_ts DESC, recipient_id DESC))[1],
+           (array_agg(recipient_id ORDER BY msg_ts DESC, recipient_id DESC))[1]
+    INTO rids, next_ts, next_rid
+    FROM (
+      SELECT msg_ts, recipient_id
+      FROM messages
+      WHERE NOT msg_quota
+        AND msg_ts < p_old_ts
+        AND (msg_ts, recipient_id) > (last_ts, last_rid)
+      ORDER BY msg_ts ASC, recipient_id ASC
+      LIMIT batch_size
+    ) m;
+
+    EXIT WHEN rids IS NULL;
+
+    FOREACH rid IN ARRAY rids
+    LOOP
+      BEGIN
+        del_count := delete_expired_msgs(rid, p_old_ts);
+        total_deleted := total_deleted + del_count;
+      EXCEPTION WHEN OTHERS THEN
+        RAISE WARNING 'STORE, expire_old_messages, error expiring queue %: %', encode(rid, 'base64'), SQLERRM;
+        CONTINUE;
+      END;
+    END LOOP;
+    COMMIT;
+
+    last_ts := next_ts;
+    last_rid := next_rid;
+  END LOOP;
+
+  r_expired_msgs_count := total_deleted;
+  r_stored_msgs_count := (SELECT COUNT(1) FROM messages);
+  r_stored_queues := (SELECT COUNT(1) FROM msg_queues WHERE deleted_at IS NULL);
+END;
+$$;
+
+CREATE OR REPLACE FUNCTION delete_expired_msgs(p_recipient_id bytea, p_old_ts bigint) RETURNS bigint
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  q_size BIGINT;
+  keep_min_id BIGINT;
+  del_count BIGINT;
+BEGIN
+  SELECT msg_queue_size INTO q_size
+  FROM msg_queues
+  WHERE recipient_id = p_recipient_id AND deleted_at IS NULL
+  FOR UPDATE SKIP LOCKED;
+
+  IF NOT FOUND OR q_size = 0 THEN
+    RETURN 0;
+  END IF;
+
+  SELECT MIN(message_id) INTO keep_min_id
+  FROM messages WHERE recipient_id = p_recipient_id AND msg_ts >= p_old_ts AND msg_quota = FALSE;
+
+  IF keep_min_id IS NULL THEN
+    DELETE FROM messages WHERE recipient_id = p_recipient_id AND msg_quota = FALSE;
+  ELSE
+    DELETE FROM messages WHERE recipient_id = p_recipient_id AND message_id < keep_min_id AND msg_quota = FALSE;
+  END IF;
+
+  GET DIAGNOSTICS del_count = ROW_COUNT;
+  IF del_count > 0 THEN
+    UPDATE msg_queues
+    SET msg_can_write = msg_can_write OR msg_queue_size <= del_count,
+        msg_queue_size = GREATEST(msg_queue_size - del_count, 0)
+    WHERE recipient_id = p_recipient_id;
+  END IF;
+  RETURN del_count;
+END;
+$$;
+
+CREATE OR REPLACE FUNCTION write_message(p_recipient_id bytea, p_msg_id bytea, p_msg_ts bigint, p_msg_quota boolean, p_msg_ntf_flag boolean, p_msg_body bytea, p_quota integer) RETURNS TABLE(quota_written boolean, was_empty boolean)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  q_can_write BOOLEAN;
+  q_size BIGINT;
+BEGIN
+  SELECT msg_can_write, msg_queue_size INTO q_can_write, q_size
+  FROM msg_queues
+  WHERE recipient_id = p_recipient_id AND deleted_at IS NULL
+  FOR UPDATE;
+
+  IF q_can_write OR q_size = 0 THEN
+    quota_written := p_msg_quota OR q_size >= p_quota;
+    was_empty := q_size = 0;
+
+    INSERT INTO messages(recipient_id, msg_id, msg_ts, msg_quota, msg_ntf_flag, msg_body)
+    VALUES (p_recipient_id, p_msg_id, p_msg_ts, quota_written, p_msg_ntf_flag AND NOT quota_written, CASE WHEN quota_written THEN '' :: BYTEA ELSE p_msg_body END);
+
+    UPDATE msg_queues
+    SET msg_can_write = NOT quota_written,
+        msg_queue_size = msg_queue_size + 1
+    WHERE recipient_id = p_recipient_id;
+
+    RETURN QUERY VALUES (quota_written, was_empty);
+  END IF;
+END;
+$$;
+
+CREATE OR REPLACE FUNCTION try_del_msg(p_recipient_id bytea, p_msg_id bytea) RETURNS TABLE(r_msg_id bytea, r_msg_ts bigint, r_msg_quota boolean, r_msg_ntf_flag boolean, r_msg_body bytea)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  q_size BIGINT;
+  msg RECORD;
+BEGIN
+  SELECT msg_queue_size INTO q_size
+  FROM msg_queues
+  WHERE recipient_id = p_recipient_id AND deleted_at IS NULL
+  FOR UPDATE;
+
+  IF NOT FOUND THEN
+    RETURN;
+  END IF;
+
+  SELECT message_id, msg_id, msg_ts, msg_quota, msg_ntf_flag, msg_body
+  INTO msg
+  FROM messages
+  WHERE recipient_id = p_recipient_id
+  ORDER BY message_id ASC LIMIT 1;
+
+  IF NOT FOUND THEN
+    IF q_size != 0 THEN
+      UPDATE msg_queues
+      SET msg_can_write = TRUE,
+          msg_queue_size = 0
+      WHERE recipient_id = p_recipient_id;
+    END IF;
+    RETURN;
+  END IF;
+
+  IF msg.msg_id = p_msg_id THEN
+    DELETE FROM messages WHERE message_id = msg.message_id;
+    IF FOUND THEN
+      UPDATE msg_queues
+      SET msg_can_write = msg_can_write OR msg_queue_size <= 1,
+          msg_queue_size = GREATEST(msg_queue_size - 1, 0)
+      WHERE recipient_id = p_recipient_id;
+      RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
+    END IF;
+  END IF;
+END;
+$$;
+
+CREATE OR REPLACE FUNCTION try_del_peek_msg(p_recipient_id bytea, p_msg_id bytea) RETURNS TABLE(r_msg_id bytea, r_msg_ts bigint, r_msg_quota boolean, r_msg_ntf_flag boolean, r_msg_body bytea)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  q_size BIGINT;
+  msg RECORD;
+  msg_deleted BOOLEAN;
+BEGIN
+  SELECT msg_queue_size INTO q_size
+  FROM msg_queues
+  WHERE recipient_id = p_recipient_id AND deleted_at IS NULL
+  FOR UPDATE;
+
+  IF NOT FOUND THEN
+    RETURN;
+  END IF;
+
+  SELECT message_id, msg_id, msg_ts, msg_quota, msg_ntf_flag, msg_body
+  INTO msg
+  FROM messages
+  WHERE recipient_id = p_recipient_id
+  ORDER BY message_id ASC LIMIT 1;
+
+  IF NOT FOUND THEN
+    IF q_size != 0 THEN
+      UPDATE msg_queues
+      SET msg_can_write = TRUE,
+          msg_queue_size = 0
+      WHERE recipient_id = p_recipient_id;
+    END IF;
+    RETURN;
+  END IF;
+
+  IF msg.msg_id = p_msg_id THEN
+    DELETE FROM messages WHERE message_id = msg.message_id;
+
+    msg_deleted := FOUND;
+    IF msg_deleted THEN
+      RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
+    END IF;
+
+    SELECT msg_id, msg_ts, msg_quota, msg_ntf_flag, msg_body
+    INTO msg
+    FROM messages
+    WHERE recipient_id = p_recipient_id
+    ORDER BY message_id ASC LIMIT 1;
+
+    IF FOUND THEN
+      RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
+      IF msg_deleted THEN
+        UPDATE msg_queues
+        SET msg_can_write = msg_can_write OR msg_queue_size <= 1,
+            msg_queue_size = GREATEST(msg_queue_size - 1, 0)
+        WHERE recipient_id = p_recipient_id;
+      END IF;
+    ELSIF msg_deleted OR q_size != 0 THEN
+      UPDATE msg_queues
+      SET msg_can_write = TRUE,
+          msg_queue_size = 0
+      WHERE recipient_id = p_recipient_id;
+    END IF;
+  ELSE
+    RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
+  END IF;
+END;
+$$;
+
+ALTER TABLE msg_queues DROP COLUMN msg_queue_expire;
+    |]
+
+down_m20260918_expire_messages :: Text
+down_m20260918_expire_messages =
+  [r|
+ALTER TABLE msg_queues ADD COLUMN msg_queue_expire boolean NOT NULL DEFAULT FALSE;
+
+-- ADD COLUMN already defaulted every row to FALSE, so the backfill only needs the rows
+-- that become TRUE, which avoids rewriting the whole table. The flag is only read to pick
+-- queues for delete_expired_msgs, and that returns 0 without deleting when msg_queue_size
+-- is 0, so excluding empty queues here cannot change what expires.
+UPDATE msg_queues q
+SET msg_queue_expire = TRUE
+WHERE msg_queue_size > 0
+  AND EXISTS (SELECT 1 FROM messages m WHERE m.recipient_id = q.recipient_id AND NOT m.msg_quota);
+
+CREATE INDEX idx_msg_queues_expire ON msg_queues (recipient_id) WHERE deleted_at IS NULL AND msg_queue_expire;
+
+CREATE INDEX idx_messages_recipient_id_msg_ts ON messages (recipient_id, msg_ts);
+
+DROP INDEX idx_messages_expire;
+
+CREATE OR REPLACE FUNCTION delete_expired_msgs(p_recipient_id bytea, p_old_ts bigint) RETURNS bigint
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  q_size BIGINT;
+  keep_min_id BIGINT;
+  del_count BIGINT;
+BEGIN
+  SELECT msg_queue_size INTO q_size
+  FROM msg_queues
+  WHERE recipient_id = p_recipient_id AND deleted_at IS NULL
+  FOR UPDATE SKIP LOCKED;
+
+  IF NOT FOUND OR q_size = 0 THEN
+    RETURN 0;
+  END IF;
+
+  SELECT MIN(message_id) INTO keep_min_id
+  FROM messages WHERE recipient_id = p_recipient_id AND msg_ts >= p_old_ts AND msg_quota = FALSE;
+
+  IF keep_min_id IS NULL THEN
+    DELETE FROM messages WHERE recipient_id = p_recipient_id AND msg_quota = FALSE;
+  ELSE
+    DELETE FROM messages WHERE recipient_id = p_recipient_id AND message_id < keep_min_id AND msg_quota = FALSE;
+  END IF;
+
+  GET DIAGNOSTICS del_count = ROW_COUNT;
+  IF del_count > 0 THEN
+    UPDATE msg_queues
+    SET msg_can_write = msg_can_write OR msg_queue_size <= del_count,
+        msg_queue_expire = msg_queue_size > del_count AND keep_min_id IS NOT NULL,
+        msg_queue_size = GREATEST(msg_queue_size - del_count, 0)
+    WHERE recipient_id = p_recipient_id;
+  END IF;
+  RETURN del_count;
+END;
+$$;
+
+CREATE OR REPLACE FUNCTION write_message(p_recipient_id bytea, p_msg_id bytea, p_msg_ts bigint, p_msg_quota boolean, p_msg_ntf_flag boolean, p_msg_body bytea, p_quota integer) RETURNS TABLE(quota_written boolean, was_empty boolean)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  q_can_write BOOLEAN;
+  q_size BIGINT;
+BEGIN
+  SELECT msg_can_write, msg_queue_size INTO q_can_write, q_size
+  FROM msg_queues
+  WHERE recipient_id = p_recipient_id AND deleted_at IS NULL
+  FOR UPDATE;
+
+  IF q_can_write OR q_size = 0 THEN
+    quota_written := p_msg_quota OR q_size >= p_quota;
+    was_empty := q_size = 0;
+
+    INSERT INTO messages(recipient_id, msg_id, msg_ts, msg_quota, msg_ntf_flag, msg_body)
+    VALUES (p_recipient_id, p_msg_id, p_msg_ts, quota_written, p_msg_ntf_flag AND NOT quota_written, CASE WHEN quota_written THEN '' :: BYTEA ELSE p_msg_body END);
+
+    UPDATE msg_queues
+    SET msg_can_write = NOT quota_written,
+        msg_queue_expire = TRUE,
+        msg_queue_size = msg_queue_size + 1
+    WHERE recipient_id = p_recipient_id;
+
+    RETURN QUERY VALUES (quota_written, was_empty);
+  END IF;
+END;
+$$;
+
+CREATE OR REPLACE FUNCTION try_del_msg(p_recipient_id bytea, p_msg_id bytea) RETURNS TABLE(r_msg_id bytea, r_msg_ts bigint, r_msg_quota boolean, r_msg_ntf_flag boolean, r_msg_body bytea)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  q_size BIGINT;
+  msg RECORD;
+BEGIN
+  SELECT msg_queue_size INTO q_size
+  FROM msg_queues
+  WHERE recipient_id = p_recipient_id AND deleted_at IS NULL
+  FOR UPDATE;
+
+  IF NOT FOUND THEN
+    RETURN;
+  END IF;
+
+  SELECT message_id, msg_id, msg_ts, msg_quota, msg_ntf_flag, msg_body
+  INTO msg
+  FROM messages
+  WHERE recipient_id = p_recipient_id
+  ORDER BY message_id ASC LIMIT 1;
+
+  IF NOT FOUND THEN
+    IF q_size != 0 THEN
+      UPDATE msg_queues
+      SET msg_can_write = TRUE,
+          msg_queue_expire = FALSE,
+          msg_queue_size = 0
+      WHERE recipient_id = p_recipient_id;
+    END IF;
+    RETURN;
+  END IF;
+
+  IF msg.msg_id = p_msg_id THEN
+    DELETE FROM messages WHERE message_id = msg.message_id;
+    IF FOUND THEN
+      UPDATE msg_queues
+      SET msg_can_write = msg_can_write OR msg_queue_size <= 1,
+          msg_queue_expire = msg_queue_size > 1,
+          msg_queue_size = GREATEST(msg_queue_size - 1, 0)
+      WHERE recipient_id = p_recipient_id;
+      RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
+    END IF;
+  END IF;
+END;
+$$;
+
+CREATE OR REPLACE FUNCTION try_del_peek_msg(p_recipient_id bytea, p_msg_id bytea) RETURNS TABLE(r_msg_id bytea, r_msg_ts bigint, r_msg_quota boolean, r_msg_ntf_flag boolean, r_msg_body bytea)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  q_size BIGINT;
+  msg RECORD;
+  msg_deleted BOOLEAN;
+BEGIN
+  SELECT msg_queue_size INTO q_size
+  FROM msg_queues
+  WHERE recipient_id = p_recipient_id AND deleted_at IS NULL
+  FOR UPDATE;
+
+  IF NOT FOUND THEN
+    RETURN;
+  END IF;
+
+  SELECT message_id, msg_id, msg_ts, msg_quota, msg_ntf_flag, msg_body
+  INTO msg
+  FROM messages
+  WHERE recipient_id = p_recipient_id
+  ORDER BY message_id ASC LIMIT 1;
+
+  IF NOT FOUND THEN
+    IF q_size != 0 THEN
+      UPDATE msg_queues
+      SET msg_can_write = TRUE,
+          msg_queue_expire = FALSE,
+          msg_queue_size = 0
+      WHERE recipient_id = p_recipient_id;
+    END IF;
+    RETURN;
+  END IF;
+
+  IF msg.msg_id = p_msg_id THEN
+    DELETE FROM messages WHERE message_id = msg.message_id;
+
+    msg_deleted := FOUND;
+    IF msg_deleted THEN
+      RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
+    END IF;
+
+    SELECT msg_id, msg_ts, msg_quota, msg_ntf_flag, msg_body
+    INTO msg
+    FROM messages
+    WHERE recipient_id = p_recipient_id
+    ORDER BY message_id ASC LIMIT 1;
+
+    IF FOUND THEN
+      RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
+      IF msg_deleted THEN
+        UPDATE msg_queues
+        SET msg_can_write = msg_can_write OR msg_queue_size <= 1,
+            msg_queue_expire = msg_queue_size > 1,
+            msg_queue_size = GREATEST(msg_queue_size - 1, 0)
+        WHERE recipient_id = p_recipient_id;
+      END IF;
+    ELSIF msg_deleted OR q_size != 0 THEN
+      UPDATE msg_queues
+      SET msg_can_write = TRUE,
+          msg_queue_expire = FALSE,
+          msg_queue_size = 0
+      WHERE recipient_id = p_recipient_id;
+    END IF;
+  ELSE
+    RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
+  END IF;
+END;
+$$;
+
+DROP PROCEDURE expire_old_messages(bigint, integer);
+
+CREATE PROCEDURE expire_old_messages(IN p_old_ts bigint, IN batch_size integer, OUT r_expired_msgs_count bigint, OUT r_stored_msgs_count bigint, OUT r_stored_queues bigint)
+    LANGUAGE plpgsql
+    AS $$
+DECLARE
+  rids BYTEA[];
+  rid BYTEA;
+  last_rid BYTEA := '\x';
+  del_count BIGINT;
+  total_deleted BIGINT := 0;
+BEGIN
+  LOOP
+    SELECT array_agg(recipient_id)
+    INTO rids
+    FROM (
+      SELECT recipient_id
+      FROM msg_queues
+      WHERE deleted_at IS NULL
+        AND msg_queue_expire = TRUE
+        AND recipient_id > last_rid
+      ORDER BY recipient_id ASC
+      LIMIT batch_size
+    ) qs;
+
+    EXIT WHEN rids IS NULL OR cardinality(rids) = 0;
+
+    FOREACH rid IN ARRAY rids
+    LOOP
+      BEGIN
+        del_count := delete_expired_msgs(rid, p_old_ts);
+        total_deleted := total_deleted + del_count;
+      EXCEPTION WHEN OTHERS THEN
+        RAISE WARNING 'STORE, expire_old_messages, error expiring queue %: %', encode(rid, 'base64'), SQLERRM;
+        CONTINUE;
+      END;
+      COMMIT;
+    END LOOP;
+    last_rid := rids[cardinality(rids)];
+  END LOOP;
+
+  r_expired_msgs_count := total_deleted;
+  r_stored_msgs_count := (SELECT COUNT(1) FROM messages);
+  r_stored_queues := (SELECT COUNT(1) FROM msg_queues WHERE deleted_at IS NULL);
+END;
+$$;
+    |]

src/Simplex/Messaging/Server/QueueStore/Postgres/server_schema.sql

@@ -1,5 +1,6 @@
 
 
+
 SET statement_timeout = 0;
 SET lock_timeout = 0;
 SET idle_in_transaction_session_timeout = 0;
@@ -46,7 +47,6 @@ BEGIN
   IF del_count > 0 THEN
     UPDATE msg_queues
     SET msg_can_write = msg_can_write OR msg_queue_size <= del_count,
-        msg_queue_expire = msg_queue_size > del_count AND keep_min_id IS NOT NULL,
         msg_queue_size = GREATEST(msg_queue_size - del_count, 0)
     WHERE recipient_id = p_recipient_id;
   END IF;
@@ -56,31 +56,39 @@ $$;
 
 
 
-CREATE PROCEDURE smp_server.expire_old_messages(IN p_old_queue bigint, IN p_old_ts bigint, IN batch_size integer, OUT r_expired_msgs_count bigint, OUT r_stored_msgs_count bigint, OUT r_stored_queues bigint)
+CREATE PROCEDURE smp_server.expire_old_messages(IN p_old_ts bigint, IN batch_size integer, OUT r_expired_msgs_count bigint, OUT r_stored_msgs_count bigint, OUT r_stored_queues bigint)
     LANGUAGE plpgsql
     AS $$
 DECLARE
   rids BYTEA[];
   rid BYTEA;
+  last_ts BIGINT := -1;
   last_rid BYTEA := '\x';
+  next_ts BIGINT;
+  next_rid BYTEA;
   del_count BIGINT;
   total_deleted BIGINT := 0;
 BEGIN
   LOOP
-    SELECT array_agg(recipient_id)
-    INTO rids
+    -- The page is scanned in (msg_ts, recipient_id) order, so its last row is the next
+    -- keyset cursor. Advancing it past every row read, including queues left unexpired
+    -- because delete_expired_msgs skipped a locked row or raised, is what terminates the
+    -- loop; re-reading from the start instead would repeat those queues forever.
+    SELECT array_agg(DISTINCT recipient_id),
+           (array_agg(msg_ts ORDER BY msg_ts DESC, recipient_id DESC))[1],
+           (array_agg(recipient_id ORDER BY msg_ts DESC, recipient_id DESC))[1]
+    INTO rids, next_ts, next_rid
     FROM (
-      SELECT recipient_id
-      FROM msg_queues
-      WHERE deleted_at IS NULL
-        AND updated_at > p_old_queue
-        AND msg_queue_expire = TRUE
-        AND recipient_id > last_rid
-      ORDER BY recipient_id ASC
+      SELECT msg_ts, recipient_id
+      FROM messages
+      WHERE NOT msg_quota
+        AND msg_ts < p_old_ts
+        AND (msg_ts, recipient_id) > (last_ts, last_rid)
+      ORDER BY msg_ts ASC, recipient_id ASC
       LIMIT batch_size
-    ) qs;
+    ) m;
 
-    EXIT WHEN rids IS NULL OR cardinality(rids) = 0;
+    EXIT WHEN rids IS NULL;
 
     FOREACH rid IN ARRAY rids
     LOOP
@@ -91,9 +99,11 @@ BEGIN
         RAISE WARNING 'STORE, expire_old_messages, error expiring queue %: %', encode(rid, 'base64'), SQLERRM;
         CONTINUE;
       END;
-      COMMIT;
     END LOOP;
-    last_rid := rids[cardinality(rids)];
+    COMMIT;
+
+    last_ts := next_ts;
+    last_rid := next_rid;
   END LOOP;
 
   r_expired_msgs_count := total_deleted;
@@ -195,7 +205,6 @@ BEGIN
     IF q_size != 0 THEN
       UPDATE msg_queues
       SET msg_can_write = TRUE,
-          msg_queue_expire = FALSE,
           msg_queue_size = 0
       WHERE recipient_id = p_recipient_id;
     END IF;
@@ -207,7 +216,6 @@ BEGIN
     IF FOUND THEN
       UPDATE msg_queues
       SET msg_can_write = msg_can_write OR msg_queue_size <= 1,
-          msg_queue_expire = msg_queue_size > 1,
           msg_queue_size = GREATEST(msg_queue_size - 1, 0)
       WHERE recipient_id = p_recipient_id;
       RETURN QUERY VALUES (msg.msg_id, msg.msg_ts, msg.msg_quota, msg.msg_ntf_flag, msg.msg_body);
@@ -245,7 +253,6 @@ BEGIN
     IF q_size != 0 THEN
       UPDATE msg_queues
       SET msg_can_write = TRUE,
-          msg_queue_expire = FALSE,
           msg_queue_size = 0
       WHERE recipient_id = p_recipient_id;
     END IF;
@@ -271,14 +278,12 @@ BEGIN
       IF msg_deleted THEN
         UPDATE msg_queues
         SET msg_can_write = msg_can_write OR msg_queue_size <= 1,
-            msg_queue_expire = msg_queue_size > 1,
             msg_queue_size = GREATEST(msg_queue_size - 1, 0)
         WHERE recipient_id = p_recipient_id;
       END IF;
     ELSIF msg_deleted OR q_size != 0 THEN
       UPDATE msg_queues
       SET msg_can_write = TRUE,
-          msg_queue_expire = FALSE,
           msg_queue_size = 0
       WHERE recipient_id = p_recipient_id;
     END IF;
@@ -348,7 +353,6 @@ BEGIN
 
     UPDATE msg_queues
     SET msg_can_write = NOT quota_written,
-        msg_queue_expire = TRUE,
         msg_queue_size = msg_queue_size + 1
     WHERE recipient_id = p_recipient_id;
 
@@ -397,7 +401,8 @@ CREATE TABLE smp_server.messages (
     msg_quota boolean NOT NULL,
     msg_ntf_flag boolean NOT NULL,
     msg_body bytea NOT NULL
-);
+)
+WITH (autovacuum_vacuum_scale_factor='0.02', autovacuum_analyze_scale_factor='0.01', toast.autovacuum_vacuum_scale_factor='0.02');
 
 
 
@@ -439,9 +444,9 @@ CREATE TABLE smp_server.msg_queues (
     rcv_service_id bytea,
     ntf_service_id bytea,
     msg_can_write boolean DEFAULT true NOT NULL,
-    msg_queue_expire boolean DEFAULT false NOT NULL,
     msg_queue_size bigint DEFAULT 0 NOT NULL
-);
+)
+WITH (fillfactor='80', autovacuum_vacuum_scale_factor='0.02', autovacuum_analyze_scale_factor='0.01', autovacuum_vacuum_cost_limit='1000');
 
 
 
@@ -453,7 +458,8 @@ CREATE TABLE smp_server.services (
     created_at bigint NOT NULL,
     queue_count bigint DEFAULT 0 NOT NULL,
     queue_ids_hash bytea DEFAULT '\x00000000000000000000000000000000'::bytea NOT NULL
-);
+)
+WITH (fillfactor='70', autovacuum_vacuum_threshold='1000', autovacuum_vacuum_scale_factor='0');
 
 
 
@@ -482,6 +488,10 @@ ALTER TABLE ONLY smp_server.services
 
 
 
+CREATE INDEX idx_messages_expire ON smp_server.messages USING btree (msg_ts, recipient_id) WHERE (NOT msg_quota);
+
+
+
 CREATE INDEX idx_messages_recipient_id_message_id ON smp_server.messages USING btree (recipient_id, message_id);
 
 
@@ -490,14 +500,14 @@ CREATE INDEX idx_messages_recipient_id_msg_quota ON smp_server.messages USING bt
 
 
 
-CREATE INDEX idx_messages_recipient_id_msg_ts ON smp_server.messages USING btree (recipient_id, msg_ts);
-
-
-
 CREATE UNIQUE INDEX idx_msg_queues_link_id ON smp_server.msg_queues USING btree (link_id);
 
 
 
+CREATE INDEX idx_msg_queues_notifier_active ON smp_server.msg_queues USING btree (notifier_id) WHERE ((deleted_at IS NULL) AND (notifier_id IS NOT NULL));
+
+
+
 CREATE UNIQUE INDEX idx_msg_queues_notifier_id ON smp_server.msg_queues USING btree (notifier_id);
 
 
@@ -514,10 +524,6 @@ CREATE UNIQUE INDEX idx_msg_queues_sender_id ON smp_server.msg_queues USING btre
 
 
 
-CREATE INDEX idx_msg_queues_updated_at_recipient_id ON smp_server.msg_queues USING btree (deleted_at, updated_at, msg_queue_expire, recipient_id);
-
-
-
 CREATE INDEX idx_services_service_role ON smp_server.services USING btree (service_role);
 
 
@@ -549,3 +555,4 @@ ALTER TABLE ONLY smp_server.msg_queues
 
 
 
+

src/Simplex/Messaging/Server/QueueStore/STM.hs

@@ -173,6 +173,7 @@ instance StoreQueueClass q => QueueStoreClass q (STMQueueStore q) where
       rId = recipientId sq
       qr = queueRec sq
       add q = case queueData q of
+        _ | isSecuredMsgQueue q -> pure $ Left AUTH
         Nothing -> addLink
         Just (lnkId', d') | lnkId' == lnkId && fst d' == fst d -> addLink
         _ -> pure $ Left AUTH

src/Simplex/Messaging/Server/Stats.hs

@@ -128,6 +128,7 @@ data ServerStats = ServerStats
     ntfServices :: ServiceStats,
     rcvServicesSubMsg :: IORef Int,
     rcvServicesSubDuplicate :: IORef Int,
+    authDelayExceeded :: IORef Int,
     qCount :: IORef Int,
     msgCount :: IORef Int,
     ntfCount :: IORef Int,
@@ -190,6 +191,7 @@ data ServerStatsData = ServerStatsData
     _rcvServices :: ServiceStatsData,
     _rcvServicesSubMsg :: Int,
     _rcvServicesSubDuplicate :: Int,
+    _authDelayExceeded :: Int,
     _qCount :: Int,
     _msgCount :: Int,
     _ntfCount :: Int,
@@ -254,6 +256,7 @@ newServerStats ts = do
   ntfServices <- newServiceStats
   rcvServicesSubMsg <- newIORef 0
   rcvServicesSubDuplicate <- newIORef 0
+  authDelayExceeded <- newIORef 0
   qCount <- newIORef 0
   msgCount <- newIORef 0
   ntfCount <- newIORef 0
@@ -315,6 +318,7 @@ newServerStats ts = do
         ntfServices,
         rcvServicesSubMsg,
         rcvServicesSubDuplicate,
+        authDelayExceeded,
         qCount,
         msgCount,
         ntfCount,
@@ -378,6 +382,7 @@ getServerStatsData s = do
   _ntfServices <- getServiceStatsData $ ntfServices s
   _rcvServicesSubMsg <- readIORef $ rcvServicesSubMsg s
   _rcvServicesSubDuplicate <- readIORef $ rcvServicesSubDuplicate s
+  _authDelayExceeded <- readIORef $ authDelayExceeded s
   _qCount <- readIORef $ qCount s
   _msgCount <- readIORef $ msgCount s
   _ntfCount <- readIORef $ ntfCount s
@@ -439,6 +444,7 @@ getServerStatsData s = do
         _ntfServices,
         _rcvServicesSubMsg,
         _rcvServicesSubDuplicate,
+        _authDelayExceeded,
         _qCount,
         _msgCount,
         _ntfCount,
@@ -503,6 +509,7 @@ setServerStats s d = do
   setServiceStats (ntfServices s) $! _ntfServices d
   writeIORef (rcvServicesSubMsg s) $! _rcvServicesSubMsg d
   writeIORef (rcvServicesSubDuplicate s) $! _rcvServicesSubDuplicate d
+  writeIORef (authDelayExceeded s) $! _authDelayExceeded d
   writeIORef (qCount s) $! _qCount d
   writeIORef (msgCount s) $! _msgCount d
   writeIORef (ntfCount s) $! _ntfCount d
@@ -705,6 +712,7 @@ instance StrEncoding ServerStatsData where
           _ntfServices,
           _rcvServicesSubMsg = 0,
           _rcvServicesSubDuplicate = 0,
+          _authDelayExceeded = 0,
           _qCount,
           _msgCount = 0,
           _ntfCount = 0,

src/Simplex/Messaging/SimplexName.hs

@@ -10,14 +10,21 @@ module Simplex.Messaging.SimplexName
     SimplexTLD (..),
     SimplexNameType (..),
     fullDomainName,
+    LabelHash (..),
+    labelHash,
+    boundedNonSpace,
     shortNameInfoStr,
   )
 where
 
 import Control.Applicative (optional, (<|>))
+import Crypto.Hash (Digest, hash)
+import Crypto.Hash.Algorithms (Keccak_256)
 import qualified Data.Aeson.TH as J
 import qualified Data.Attoparsec.ByteString.Char8 as A
 import qualified Data.Attoparsec.Text as AT
+import qualified Data.ByteArray as BA
+import qualified Data.ByteArray.Encoding as BAE
 import Data.ByteString.Char8 (ByteString)
 import qualified Data.ByteString.Char8 as B
 import Data.Char (isDigit)
@@ -70,6 +77,20 @@ nameLabelP = do
     -- (Cyrillic а vs ASCII a hash to different on-chain records).
     isNameLetter c = c >= 'a' && c <= 'z' || c >= 'A' && c <= 'Z'
 
+-- | The registry's key for a label: 32 bytes, as labelOf takes it.
+newtype LabelHash = LabelHash ByteString
+  deriving (Eq, Show)
+
+-- | keccak-256 of the lowercased label, as the registry keys it.
+labelHash :: Text -> LabelHash
+labelHash label = LabelHash $ BA.convert (hash (encodeUtf8 (T.toLower label)) :: Digest Keccak_256)
+
+instance StrEncoding LabelHash where
+  strEncode (LabelHash h) = '[' `B.cons` (BAE.convertToBase BAE.Base16 h `B.snoc` ']')
+  strP = do
+    h <- BAE.convertFromBase BAE.Base16 <$?> (A.char '[' *> A.takeWhile (/= ']') <* A.char ']')
+    if B.length h == 32 then pure $ LabelHash h else fail "bad LabelHash"
+
 -- | Cap the name at 253 bytes (DNS full-domain limit)
 boundedNonSpace :: A.Parser ByteString
 boundedNonSpace = do
@@ -108,12 +129,15 @@ instance Encoding SimplexDomain where
   smpP = strP
 
 fullDomainName :: SimplexDomain -> Text
-fullDomainName SimplexDomain {nameTLD, domain, subDomain} = T.intercalate "." (reverse subDomain ++ [domain] ++ tld')
-  where
-    tld' = case nameTLD of
-      TLDSimplex -> ["simplex"]
-      TLDTesting -> ["testing"]
-      TLDWeb -> []
+fullDomainName SimplexDomain {nameTLD, domain, subDomain} = T.intercalate "." (reverse subDomain ++ [domain]) <> decodeLatin1 (strEncode nameTLD)
+
+instance StrEncoding SimplexTLD where
+  strEncode = \case
+    TLDSimplex -> ".simplex"
+    TLDTesting -> ".testing"
+    TLDWeb -> ""
+  strP =
+    ".simplex" $> TLDSimplex <|> ".testing" $> TLDTesting <|> pure TLDWeb
 
 shortNameInfoStr :: SimplexNameInfo -> Text
 shortNameInfoStr = \case

src/Simplex/Messaging/Transport.hs

@@ -46,13 +46,14 @@ module Simplex.Messaging.Transport
     minServerSMPRelayVersion,
     currentClientSMPRelayVersion,
     currentServerSMPRelayVersion,
-    shortLinksSMPVersion,
     serviceCertsSMPVersion,
     newNtfCredsSMPVersion,
     clientNoticesSMPVersion,
     rcvServiceSMPVersion,
     namesSMPVersion,
     serverInfoSMPVersion,
+    nameAvailSMPVersion,
+    fwdNoncesSMPVersion,
     simplexMQVersion,
     smpBlockSize,
     TransportConfig (..),
@@ -175,6 +176,8 @@ smpBlockSize = 16384
 -- 19 - service subscriptions to messages (10/20/2025)
 -- 20 - public namespaces resolver, RSLV command (6/20/2026)
 -- 21 - server public information in handshake (7/5/2026)
+-- 22 - RNAME answers name availability as well as the record (7/25/2026)
+-- 23 - version in forwarded command nonce, random nonce in forwarded responses (10/2/2026)
 
 data SMPVersion
 
@@ -187,11 +190,8 @@ type VersionRangeSMP = VersionRange SMPVersion
 pattern VersionSMP :: Word16 -> VersionSMP
 pattern VersionSMP v = Version v
 
-_proxyServerHandshakeSMPVersion :: VersionSMP
-_proxyServerHandshakeSMPVersion = VersionSMP 14
-
-shortLinksSMPVersion :: VersionSMP
-shortLinksSMPVersion = VersionSMP 15
+_shortLinksSMPVersion :: VersionSMP
+_shortLinksSMPVersion = VersionSMP 15
 
 serviceCertsSMPVersion :: VersionSMP
 serviceCertsSMPVersion = VersionSMP 16
@@ -211,29 +211,37 @@ namesSMPVersion = VersionSMP 20
 serverInfoSMPVersion :: VersionSMP
 serverInfoSMPVersion = VersionSMP 21
 
+-- | RNAME carries availability. A server below this answers RSLV with the
+-- record alone, and ERR NAME NOT_FOUND for a name that does not resolve.
+nameAvailSMPVersion :: VersionSMP
+nameAvailSMPVersion = VersionSMP 22
+
+fwdNoncesSMPVersion :: VersionSMP
+fwdNoncesSMPVersion = VersionSMP 23
+
 minClientSMPRelayVersion :: VersionSMP
-minClientSMPRelayVersion = VersionSMP 14
+minClientSMPRelayVersion = VersionSMP 15
 
 minServerSMPRelayVersion :: VersionSMP
-minServerSMPRelayVersion = VersionSMP 14
+minServerSMPRelayVersion = VersionSMP 15
 
 currentClientSMPRelayVersion :: VersionSMP
-currentClientSMPRelayVersion = VersionSMP 21
+currentClientSMPRelayVersion = VersionSMP 23
 
 currentServerSMPRelayVersion :: VersionSMP
-currentServerSMPRelayVersion = VersionSMP 21
+currentServerSMPRelayVersion = VersionSMP 23
 
 -- Max SMP protocol version to be used in e2e encrypted connection between
 -- client and server, as defined by SMP proxy. Normally set below the current
 -- version to prevent client version fingerprinting by the destination relays
--- when clients upgrade at different times. Pinned to the current version (20)
--- for this release because proxied name resolution is gated on namesSMPVersion
--- (20), so the one-version anti-fingerprinting buffer does not apply yet; it
--- reappears once the current version advances past 20.
+-- when clients upgrade at different times. Pinned to the current version (23)
+-- for this release because forwarded commands use the nonces from
+-- fwdNoncesSMPVersion (23), so the one-version anti-fingerprinting buffer does
+-- not apply yet; it reappears once the current version advances past 23.
 proxiedSMPRelayVersion :: VersionSMP
-proxiedSMPRelayVersion = VersionSMP 20
+proxiedSMPRelayVersion = VersionSMP 23
 
--- minimal supported protocol version is 14
+-- minimal supported protocol version is 15
 supportedClientSMPRelayVRange :: VersionRangeSMP
 supportedClientSMPRelayVRange = mkVersionRange minClientSMPRelayVersion currentClientSMPRelayVersion
 
@@ -858,7 +866,6 @@ forceCertChain :: CertChainPubKey -> CertChainPubKey
 forceCertChain cert@(CertChainPubKey (X.CertificateChain cc) signedKey) = length (show cc) `seq` show signedKey `seq` cert
 {-# INLINE forceCertChain #-}
 
--- This function is only used with v >= 8, so currently it's a simple record update.
 -- * Note: it requires updating version-based parameters, to be consistent with smpTHandle_.
 smpTHParamsSetVersion :: VersionSMP -> THandleParams SMPVersion p -> THandleParams SMPVersion p
 smpTHParamsSetVersion v params =

src/Simplex/RemoteControl/Client.hs

@@ -26,6 +26,7 @@ module Simplex.RemoteControl.Client
     -- for tests only
     sendRCPacket,
     receiveRCPacket,
+    findRCCtrlPairing,
   ) where
 
 import Control.Applicative ((<|>))
@@ -46,7 +47,7 @@ import Data.List.NonEmpty (NonEmpty (..))
 import qualified Data.List.NonEmpty as L
 import Data.Maybe (isNothing)
 import qualified Data.Text as T
-import Data.Time.Clock.System (getSystemTime)
+import Data.Time.Clock.System (SystemTime (..), getSystemTime)
 import Data.Tuple (swap)
 import Data.Word (Word16)
 import qualified Data.X509 as X
@@ -395,8 +396,10 @@ findRCCtrlPairing :: NonEmpty RCCtrlPairing -> RCEncInvitation -> ExceptT RCErro
 findRCCtrlPairing pairings RCEncInvitation {dhPubKey, nonce, encInvitation} = do
   (pairing, signedInvStr) <- liftEither $ decrypt (L.toList pairings)
   signedInv <- liftEitherWith RCESyntax $ strDecode signedInvStr
-  inv@(RCVerifiedInvitation RCInvitation {dh = invDh}) <- maybe (throwE RCEInvitation) pure $ verifySignedInvitation signedInv
+  inv@(RCVerifiedInvitation RCInvitation {dh = invDh, ts}) <- maybe (throwE RCEInvitation) pure $ verifySignedInvitation signedInv
   unless (invDh == dhPubKey) $ throwE RCEInvitation
+  now <- systemSeconds <$> liftIO getSystemTime
+  unless (now - 3660 <= systemSeconds ts && systemSeconds ts <= now + 3600) $ throwE RCEInvitation
   pure (pairing, inv)
   where
     decrypt :: [RCCtrlPairing] -> Either RCErrorType (RCCtrlPairing, ByteString)

src/Simplex/RemoteControl/Discovery.hs

@@ -122,18 +122,22 @@ closeListener subscribers sock =
 joinMulticast :: TMVar Int -> N.Socket -> N.HostAddress -> IO ()
 joinMulticast subscribers sock group = do
   now <- atomically $ takeTMVar subscribers
-  when (now == 0) $ do
-    setMembership sock group True >>= \case
-      Left e -> atomically (putTMVar subscribers now) >> logError ("setMembership failed " <> tshow e)
-      Right () -> atomically $ putTMVar subscribers (now + 1)
+  if now == 0
+    then
+      setMembership sock group True >>= \case
+        Left e -> atomically (putTMVar subscribers now) >> logError ("setMembership failed " <> tshow e)
+        Right () -> atomically $ putTMVar subscribers (now + 1)
+    else atomically $ putTMVar subscribers (now + 1)
 
 partMulticast :: TMVar Int -> N.Socket -> N.HostAddress -> IO ()
 partMulticast subscribers sock group = do
   now <- atomically $ takeTMVar subscribers
-  when (now == 1) $
-    setMembership sock group False >>= \case
-      Left e -> atomically (putTMVar subscribers now) >> logError ("setMembership failed " <> tshow e)
-      Right () -> atomically $ putTMVar subscribers (now - 1)
+  if now == 1
+    then
+      setMembership sock group False >>= \case
+        Left e -> atomically (putTMVar subscribers (now - 1)) >> logError ("setMembership failed " <> tshow e)
+        Right () -> atomically $ putTMVar subscribers (now - 1)
+    else atomically $ putTMVar subscribers (max 0 (now - 1))
 
 listenerHostAddr4 :: UDP.ListenSocket -> N.HostAddress
 listenerHostAddr4 sock = case UDP.mySockAddr sock of

tests/AgentTests/ConnectionRequestTests.hs

@@ -20,6 +20,8 @@ module AgentTests.ConnectionRequestTests
 
 import AgentTests.EqInstances ()
 import Data.ByteString (ByteString)
+import qualified Data.ByteString.Char8 as B
+import Data.Either (isLeft)
 import Network.HTTP.Types (urlEncode)
 import Simplex.Messaging.Agent.Protocol
 import qualified Simplex.Messaging.Crypto as C
@@ -285,6 +287,9 @@ connectionRequestTests =
       contactAddressV6 #== ("https://simplex.chat/contact#/?v=1-2&smp=" <> url queueStr) -- adjusted to v6
       contactAddressV6 #== ("https://simplex.chat/contact#/?v=2-2&smp=" <> url queueStr)
       contactAddressClientData #==# ("simplex:/contact#/?v=6-8&smp=" <> url queueStr <> "&data=" <> url "{\"type\":\"group_link\", \"group_link_id\":\"abc\"}")
+    it "should reject KEM ciphertext without KEM key in e2e params" $
+      strDecode @(RcvE2ERatchetParamsUri 'C.X448) (strEncode testE2ERatchetParams <> "&kem_ct=" <> strEncode (B.replicate 1039 '\0'))
+        `shouldSatisfy` isLeft
     it "should serialize / parse queue address, connection invitations and contact addresses as binary" $ do
       smpEncodingTest queue
       smpEncodingTest queueNoQM -- this passes, no queue mode patch in SMPQueueUri encoding
@@ -357,6 +362,10 @@ connectionRequestTests =
       smpEncodingTest $ AgentServiceRequest [qInfo] Nothing "service request payload"
       smpEncodingTest $ AgentServiceResponse "service response payload"
       smpEncodingTest $ AgentRejection "rejected: not allowed"
+    it "should serialize and parse ratchet key info" $ do
+      smpDecode "R" `shouldBe` Right (AgentRatchetInfo RatchetInfo {answeredKeyHash = Nothing})
+      smpDecode "R1\3abcdef" `shouldBe` Right (AgentRatchetInfo RatchetInfo {answeredKeyHash = Just "abc"})
+      smpEncodingTest $ AgentRatchetInfo RatchetInfo {answeredKeyHash = Just "0123456789abcdef0123456789abcdef"}
   where
     smpEncodingTest :: (Encoding a, Eq a, Show a, HasCallStack) => a -> Expectation
     smpEncodingTest a = smpDecode (smpEncode a) `shouldBe` Right a

tests/AgentTests/DoubleRatchetTests.hs

@@ -53,6 +53,7 @@ doubleRatchetTests = do
       it "should propose KEM during agreement, but no shared secret" $ testAlgs testPqX3dhProposeInReply
       it "should agree shared secret using KEM" $ testAlgs testPqX3dhProposeAccept
       it "should reject proposed KEM in reply" $ testAlgs testPqX3dhProposeReject
+      it "should agree different PQ associated data with substituted KEM key" $ testAlgs testPqX3dhSubstitutedKem
       it "should allow second proposal in reply" $ testAlgs testPqX3dhProposeAgain
     describe "hybrid KEM key agreement errors" $ do
       it "should fail if reply contains acceptance without proposal" $ testAlgs testPqX3dhAcceptWithoutProposalError
@@ -362,6 +363,7 @@ testDecodeV2RatchetJSON :: IO ()
 testDecodeV2RatchetJSON = do
   let v2RatchetJSON = "{\"rcVersion\":[2,2],\"rcAD\":\"2GEJrq48TmQse6NR16I-hrI0tSySZQ57E_g46nDceAPRAiF6j0drq26RTE7be6X7uiB4RaGJGf4QRXzcYuVtWw==\",\"rcDHRs\":\"TUM0Q0FRQXdCUVlESzJWdUJDSUVJRkNYbUxtSHQ3SUNfeHpGTi1Qb3ZqTVQ3S2p6XzZlZlBjOG9fRFY2RWxKOQ==\",\"rcRK\":\"BOX2X7YW5qDSp2XknY_lqacSrtDqQNPvS6iJlZIs3G0=\",\"rcNs\":0,\"rcNr\":0,\"rcPN\":0,\"rcNHKs\":\"IMouSkXUvzT_mo0WM-pqEUK09-HTLk9WOTCFQglyQxU=\",\"rcNHKr\":\"g-tus1clYPV0rGlzkf5a959tUqDYQVZ1FpcPeXdKwxI=\"}"
   Right (r :: Ratchet X25519) <- pure $ J.eitherDecodeStrict' v2RatchetJSON
+  rcVCPQ r `shouldBe` Nothing
   rcSupportKEM r `shouldBe` PQSupportOff
   rcEnableKEM r `shouldBe` PQEncOff
   rcSndKEM r `shouldBe` PQEncOff
@@ -417,6 +419,18 @@ testPqX3dhProposeAccept _ = do
   Right paramsAlice <- runExceptT $ pqX3dhRcv pksAlice e2eBob
   paramsAlice `compatibleRatchets` paramsBob
 
+testPqX3dhSubstitutedKem :: forall a. (AlgorithmI a, DhAlgorithm a) => C.SAlgorithm a -> IO ()
+testPqX3dhSubstitutedKem _ = do
+  g <- C.newRandom
+  let v = currentE2EEncryptVersion
+  (pksAlice@(_, _, Just _), e2eAlice) <- liftIO $ generateRcvE2EParams @a g v PQSupportOn
+  (_, E2ERatchetParams _ _ _ (Just (RKParamsProposed mallorysKem))) <- liftIO $ generateRcvE2EParams @a g v PQSupportOn
+  (pksBob@(_, _, Just _), AE2ERatchetParams _ e2eBob) <- liftIO $ generateSndE2EParams @a g v (Just $ AUseKEM SRKSAccepted $ AcceptKEM mallorysKem)
+  Right (paramsBob, _) <- pure $ pqX3dhSnd pksBob e2eAlice
+  Right (paramsAlice, _) <- runExceptT $ pqX3dhRcv pksAlice e2eBob
+  assocData paramsAlice `shouldBe` assocData paramsBob
+  rcVerifyCodePQ paramsAlice `shouldNotBe` rcVerifyCodePQ paramsBob
+
 testPqX3dhProposeReject :: forall a. (AlgorithmI a, DhAlgorithm a) => C.SAlgorithm a -> IO ()
 testPqX3dhProposeReject _ = do
   g <- C.newRandom
@@ -459,9 +473,9 @@ testPqX3dhProposeAgain _ = do
 
 compatibleRatchets :: (RatchetInitParams, x) -> (RatchetInitParams, x) -> Expectation
 compatibleRatchets
-  (RatchetInitParams {assocData, ratchetKey, sndHK, rcvNextHK, kemAccepted}, _)
-  (RatchetInitParams {assocData = ad, ratchetKey = rk, sndHK = shk, rcvNextHK = rnhk, kemAccepted = ka}, _) = do
-    assocData == ad && ratchetKey == rk && sndHK == shk && rcvNextHK == rnhk `shouldBe` True
+  (RatchetInitParams {assocData, rcVerifyCodePQ, ratchetKey, sndHK, rcvNextHK, kemAccepted}, _)
+  (RatchetInitParams {assocData = ad, rcVerifyCodePQ = vcPQ, ratchetKey = rk, sndHK = shk, rcvNextHK = rnhk, kemAccepted = ka}, _) = do
+    assocData == ad && rcVerifyCodePQ == vcPQ && ratchetKey == rk && sndHK == shk && rcvNextHK == rnhk `shouldBe` True
     case (kemAccepted, ka) of
       (Just RatchetKEMAccepted {rcPQRr, rcPQRss, rcPQRct}, Just RatchetKEMAccepted {rcPQRr = pqk, rcPQRss = pqss, rcPQRct = pqct}) ->
         pqk /= rcPQRr && pqss == rcPQRss && pqct == rcPQRct `shouldBe` True

tests/AgentTests/FunctionalAPITests.hs

@@ -70,8 +70,8 @@ import Data.ByteString.Char8 (ByteString)
 import qualified Data.ByteString.Char8 as B
 import Data.Either (isRight)
 import Data.Int (Int64)
-import Data.List (find, isPrefixOf, isSuffixOf)
-import Data.List.NonEmpty (NonEmpty)
+import Data.List (find, isInfixOf, isPrefixOf, isSuffixOf)
+import Data.List.NonEmpty (NonEmpty (..))
 import qualified Data.Map as M
 import Data.Maybe (isJust, isNothing)
 import qualified Data.Set as S
@@ -87,12 +87,12 @@ import SMPAgentClient
 import SMPClient
 import Simplex.Messaging.Agent hiding (acceptContact, createConnection, deleteConnection, deleteConnections, getConnShortLink, joinConnection, sendMessage, setConnShortLink, subscribeConnection, suspendConnection)
 import qualified Simplex.Messaging.Agent as A
-import Simplex.Messaging.Agent.Client (ProtocolTestFailure (..), ProtocolTestStep (..), ServerQueueInfo (..), UserNetworkInfo (..), UserNetworkType (..), waitForUserNetwork)
+import Simplex.Messaging.Agent.Client (ProtocolTestFailure (..), ProtocolTestStep (..), ServerQueueInfo (..), UserNetworkInfo (..), UserNetworkType (..), sendAgentMessage, waitForUserNetwork)
 import Simplex.Messaging.Agent.Env.SQLite (AgentConfig (..), Env (..), InitialAgentServers (..), createAgentStore)
 import Simplex.Messaging.Agent.Protocol hiding (CON, CONF, INFO, REQ, SENT)
 import qualified Simplex.Messaging.Agent.Protocol as A
 import Simplex.Messaging.Agent.Store (Connection' (..), SomeConn' (..), StoredRcvQueue (..))
-import Simplex.Messaging.Agent.Store.AgentStore (getConn)
+import Simplex.Messaging.Agent.Store.AgentStore (deleteRatchetKeyHashesExpired, getConn, getConnIds, getRatchetX3dhKeys)
 import Simplex.Messaging.Agent.Store.Common (DBStore (..), withTransaction)
 import Simplex.Messaging.Agent.Store.Interface
 import qualified Simplex.Messaging.Agent.Store.DB as DB
@@ -112,7 +112,7 @@ import Simplex.Messaging.Server.Information (ServerPublicInfo (..))
 import Simplex.Messaging.Server.MsgStore.Types (SMSType (..), SQSType (..))
 import Simplex.Messaging.Server.QueueStore.QueueInfo
 import Simplex.Messaging.Server.StoreLog (StoreLogRecord (..))
-import Simplex.Messaging.Transport (ASrvTransport, SMPVersion, VersionSMP, currentServerSMPRelayVersion, minClientSMPRelayVersion, minServerSMPRelayVersion, alpnSupportedSMPHandshakes, supportedServerSMPRelayVRange)
+import Simplex.Messaging.Transport (ASrvTransport, SMPVersion, VersionSMP, currentServerSMPRelayVersion, minClientSMPRelayVersion, minServerSMPRelayVersion, alpnSupportedSMPHandshakes)
 import Simplex.Messaging.Transport.Server (TransportServerConfig (..))
 import Simplex.Messaging.Util (bshow, diffToMicroseconds)
 import Simplex.Messaging.Version (VersionRange (..))
@@ -204,7 +204,7 @@ pattern INFO :: ConnInfo -> AEvent 'AEConn
 pattern INFO connInfo = A.INFO PQSupportOn connInfo
 
 pattern REQ :: InvitationId -> NonEmpty SMPServer -> ConnInfo -> AEvent e
-pattern REQ invId srvs connInfo <- A.REQ invId PQSupportOn srvs connInfo _
+pattern REQ invId srvs connInfo <- A.REQ invId PQSupportOn srvs connInfo _ _
 
 pattern CON :: AEvent 'AEConn
 pattern CON = A.CON PQEncOn
@@ -231,7 +231,11 @@ pattern Rcvd' :: AgentMsgId -> AgentMsgId -> AEvent 'AEConn
 pattern Rcvd' aMsgId rcvdMsgId <- RCVD MsgMeta {integrity = MsgOk, recipient = (aMsgId, _)} [MsgReceipt {agentMsgId = rcvdMsgId, msgRcptStatus = MROk}]
 
 smpCfgVPrev :: ProtocolClientConfig SMPVersion
-smpCfgVPrev = (smpCfg agentCfg) {serverVRange = prevRange $ serverVRange $ smpCfg agentCfg}
+smpCfgVPrev =
+  (smpCfg agentCfg)
+    { serverVRange = prevRange $ serverVRange $ smpCfg agentCfg,
+      proxiedRelayVRange = prevRange $ proxiedRelayVRange $ smpCfg agentCfg
+    }
 
 -- ntfCfgVPrev :: ProtocolClientConfig NTFVersion
 -- ntfCfgVPrev = (ntfCfg agentCfg) {clientALPN = Nothing, serverVRange = V.mkVersionRange (VersionNTF 1) (VersionNTF 1)}
@@ -293,7 +297,7 @@ createConnection c userId enableNtfs cMode clientData subMode = do
 
 joinConnection :: AgentClient -> UserId -> Bool -> ConnectionRequestUri c -> ConnInfo -> SubscriptionMode -> AE (ConnId, SndQueueSecured)
 joinConnection c userId enableNtfs cReq connInfo subMode = do
-  connId <- A.prepareConnectionToJoin c userId enableNtfs cReq PQSupportOn
+  (connId, _) <- A.prepareConnectionToJoin c userId enableNtfs cReq PQSupportOn
   sndSecure <- A.joinConnection c NRMInteractive userId connId enableNtfs cReq connInfo PQSupportOn subMode
   pure (connId, sndSecure)
 
@@ -367,6 +371,10 @@ functionalAPITests ps = do
       testAddDRViaSetConnShortLink ps
     it "should resume DR accept after a transient failure (reuse the send queue and ratchet)" $
       testAcceptContactDRResumeAfterOffline ps
+    it "should store the ratchet when DR accept is prepared" $
+      withSmpServer ps $ testAcceptContactDRPrepared False
+    it "should create the ratchet at DR accept when the prepared connection has none" $
+      withSmpServer ps $ testAcceptContactDRPrepared True
     it "should support rejecting contact request" $
       withSmpServer ps testRejectContactRequest
     it "should communicate rejection reason via double ratchet" $
@@ -411,13 +419,16 @@ functionalAPITests ps = do
     describe "should connect via 1-time short link with async join" $ testProxyMatrix ps testInvitationShortLinkAsync
     describe "should connect via contact short link" $ testProxyMatrix ps testContactShortLink
     describe "should add short link to existing contact and connect" $ testProxyMatrix ps testAddContactShortLink
-    xdescribe "try to create 1-time short link with prev versions" $ testProxyMatrixWithPrev ps testInvitationShortLinkPrev
     describe "server restart" $ do
       it "should get 1-time link data after restart" $ testInvitationShortLinkRestart ps
       it "should connect via contact short link after restart" $ testContactShortLinkRestart ps
       it "should connect via added contact short link after restart" $ testAddContactShortLinkRestart ps
     it "should create and get short links with the old contact queues" $ testOldContactQueueShortLink ps
     it "should connect via prepared connection link" $ testPrepareCreateConnectionLink ps
+    it "should connect via prepared invitation link" $ testPrepareCreateInvitationLink ps
+    it "should not create connection for prepared invitation link with too large data" $ testPrepareInvitationLinkTooLarge ps
+    it "should not create connection for prepared contact link with too large data" $ testPrepareContactLinkTooLarge ps
+    it "should connect via short link prepared for existing contact connection" $ testPrepareConnShortLink ps
   describe "Message delivery" $ do
     describe "update connection agent version on received messages" $ do
       it "should increase if compatible, shouldn'ps decrease" $
@@ -442,6 +453,9 @@ functionalAPITests ps = do
       it "should expire multiple messages" $ testExpireManyMessages ps
       it "should expire one message if quota is exceeded" $ testExpireMessageQuota ps
       it "should expire multiple messages if quota is exceeded" $ testExpireManyMessagesQuota ps
+    describe "joining a full contact address" $ do
+      it "should retry joining until the address has space" $ testJoinFullContactAsync ps
+      it "should fail joining after quota exceeded timeout" $ testJoinFullContactAsyncExpire ps
     it "should drop message after too many receive attempts" $ testDropMsgAfterRcvAttempts ps
 #if !defined(dbPostgres)
     -- TODO [postgres] restore from outdated db backup (we use copyFile/renameFile for sqlite)
@@ -456,6 +470,16 @@ functionalAPITests ps = do
         testRatchetSyncSuspendForeground ps
       it "should synchronize ratchets when clients start synchronization simultaneously" $
         testRatchetSyncSimultaneous ps
+      it "should ignore replayed ratchet key after expired hashes are deleted" $
+        testRatchetSyncReplayedKey ps
+      it "should synchronize ratchets when synchronization is forced again" $
+        testRatchetSyncRepeated ps
+      it "should not mark ratchet key as processed when ratchet recreation fails" $
+        testRatchetSyncFailedKeyNotProcessed ps
+      it "should not store reply ratchet key when ratchet recreation fails" $
+        testRatchetSyncFailedRecreationNoReply ps
+      it "should not store ratchet key when starting synchronization fails" $
+        testRatchetSyncStartFailedNoKey ps
 #endif
     describe "Subscription mode OnlyCreate" $ do
       it "messages delivered only when polled" $
@@ -534,6 +558,10 @@ functionalAPITests ps = do
       testServerMatrix2 ps testFastSwitchConnection
     it "should switch delivery to the new queue when the old server is down" $
       testFastSwitchDeadOldServer ps
+    it "should drain the old queue after QEND on the new queue" $
+      testFastSwitchDrainOldQueue ps
+    it "should drain the old queue after QEND on the new queue and restart" $
+      testFastSwitchDrainOldQueueRestart ps
     describe "should switch delivery to the new queue" $
       testServerMatrix2 ps testSwitchConnection
     describe "should switch to new queue asynchronously" $
@@ -581,9 +609,9 @@ functionalAPITests ps = do
       it "should pass with correct password" $ testSMPServerConnectionTest ps auth (srv auth) `shouldReturn` Right (Just (Right testServerInformation))
       it "should fail without password" $ testSMPServerConnectionTest ps auth (srv Nothing) `shouldReturn` Left authErr
       it "should fail with incorrect password" $ testSMPServerConnectionTest ps auth (srv $ Just "wrong") `shouldReturn` Left authErr
-  describe "getRatchetAdHash" $
-    it "should return the same data for both peers" $
-      withSmpServer ps testRatchetAdHash
+  describe "getConnectionVerifyCodes" $
+    it "should return the same codes for both peers" $
+      withSmpServer ps testConnectionVerifyCodes
   describe "Delivery receipts" $ do
     it "should send and receive delivery receipt" $ withSmpServer ps testDeliveryReceipts
     it "send delivery receipts concurrently with messages" $ testDeliveryReceiptsConcurrent ps
@@ -660,19 +688,6 @@ testProxyMatrix ps runTest = do
   it "2 servers, directly" $ withSmpServers2 ps $ withAgentClientsServers2 (agentCfg, initAgentServers) (agentCfg, initAgentServers2) $ runTest False
   it "2 servers, via proxy" $ withSmpServersProxy2 ps $ withAgentClientsServers2 (agentCfg, initAgentServersProxy) (agentCfg, initAgentServersProxy2) $ runTest True
 
-testProxyMatrixWithPrev :: HasCallStack => (ASrvTransport, AStoreType) -> (Bool -> Bool -> AgentClient -> AgentClient -> IO ()) -> Spec
-testProxyMatrixWithPrev ps@(t, msType@(ASType qs _ms)) runTest = do
-  it "2 servers, directly, curr clients, prev servers" $ withSmpServers2Prev $ withAgentClientsServers2 (agentCfg, initAgentServers) (agentCfg, initAgentServers2) $ runTest False True
-  it "2 servers, via proxy, curr clients, prev servers" $ withSmpServersProxy2Prev $ withAgentClientsServers2 (agentCfg, initAgentServersProxy) (agentCfg, initAgentServersProxy2) $ runTest True True
-  it "2 servers, directly, prev clients, curr servers" $ withSmpServers2 ps $ withAgentClientsServers2 (agentCfgVPrevPQ, initAgentServers) (agentCfgVPrevPQ, initAgentServers2) $ runTest False False
-  it "2 servers, via proxy, prev clients, curr servers" $ withSmpServersProxy2 ps $ withAgentClientsServers2 (agentCfgVPrevPQ, initAgentServersProxy) (agentCfgVPrevPQ, initAgentServersProxy2) $ runTest True False
-  where
-    prev cfg' = updateCfg cfg' $ \cfg_ -> cfg_ {smpServerVRange = prevRange supportedServerSMPRelayVRange}
-    withSmpServers2Prev a = withServers2 (prev $ cfgMS msType) (prev $ cfgJ2QS qs) a
-    withSmpServersProxy2Prev a = withServers2 (prev $ proxyCfgMS msType) (prev $ proxyCfgJ2QS qs) a
-    withServers2 cfg1 cfg2 a =
-      withSmpServerConfigOn t cfg1 testPort $ \_ -> withSmpServerConfigOn t cfg2 testPort2 $ \_ -> a
-
 testPQMatrix2 :: HasCallStack => (ASrvTransport, AStoreType) -> (HasCallStack => (AgentClient, InitialKeys) -> (AgentClient, PQSupport) -> AgentMsgId -> IO ()) -> Spec
 testPQMatrix2 = pqMatrix2_ True
 
@@ -755,7 +770,7 @@ runAgentClientTestPQ :: HasCallStack => Bool -> (AgentClient, InitialKeys) -> (A
 runAgentClientTestPQ viaProxy (alice, aPQ) (bob, bPQ) baseId =
   runRight_ $ do
     (bobId, CCLink qInfo Nothing) <- A.createConnection alice NRMInteractive 1 True True SCMInvitation Nothing Nothing aPQ False SMSubscribe
-    aliceId <- A.prepareConnectionToJoin bob 1 True qInfo bPQ
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True qInfo bPQ
     sqSecured' <- A.joinConnection bob NRMInteractive 1 aliceId True qInfo "bob's connInfo" bPQ SMSubscribe
     liftIO $ sqSecured' `shouldBe` True
     ("", _, A.CONF confId pqSup' _ "bob's connInfo") <- get alice
@@ -984,12 +999,12 @@ runAgentClientContactTestPQ :: HasCallStack => SndQueueSecured -> Bool -> (Agent
 runAgentClientContactTestPQ sqSecured viaProxy (alice, aPQ) (bob, bPQ) baseId =
   runRight_ $ do
     (_, CCLink qInfo Nothing) <- A.createConnection alice NRMInteractive 1 True True SCMContact Nothing Nothing aPQ False SMSubscribe
-    aliceId <- A.prepareConnectionToJoin bob 1 True qInfo bPQ
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True qInfo bPQ
     sqSecuredJoin <- A.joinConnection bob NRMInteractive 1 aliceId True qInfo "bob's connInfo" bPQ SMSubscribe
     liftIO $ sqSecuredJoin `shouldBe` False -- joining via contact address connection
-    ("", _, A.REQ invId pqSup' _ "bob's connInfo" _) <- get alice
+    ("", _, A.REQ invId pqSup' _ "bob's connInfo" _ _) <- get alice
     liftIO $ pqSup' `shouldBe` PQSupportOn
-    bobId <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption aPQ)
+    (bobId, _) <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption aPQ)
     sqSecured' <- acceptContact alice 1 bobId True invId "alice's connInfo" (CR.connPQEncryption aPQ) SMSubscribe
     liftIO $ sqSecured' `shouldBe` sqSecured
     ("", _, A.CONF confId pqSup'' _ "alice's connInfo") <- get bob
@@ -1053,7 +1068,7 @@ runAgentClientContactDRTest_ asyncAccept asyncJoin addrIK useDR bPQ ps = withSmp
       Nothing -> expectationFailure "address must advertise DR ratchet keys"
     -- classic (non-DR) join drops the address keys from the request
     let connReqJoin = if useDR then connReq' else case connReq' of CRContactUri d _ -> CRContactUri d Nothing
-    aliceId <- A.prepareConnectionToJoin bob 1 True connReqJoin bPQ
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True connReqJoin bPQ
     if asyncJoin
       then do
         A.joinConnectionAsync bob "join" False aliceId True connReqJoin "bob's connInfo" bPQ SMSubscribe
@@ -1061,9 +1076,9 @@ runAgentClientContactDRTest_ asyncAccept asyncJoin addrIK useDR bPQ ps = withSmp
       else do
         sqSecuredJoin <- A.joinConnection bob NRMInteractive 1 aliceId True connReqJoin "bob's connInfo" bPQ SMSubscribe
         liftIO $ sqSecuredJoin `shouldBe` False
-    ("", _, A.REQ invId reqPQ _ "bob's connInfo" _) <- get alice
+    ("", _, A.REQ invId reqPQ _ "bob's connInfo" _ _) <- get alice
     liftIO $ reqPQ `shouldBe` PQSupportOn
-    bobId <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
+    (bobId, _) <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
     if asyncAccept
       then do
         acceptContactAsync alice "accept" bobId True invId "alice's connInfo" (CR.connPQEncryption addrIK) SMSubscribe
@@ -1093,10 +1108,10 @@ runCreateConnectionDRTest_ asyncNew addrIK bPQ ps = withSmpServer ps $ withAgent
     liftIO $ case addrKeys_ of
       Just (_, CR.E2ERatchetParamsUri _ _ _ kem_) -> isJust kem_ `shouldBe` supportPQ (CR.initialPQEncryption False addrIK)
       Nothing -> expectationFailure "createConnection must advertise DR ratchet keys"
-    aliceId <- A.prepareConnectionToJoin bob 1 True connReq bPQ
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True connReq bPQ
     void $ A.joinConnection bob NRMInteractive 1 aliceId True connReq "bob's connInfo" bPQ SMSubscribe
-    ("", _, A.REQ invId _ _ "bob's connInfo" _) <- get alice
-    bobId <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
+    ("", _, A.REQ invId _ _ "bob's connInfo" _ _) <- get alice
+    (bobId, _) <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
     void $ acceptContact alice 1 bobId True invId "alice's connInfo" (CR.connPQEncryption addrIK) SMSubscribe
     ("", _, A.CONF confId _ _ "alice's connInfo") <- get bob
     allowConfirmGreet alice bobId bob aliceId confId addrIK pqEnc
@@ -1118,10 +1133,10 @@ testContactDRMatrix ps = do
 
 joinContactDR :: HasCallStack => AgentClient -> AgentClient -> ConnectionRequestUri 'CMContact -> InitialKeys -> PQEncryption -> ExceptT AgentErrorType IO ()
 joinContactDR alice requester connReq addrIK pqEnc = do
-  aliceId <- A.prepareConnectionToJoin requester 1 True connReq PQSupportOn
+  (aliceId, _) <- A.prepareConnectionToJoin requester 1 True connReq PQSupportOn
   void $ A.joinConnection requester NRMInteractive 1 aliceId True connReq "bob's connInfo" PQSupportOn SMSubscribe
-  ("", _, A.REQ invId _ _ "bob's connInfo" _) <- get alice
-  reqId <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
+  ("", _, A.REQ invId _ _ "bob's connInfo" _ _) <- get alice
+  (reqId, _) <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
   void $ acceptContact alice 1 reqId True invId "alice's connInfo" (CR.connPQEncryption addrIK) SMSubscribe
   ("", _, A.CONF confId _ _ "alice's connInfo") <- get requester
   allowConfirmGreet alice reqId requester aliceId confId addrIK pqEnc
@@ -1149,7 +1164,7 @@ testAddressKeyRotation ps = withSmpServer ps $ withAgentClients3 $ \alice bob ca
     joinContactDR alice carol connReq2 addrIK pqEnc
     void $ A.setConnShortLink alice NRMInteractive addrConnId SCMContact userLinkData Nothing True (Just addrIK)
     void $ A.setConnShortLink alice NRMInteractive addrConnId SCMContact userLinkData Nothing True (Just addrIK)
-    aId <- A.prepareConnectionToJoin bob 1 True connReq1 PQSupportOn
+    (aId, _) <- A.prepareConnectionToJoin bob 1 True connReq1 PQSupportOn
     void $ A.joinConnection bob NRMInteractive 1 aId True connReq1 "bob's connInfo" PQSupportOn SMSubscribe
     get alice =##> \case ("", _, A.ERR _) -> True; _ -> False
 
@@ -1192,11 +1207,11 @@ testAddressUpdatePreservesDRKeys ps = withSmpServer ps $ withAgentClients2 $ \al
     liftIO $ shortLink' `shouldBe` shortLink
     (_, ContactLinkData _ updated, connReq') <- getConnShortLink bob 1 shortLink
     liftIO $ ratchetKeys updated `shouldBe` ratchetKeys published
-    aliceId <- A.prepareConnectionToJoin bob 1 True connReq' bPQ
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True connReq' bPQ
     sqSecuredJoin <- A.joinConnection bob NRMInteractive 1 aliceId True connReq' "bob's connInfo" bPQ SMSubscribe
     liftIO $ sqSecuredJoin `shouldBe` False
-    ("", _, A.REQ invId _ _ "bob's connInfo" _) <- get alice
-    bobId <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
+    ("", _, A.REQ invId _ _ "bob's connInfo" _ _) <- get alice
+    (bobId, _) <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
     _ <- acceptContact alice 1 bobId True invId "alice's connInfo" (CR.connPQEncryption addrIK) SMSubscribe
     ("", _, A.CONF confId _ _ "alice's connInfo") <- get bob
     allowConfirmGreet alice bobId bob aliceId confId addrIK pqEnc
@@ -1216,10 +1231,10 @@ testAcceptContactDRResumeAfterOffline ps = withAgentClients2 $ \alice bob -> do
     (ccLink@(CCLink _ (Just shortLink)), preparedParams) <- A.prepareConnectionLink alice 1 rootKey linkEntId True Nothing connIK True Nothing
     _ <- A.createConnectionForLink alice NRMInteractive 1 True ccLink preparedParams (UserContactLinkData userCtData) SMSubscribe
     (_, _, connReq') <- getConnShortLink bob 1 shortLink
-    aId <- A.prepareConnectionToJoin bob 1 True connReq' PQSupportOn
+    (aId, _) <- A.prepareConnectionToJoin bob 1 True connReq' PQSupportOn
     _ <- A.joinConnection bob NRMInteractive 1 aId True connReq' "bob's connInfo" PQSupportOn SMSubscribe
-    ("", _, A.REQ invId _ _ "bob's connInfo" _) <- get alice
-    bId <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
+    ("", _, A.REQ invId _ _ "bob's connInfo" _ _) <- get alice
+    (bId, _) <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption addrIK)
     pure (bId, invId, aId)
   ("", "", DOWN _ _) <- nGet alice
   ("", "", DOWN _ _) <- nGet bob
@@ -1234,6 +1249,30 @@ testAcceptContactDRResumeAfterOffline ps = withAgentClients2 $ \alice bob -> do
     ("", _, A.CONF confId _ _ "alice's connInfo") <- get bob
     allowConfirmGreet alice bobId bob aliceId confId addrIK pqEnc
 
+testAcceptContactDRPrepared :: HasCallStack => Bool -> IO ()
+testAcceptContactDRPrepared ratchetDeleted =
+  withAgentClients2 $ \alice bob -> runRight_ $ do
+    let userLinkData = UserContactLinkData UserContactData {direct = True, owners = [], relays = [], userData = UserLinkData "test user data", ratchetKeys = Nothing}
+        pqEnc = PQEncryption $ pqConnectionMode IKPQOn PQSupportOn
+    (_, CCLink connReq _) <- A.createConnection alice NRMInteractive 1 True True SCMContact (Just userLinkData) Nothing IKPQOn True SMSubscribe
+    (aliceId, CRBRatchet codes) <- A.prepareConnectionToJoin bob 1 True connReq PQSupportOn
+    void $ A.joinConnection bob NRMInteractive 1 aliceId True connReq "bob's connInfo" PQSupportOn SMSubscribe
+    ("", _, A.REQ invId _ _ "bob's connInfo" (CRBRatchet reqCodes) _) <- get alice
+    (bobId, CRBRatchet acceptCodes) <- A.prepareConnectionToAccept alice 1 True invId PQSupportOn
+    liftIO $ do
+      reqCodes `shouldBe` codes
+      acceptCodes `shouldBe` codes
+    if ratchetDeleted
+      then do
+        liftIO $ withTransaction (store $ agentEnv alice) (`DB.execute_` "DELETE FROM ratchets")
+        Left (CONN NOT_FOUND _) <- tryError $ getConnectionVerifyCodes alice bobId
+        pure ()
+      else getConnectionVerifyCodes alice bobId >>= liftIO . (`shouldBe` codes)
+    void $ acceptContact alice 1 bobId True invId "alice's connInfo" PQSupportOn SMSubscribe
+    ("", _, A.CONF confId _ _ "alice's connInfo") <- get bob
+    allowConfirmGreet alice bobId bob aliceId confId IKPQOn pqEnc
+    getConnectionVerifyCodes alice bobId >>= liftIO . (`shouldBe` codes)
+
 runAgentClientContactTestPQ3 :: HasCallStack => Bool -> (AgentClient, InitialKeys) -> (AgentClient, PQSupport) -> (AgentClient, PQSupport) -> AgentMsgId -> IO ()
 runAgentClientContactTestPQ3 viaProxy (alice, aPQ) (bob, bPQ) (tom, tPQ) baseId = runRight_ $ do
   (_, CCLink qInfo Nothing) <- A.createConnection alice NRMInteractive 1 True True SCMContact Nothing Nothing aPQ False SMSubscribe
@@ -1244,12 +1283,12 @@ runAgentClientContactTestPQ3 viaProxy (alice, aPQ) (bob, bPQ) (tom, tPQ) baseId
   where
     msgId = subtract baseId . fst
     connectViaContact b pq qInfo = do
-      aId <- A.prepareConnectionToJoin b 1 True qInfo pq
+      (aId, _) <- A.prepareConnectionToJoin b 1 True qInfo pq
       sqSecuredJoin <- A.joinConnection b NRMInteractive 1 aId True qInfo "bob's connInfo" pq SMSubscribe
       liftIO $ sqSecuredJoin `shouldBe` False -- joining via contact address connection
-      ("", _, A.REQ invId pqSup' _ "bob's connInfo" _) <- get alice
+      ("", _, A.REQ invId pqSup' _ "bob's connInfo" _ _) <- get alice
       liftIO $ pqSup' `shouldBe` PQSupportOn
-      bId <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption aPQ)
+      (bId, _) <- A.prepareConnectionToAccept alice 1 True invId (CR.connPQEncryption aPQ)
       sqSecuredAccept <- acceptContact alice 1 bId True invId "alice's connInfo" (CR.connPQEncryption aPQ) SMSubscribe
       liftIO $ sqSecuredAccept `shouldBe` True
       ("", _, A.CONF confId pqSup'' _ "alice's connInfo") <- get b
@@ -1290,10 +1329,10 @@ testRejectContactRequest :: HasCallStack => IO ()
 testRejectContactRequest =
   withAgentClients2 $ \alice bob -> runRight_ $ do
     (_addrConnId, CCLink qInfo Nothing) <- A.createConnection alice NRMInteractive 1 True True SCMContact Nothing Nothing IKPQOn False SMSubscribe
-    aliceId <- A.prepareConnectionToJoin bob 1 True qInfo PQSupportOn
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True qInfo PQSupportOn
     sqSecured <- A.joinConnection bob NRMInteractive 1 aliceId True qInfo "bob's connInfo" PQSupportOn SMSubscribe
     liftIO $ sqSecured `shouldBe` False -- joining via contact address connection
-    ("", _, A.REQ invId PQSupportOn _ "bob's connInfo" _) <- get alice
+    ("", _, A.REQ invId PQSupportOn _ "bob's connInfo" _ _) <- get alice
     Left (A.CMD PROHIBITED _) <- tryError $ rejectContact alice NRMInteractive 1 invId (Just "no rejection without double ratchet")
     rejectContact alice NRMInteractive 1 invId Nothing
     liftIO $ noMessages bob "nothing delivered to bob"
@@ -1303,9 +1342,9 @@ testRejectContactRequestDR =
   withAgentClients2 $ \alice bob -> runRight_ $ do
     let userLinkData = UserContactLinkData UserContactData {direct = True, owners = [], relays = [], userData = UserLinkData "test user data", ratchetKeys = Nothing}
     (_addrConnId, CCLink connReq _) <- A.createConnection alice NRMInteractive 1 True True SCMContact (Just userLinkData) Nothing IKPQOn True SMSubscribe
-    aliceId <- A.prepareConnectionToJoin bob 1 True connReq PQSupportOn
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True connReq PQSupportOn
     void $ A.joinConnection bob NRMInteractive 1 aliceId True connReq "bob's connInfo" PQSupportOn SMSubscribe
-    ("", _, A.REQ invId _ _ "bob's connInfo" _) <- get alice
+    ("", _, A.REQ invId _ _ "bob's connInfo" _ _) <- get alice
     rejectContact alice NRMInteractive 1 invId (Just "not now")
     ("", _, A.RJCT "not now") <- get bob
     pure ()
@@ -1315,9 +1354,9 @@ testRejectContactRequestDRAsync =
   withAgentClients2 $ \alice bob -> runRight_ $ do
     let userLinkData = UserContactLinkData UserContactData {direct = True, owners = [], relays = [], userData = UserLinkData "test user data", ratchetKeys = Nothing}
     (_addrConnId, CCLink connReq _) <- A.createConnection alice NRMInteractive 1 True True SCMContact (Just userLinkData) Nothing IKPQOn True SMSubscribe
-    aliceId <- A.prepareConnectionToJoin bob 1 True connReq PQSupportOn
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True connReq PQSupportOn
     void $ A.joinConnection bob NRMInteractive 1 aliceId True connReq "bob's connInfo" PQSupportOn SMSubscribe
-    ("", _, A.REQ invId _ _ "bob's connInfo" _) <- get alice
+    ("", _, A.REQ invId _ _ "bob's connInfo" _ _) <- get alice
     rejectContactAsync alice "1" 1 invId (Just "not now")
     ("", _, A.RJCT "not now") <- get bob
     pure ()
@@ -1430,7 +1469,7 @@ testUpdateConnectionUserId =
     (connId, qInfo) <- createConnection alice 1 True SCMInvitation Nothing SMSubscribe
     newUserId <- createUser alice False [noAuthSrvCfg testSMPServer] [noAuthSrvCfg testXFTPServer]
     _ <- changeConnectionUser alice 1 connId newUserId
-    aliceId <- A.prepareConnectionToJoin bob 1 True qInfo PQSupportOn
+    (aliceId, _) <- A.prepareConnectionToJoin bob 1 True qInfo PQSupportOn
     sqSecured' <- A.joinConnection bob NRMInteractive 1 aliceId True qInfo "bob's connInfo" PQSupportOn SMSubscribe
     liftIO $ sqSecured' `shouldBe` True
     ("", _, A.CONF confId pqSup' _ "bob's connInfo") <- get alice
@@ -1589,7 +1628,7 @@ testInvitationErrors ps restart = do
   b <- getAgentB
   (bId, cReq) <- withServer1 ps $ runRight $ createConnection a 1 True SCMInvitation Nothing SMSubscribe
   ("", "", DOWN _ [_]) <- nGet a
-  aId <- runRight $ A.prepareConnectionToJoin b 1 True cReq PQSupportOn
+  (aId, _) <- runRight $ A.prepareConnectionToJoin b 1 True cReq PQSupportOn
   -- fails to secure the queue on testPort
   BROKER srv (NETWORK _) <- runLeft $ A.joinConnection b NRMInteractive 1 aId True cReq "bob's connInfo" PQSupportOn SMSubscribe
   (testPort `isSuffixOf` srv) `shouldBe` True
@@ -1659,7 +1698,7 @@ testContactErrors ps restart = do
   b <- getAgentB
   (contactId, cReq) <- withServer1 ps $ runRight $ createConnection a 1 True SCMContact Nothing SMSubscribe
   ("", "", DOWN _ [_]) <- nGet a
-  aId <- runRight $ A.prepareConnectionToJoin b 1 True cReq PQSupportOn
+  (aId, _) <- runRight $ A.prepareConnectionToJoin b 1 True cReq PQSupportOn
   -- fails to create queue on testPort2
   BROKER srv2 (NETWORK _) <- runLeft $ A.joinConnection b NRMInteractive 1 aId True cReq "bob's connInfo" PQSupportOn SMSubscribe
   (testPort2 `isSuffixOf` srv2) `shouldBe` True
@@ -1681,10 +1720,10 @@ testContactErrors ps restart = do
             Right r -> error $ "unexpected result " <> show r
             Left _ -> putStrLn "retrying send" >> threadDelay 200000 >> loopSend
     loopSend
-    ("", _, A.REQ invId PQSupportOn _ "bob's connInfo" _) <- get a
+    ("", _, A.REQ invId PQSupportOn _ "bob's connInfo" _ _) <- get a
     pure invId
   ("", "", DOWN _ [_]) <- nGet a
-  bId <- runRight $ A.prepareConnectionToAccept a 1 True invId PQSupportOn
+  (bId, _) <- runRight $ A.prepareConnectionToAccept a 1 True invId PQSupportOn
   withServer2 ps $ do
     ("", "", UP _ [_]) <- nGet b''
     let loopSecure = do
@@ -1764,7 +1803,7 @@ testInvitationShortLink viaProxy a b =
 
 testJoinConn_ :: Bool -> Bool -> AgentClient -> ConnId -> AgentClient -> ConnectionRequestUri c -> ExceptT AgentErrorType IO ()
 testJoinConn_ viaProxy sndSecure a bId b connReq = do
-  aId <- A.prepareConnectionToJoin b 1 True connReq PQSupportOn
+  (aId, _) <- A.prepareConnectionToJoin b 1 True connReq PQSupportOn
   sndSecure' <- A.joinConnection b NRMInteractive 1 aId True connReq "bob's connInfo" PQSupportOn SMSubscribe
   liftIO $ sndSecure' `shouldBe` sndSecure
   ("", _, CONF confId _ "bob's connInfo") <- get a
@@ -1774,14 +1813,6 @@ testJoinConn_ viaProxy sndSecure a bId b connReq = do
   get b ##> ("", aId, CON)
   exchangeGreetingsViaProxy viaProxy a bId b aId
 
-testInvitationShortLinkPrev :: HasCallStack => Bool -> Bool -> AgentClient -> AgentClient -> IO ()
-testInvitationShortLinkPrev viaProxy sndSecure a b = runRight_ $ do
-  let userData = UserLinkData "some user data"
-      newLinkData = UserInvLinkData userData
-  -- can't create short link with previous version
-  (bId, CCLink connReq Nothing) <- A.createConnection a NRMInteractive 1 True True SCMInvitation (Just newLinkData) Nothing CR.IKPQOn False SMSubscribe
-  testJoinConn_ viaProxy sndSecure a bId b connReq
-
 testInvitationShortLinkAsync :: HasCallStack => Bool -> AgentClient -> AgentClient -> IO ()
 testInvitationShortLinkAsync viaProxy a b = do
   let userData = UserLinkData "some user data"
@@ -1792,7 +1823,7 @@ testInvitationShortLinkAsync viaProxy a b = do
   connReq' `shouldBe` connReq
   linkUserData connData' `shouldBe` userData
   runRight $ do
-    aId <- A.prepareConnectionToJoin b 1 True connReq PQSupportOn
+    (aId, _) <- A.prepareConnectionToJoin b 1 True connReq PQSupportOn
     A.joinConnectionAsync b "123" False aId True connReq "bob's connInfo" PQSupportOn SMSubscribe
     get b =##> \case ("123", c, JOINED sndSecure) -> c == aId && sndSecure; _ -> False
     ("", _, CONF confId _ "bob's connInfo") <- get a
@@ -1834,7 +1865,7 @@ testContactShortLink viaProxy a b =
       (aId, sndSecure) <- joinConnection b 1 True connReq "bob's connInfo" SMSubscribe
       liftIO $ sndSecure `shouldBe` False
       ("", _, REQ invId _ "bob's connInfo") <- get a
-      bId <- A.prepareConnectionToAccept a 1 True invId PQSupportOn
+      (bId, _) <- A.prepareConnectionToAccept a 1 True invId PQSupportOn
       sndSecure' <- acceptContact a 1 bId True invId "alice's connInfo" PQSupportOn SMSubscribe
       liftIO $ sndSecure' `shouldBe` True
       ("", _, CONF confId  _ "alice's connInfo") <- get b
@@ -1887,7 +1918,7 @@ testAddContactShortLink viaProxy a b =
       (aId, sndSecure) <- joinConnection b 1 True connReq "bob's connInfo" SMSubscribe
       liftIO $ sndSecure `shouldBe` False
       ("", _, REQ invId _ "bob's connInfo") <- get a
-      bId <- A.prepareConnectionToAccept a 1 True invId PQSupportOn
+      (bId, _) <- A.prepareConnectionToAccept a 1 True invId PQSupportOn
       sndSecure' <- acceptContact a 1 bId True invId "alice's connInfo" PQSupportOn SMSubscribe
       liftIO $ sndSecure' `shouldBe` True
       ("", _, CONF confId  _ "alice's connInfo") <- get b
@@ -2046,7 +2077,7 @@ testPrepareCreateConnectionLink ps = withSmpServer ps $ withAgentClients2 $ \a b
     (bId, sndSecure) <- joinConnection b 1 True connReq' "bob's connInfo" SMSubscribe
     liftIO $ sndSecure `shouldBe` False
     ("", _, REQ invId _ "bob's connInfo") <- get a
-    aId <- A.prepareConnectionToAccept a 1 True invId PQSupportOn
+    (aId, _) <- A.prepareConnectionToAccept a 1 True invId PQSupportOn
     sndSecure' <- acceptContact a 1 aId True invId "alice's connInfo" PQSupportOn SMSubscribe
     liftIO $ sndSecure' `shouldBe` True
     ("", _, CONF confId _ "alice's connInfo") <- get b
@@ -2056,6 +2087,64 @@ testPrepareCreateConnectionLink ps = withSmpServer ps $ withAgentClients2 $ \a b
     get b ##> ("", bId, CON)
     exchangeGreetings a aId b bId
 
+testPrepareCreateInvitationLink :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testPrepareCreateInvitationLink ps = withSmpServer ps $ withAgentClients2 $ \a b -> do
+  let userData = UserLinkData "test user data"
+  runRight_ $ do
+    (preparedReq, preparedParams@PreparedLinkParams {plpLinkKey}) <- A.prepareInvitationLink a 1 CR.IKPQOn Nothing
+    (bId, CCLink connReq (Just shortLink@(CSLInvitation _ _ _ linkKey))) <- A.createInvitationForLink a NRMInteractive 1 True preparedReq preparedParams (UserInvLinkData userData) SMSubscribe
+    liftIO $ linkKey `shouldBe` plpLinkKey
+    (_, connData', connReq') <- getConnShortLink b 1 shortLink
+    liftIO $ connReq' `shouldBe` preparedReq
+    liftIO $ connReq `shouldNotBe` preparedReq
+    liftIO $ linkUserData connData' `shouldBe` userData
+    testJoinConn_ False True a bId b connReq
+
+testPrepareInvitationLinkTooLarge :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testPrepareInvitationLinkTooLarge ps = withSmpServer ps $ withAgent 1 agentCfg initAgentServers testDB $ \a -> do
+  runRight_ $ do
+    (preparedReq, preparedParams) <- A.prepareInvitationLink a 1 CR.IKPQOn Nothing
+    Left (A.CMD LARGE _) <- tryError $ A.createInvitationForLink a NRMInteractive 1 True preparedReq preparedParams (UserInvLinkData $ UserLinkData $ B.replicate 14000 'a') SMSubscribe
+    pure ()
+  withTransaction (store $ agentEnv a) getConnIds `shouldReturn` []
+
+testPrepareContactLinkTooLarge :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testPrepareContactLinkTooLarge ps = withSmpServer ps $ withAgent 1 agentCfg initAgentServers testDB $ \a -> do
+  let userCtData = UserContactData {direct = True, owners = [], relays = [], userData = UserLinkData $ B.replicate 14000 'a', ratchetKeys = Nothing}
+  g <- C.newRandom
+  rootKey <- atomically $ C.generateKeyPair g
+  linkEntId <- atomically $ C.randomBytes 32 g
+  runRight_ $ do
+    (ccLink, preparedParams) <- A.prepareConnectionLink a 1 rootKey linkEntId True Nothing CR.IKPQOn True Nothing
+    Left (A.CMD LARGE _) <- tryError $ A.createConnectionForLink a NRMInteractive 1 True ccLink preparedParams (UserContactLinkData userCtData) SMSubscribe
+    pure ()
+  withTransaction (store $ agentEnv a) getConnIds `shouldReturn` []
+
+testPrepareConnShortLink :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testPrepareConnShortLink ps = withSmpServer ps $ withAgentClients2 $ \a b -> do
+  let userCtData = UserContactData {direct = True, owners = [], relays = [], userData = UserLinkData "test user data", ratchetKeys = Nothing}
+  runRight_ $ do
+    (contactId, CCLink _ Nothing) <- A.createConnection a NRMInteractive 1 True True SCMContact Nothing Nothing CR.IKPQOn False SMSubscribe
+    shortLink <- A.prepareConnShortLink a contactId Nothing
+    shortLink' <- A.prepareConnShortLink a contactId Nothing
+    liftIO $ shortLink' `shouldBe` shortLink
+    shortLink'' <- A.setConnShortLink a NRMInteractive contactId SCMContact (UserContactLinkData userCtData) Nothing False Nothing
+    liftIO $ shortLink'' `shouldBe` shortLink
+    (_, ContactLinkData _ userCtData', connReq) <- getConnShortLink b 1 shortLink
+    liftIO $ userCtData' `shouldBe` userCtData
+    (aId, sndSecure) <- joinConnection b 1 True connReq "bob's connInfo" SMSubscribe
+    liftIO $ sndSecure `shouldBe` False
+    ("", _, REQ invId _ "bob's connInfo") <- get a
+    (bId, _) <- A.prepareConnectionToAccept a 1 True invId PQSupportOn
+    sndSecure' <- acceptContact a 1 bId True invId "alice's connInfo" PQSupportOn SMSubscribe
+    liftIO $ sndSecure' `shouldBe` True
+    ("", _, CONF confId _ "alice's connInfo") <- get b
+    allowConnection b aId confId "bob's connInfo"
+    get a ##> ("", bId, INFO "bob's connInfo")
+    get a ##> ("", bId, CON)
+    get b ##> ("", aId, CON)
+    exchangeGreetings a bId b aId
+
 connReqWithKeys :: BinaryConnectionRequestUri m -> Maybe AddressRatchetKeys -> ConnectionRequestUri m
 connReqWithKeys cr rk = case cr of
   BCRInvitationUri crData e2eParams -> CRInvitationUri crData e2eParams
@@ -2484,6 +2573,49 @@ testExpireManyMessagesQuota (t, msType) = withSmpServerConfigOn t cfg' testPort
   where
     cfg' = updateCfg (cfgMS msType) $ \cfg_ -> cfg_ {msgQueueQuota = 1, maxJournalMsgCount = 2}
 
+testJoinFullContactAsync :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testJoinFullContactAsync (t, msType) = withSmpServerConfigOn t cfg' testPort $ \_ -> do
+  (contactId, qInfo) <- fillContactAddress
+  withAgent 2 agentCfg {commandQuotaRetryInterval = fastRetryInterval} initAgentServers testDB2 $ \bob -> do
+    aliceId <- runRight $ do
+      (aliceId, _) <- A.prepareConnectionToJoin bob 1 True qInfo PQSupportOn
+      A.joinConnectionAsync bob "2" False aliceId True qInfo "bob's connInfo" PQSupportOn SMSubscribe
+      pure aliceId
+    threadDelay 500000
+    noMessages bob "joining should be retried while the address is full"
+    withAgent 1 agentCfg initAgentServers testDB $ \alice -> runRight_ $ do
+      subscribeConnection alice contactId
+      ("", _, A.REQ _ _ _ "carol's connInfo" _ _) <- get alice
+      get bob =##> \case ("2", c, JOINED False) -> c == aliceId; _ -> False
+      ("", _, A.REQ _ _ _ "bob's connInfo" _ _) <- get alice
+      pure ()
+  where
+    cfg' = updateCfg (cfgMS msType) $ \cfg_ -> cfg_ {msgQueueQuota = 1, maxJournalMsgCount = 2}
+
+testJoinFullContactAsyncExpire :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testJoinFullContactAsyncExpire (t, msType) = withSmpServerConfigOn t cfg' testPort $ \_ -> do
+  (_, qInfo) <- fillContactAddress
+  withAgent 2 agentCfg {quotaExceededTimeout = 2, commandQuotaRetryInterval = fastRetryInterval} initAgentServers testDB2 $ \bob -> do
+    aliceId <- runRight $ do
+      (aliceId, _) <- A.prepareConnectionToJoin bob 1 True qInfo PQSupportOn
+      A.joinConnectionAsync bob "2" False aliceId True qInfo "bob's connInfo" PQSupportOn SMSubscribe
+      pure aliceId
+    threadDelay 500000
+    noMessages bob "joining should be retried until quota exceeded timeout"
+    get bob =##> \case ("2", c, ERR (SMP _ QUOTA)) -> c == aliceId; _ -> False
+  where
+    cfg' = updateCfg (cfgMS msType) $ \cfg_ -> cfg_ {msgQueueQuota = 1, maxJournalMsgCount = 2}
+
+fillContactAddress :: HasCallStack => IO (ConnId, ConnectionRequestUri 'CMContact)
+fillContactAddress = do
+  (contactId, qInfo) <- withAgent 1 agentCfg initAgentServers testDB $ \alice -> runRight $ do
+    (contactId, CCLink qInfo Nothing) <- A.createConnection alice NRMInteractive 1 True True SCMContact Nothing Nothing IKPQOn False SMSubscribe
+    pure (contactId, qInfo)
+  withAgent 3 agentCfg initAgentServers testDB3 $ \carol -> runRight_ $ do
+    (aliceId, _) <- A.prepareConnectionToJoin carol 1 True qInfo PQSupportOn
+    void $ A.joinConnection carol NRMInteractive 1 aliceId True qInfo "carol's connInfo" PQSupportOn SMSubscribe
+  pure (contactId, qInfo)
+
 testDropMsgAfterRcvAttempts :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
 testDropMsgAfterRcvAttempts ps =
   withSmpServerStoreLogOn ps testPort $ \_ -> do
@@ -2698,6 +2830,117 @@ testRatchetSyncSimultaneous ps = do
   disposeAgentClient bob
   disposeAgentClient bob2
 
+testRatchetSyncReplayedKey :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testRatchetSyncReplayedKey ps = withAgentClients2 $ \alice bob -> do
+  (aliceId, bobId, bob2) <- withSmpServerStoreMsgLogOn ps testPort $ \_ ->
+    setupDesynchronizedRatchet alice bob
+  ("", "", DOWN _ _) <- nGet alice
+  ("", "", DOWN _ _) <- nGet bob2
+  _ <- runRight $ synchronizeRatchet bob2 aliceId PQSupportOn False
+  Right pks <- withTransaction (store $ agentEnv bob2) (`getRatchetX3dhKeys` aliceId)
+  Right (SomeConn _ (DuplexConnection _ _ (sq :| _))) <- withTransaction (store $ agentEnv bob2) (`getConn` aliceId)
+  withSmpServerStoreMsgLogOn ps testPort $ \_ -> do
+    concurrently_
+      (getInAnyOrder alice [ratchetSyncP' bobId RSAgreed, serverUpP])
+      (getInAnyOrder bob2 [ratchetSyncP' aliceId RSAgreed, serverUpP])
+    get alice =##> ratchetSyncP bobId RSOk
+    get bob2 =##> ratchetSyncP aliceId RSOk
+    withTransaction (store $ agentEnv alice) $ \db -> deleteRatchetKeyHashesExpired db 0 100
+    let keyMsg = AgentRatchetKey {agentVersion = currentSMPAgentVersion, e2eEncryption = CR.mkRcvE2ERatchetParams CR.currentE2EEncryptVersion pks, info = ""}
+    Right _ <- runReaderT (runExceptT $ sendAgentMessage bob2 sq SMP.noMsgFlags $ smpEncode keyMsg) (agentEnv bob2)
+    runRight_ $ exchangeGreetingsMsgIds alice bobId 10 bob2 aliceId 7
+  disposeAgentClient bob2
+
+testRatchetSyncRepeated :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testRatchetSyncRepeated ps = withAgentClients2 $ \alice bob -> do
+  (aliceId, bobId, bob2) <- startRatchetSyncOffline ps alice bob
+  ConnectionStats {ratchetSyncState = rss2} <- runRight $ synchronizeRatchet bob2 aliceId PQSupportOn True
+  rss2 `shouldBe` RSStarted
+
+  withSmpServerStoreMsgLogOn ps testPort $ \_ -> do
+    concurrently_
+      (getInAnyOrder alice [ratchetSyncP' bobId RSAgreed, serverUpP])
+      (getInAnyOrder bob2 [ratchetSyncP' aliceId RSAgreed, serverUpP])
+    runRight_ $ do
+      get alice =##> ratchetSyncP bobId RSAgreed
+      get alice =##> ratchetSyncP bobId RSOk
+      get bob2 =##> ratchetSyncP aliceId RSOk
+      msgId <- sendMessage alice bobId SMP.noMsgFlags "hello"
+      get alice ##> ("", bobId, SENT msgId)
+      get bob2 =##> \case ("", c, Msg "hello") -> c == aliceId; _ -> False
+      ackMessage bob2 aliceId 8 Nothing
+  map fst <$> processedRatchetKeyHashes bob2 `shouldReturn` [aliceId, aliceId]
+  disposeAgentClient bob2
+
+testRatchetSyncFailedKeyNotProcessed :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testRatchetSyncFailedKeyNotProcessed ps = withAgentClients2 $ \alice bob -> do
+  (aliceId, bobId, bob2) <- startRatchetSyncOffline ps alice bob
+  withTransaction (store $ agentEnv bob2) $ \db ->
+    DB.execute_ db "UPDATE ratchets SET x3dh_priv_key_1 = NULL"
+
+  withSmpServerStoreMsgLogOn ps testPort $ \_ ->
+    concurrently_
+      (getInAnyOrder alice [ratchetSyncP' bobId RSAgreed, serverUpP])
+      (getInAnyOrder bob2 [x3dhKeysNotFoundP aliceId, serverUpP])
+  map fst <$> processedRatchetKeyHashes alice `shouldReturn` [bobId]
+  processedRatchetKeyHashes bob2 `shouldReturn` []
+  disposeAgentClient bob2
+  where
+    x3dhKeysNotFoundP :: ConnId -> ATransmission -> Bool
+    x3dhKeysNotFoundP cId = \case
+      (_, cId', AEvt SAEConn (ERR (A.INTERNAL e))) -> cId' == cId && "SEX3dhKeysNotFound" `isPrefixOf` e
+      _ -> False
+
+testRatchetSyncFailedRecreationNoReply :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testRatchetSyncFailedRecreationNoReply ps = withAgentClients2 $ \alice bob -> do
+  (_, bobId, bob2) <- startRatchetSyncOffline ps alice bob
+  aliceSndMsgs <- sndMessages alice
+  withTransaction (store $ agentEnv alice) $ \db ->
+    DB.execute_ db "CREATE TRIGGER fail_ratchet_insert BEFORE INSERT ON ratchets BEGIN SELECT RAISE(ABORT, 'ratchet insert failed'); END"
+  withSmpServerStoreMsgLogOn ps testPort $ \_ ->
+    concurrently_
+      (getInAnyOrder alice [ratchetInsertFailedP bobId, serverUpP])
+      (getInAnyOrder bob2 [serverUpP])
+  processedRatchetKeyHashes alice `shouldReturn` []
+  sndMessages alice `shouldReturn` aliceSndMsgs
+  disposeAgentClient bob2
+  where
+    ratchetInsertFailedP :: ConnId -> ATransmission -> Bool
+    ratchetInsertFailedP cId = \case
+      (_, cId', AEvt SAEConn (ERR (A.INTERNAL e))) -> cId' == cId && "ratchet insert failed" `isInfixOf` e
+      _ -> False
+
+testRatchetSyncStartFailedNoKey :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testRatchetSyncStartFailedNoKey ps = withAgentClients2 $ \alice bob -> do
+  (aliceId, _, bob2) <- withSmpServerStoreMsgLogOn ps testPort $ \_ ->
+    setupDesynchronizedRatchet alice bob
+  bobSndMsgs <- sndMessages bob2
+  withTransaction (store $ agentEnv bob2) $ \db ->
+    DB.execute_ db "CREATE TRIGGER fail_ratchet_update BEFORE UPDATE ON ratchets BEGIN SELECT RAISE(ABORT, 'ratchet update failed'); END"
+  Left (A.INTERNAL e) <- runExceptT $ synchronizeRatchet bob2 aliceId PQSupportOff False
+  e `shouldContain` "ratchet update failed"
+  ConnectionStats {ratchetSyncState} <- runRight $ getConnectionServers bob2 aliceId
+  ratchetSyncState `shouldBe` RSRequired
+  withTransaction (store $ agentEnv bob2) (`DB.query_` "SELECT conn_id, pq_support FROM connections") `shouldReturn` [(aliceId, PQSupportOn)]
+  sndMessages bob2 `shouldReturn` bobSndMsgs
+  disposeAgentClient bob2
+
+startRatchetSyncOffline :: HasCallStack => (ASrvTransport, AStoreType) -> AgentClient -> AgentClient -> IO (ConnId, ConnId, AgentClient)
+startRatchetSyncOffline ps alice bob = do
+  (aliceId, bobId, bob2) <- withSmpServerStoreMsgLogOn ps testPort $ \_ ->
+    setupDesynchronizedRatchet alice bob
+  ("", "", DOWN _ _) <- nGet alice
+  ("", "", DOWN _ _) <- nGet bob2
+  ConnectionStats {ratchetSyncState} <- runRight $ synchronizeRatchet bob2 aliceId PQSupportOn False
+  ratchetSyncState `shouldBe` RSStarted
+  pure (aliceId, bobId, bob2)
+
+processedRatchetKeyHashes :: AgentClient -> IO [(ConnId, ByteString)]
+processedRatchetKeyHashes c = withTransaction (store $ agentEnv c) (`DB.query_` "SELECT conn_id, hash FROM processed_ratchet_key_hashes")
+
+sndMessages :: AgentClient -> IO [(ConnId, Int64)]
+sndMessages c = withTransaction (store $ agentEnv c) (`DB.query_` "SELECT conn_id, internal_id FROM snd_messages ORDER BY conn_id, internal_id")
+
 getMsg :: AgentClient -> ConnId -> ExceptT AgentErrorType IO a -> ExceptT AgentErrorType IO a
 getMsg c cId action = do
   liftIO $ noMessages c "nothing should be delivered before GET"
@@ -2749,7 +2992,7 @@ makeConnectionForUsers = makeConnectionForUsers_ PQSupportOn
 makeConnectionForUsers_ :: HasCallStack => PQSupport -> AgentClient -> UserId -> AgentClient -> UserId -> ExceptT AgentErrorType IO (ConnId, ConnId)
 makeConnectionForUsers_ pqSupport alice aliceUserId bob bobUserId = do
   (bobId, CCLink qInfo Nothing) <- A.createConnection alice NRMInteractive aliceUserId True True SCMInvitation Nothing Nothing (IKLinkPQ pqSupport) False SMSubscribe
-  aliceId <- A.prepareConnectionToJoin bob bobUserId True qInfo pqSupport
+  (aliceId, _) <- A.prepareConnectionToJoin bob bobUserId True qInfo pqSupport
   sqSecured' <- A.joinConnection bob NRMInteractive bobUserId aliceId True qInfo "bob's connInfo" pqSupport SMSubscribe
   liftIO $ sqSecured' `shouldBe` True
   ("", _, A.CONF confId pqSup' _ "bob's connInfo") <- get alice
@@ -3031,7 +3274,7 @@ testAsyncCommands alice bob baseId =
     createConnectionAsync alice "1" bobId True SCMInvitation IKPQOn False SMSubscribe
     ("1", bobId', INV (ACR _ qInfo)) <- get alice
     liftIO $ bobId' `shouldBe` bobId
-    aliceId <- prepareConnectionToJoin bob 1 True qInfo PQSupportOn
+    (aliceId, _) <- prepareConnectionToJoin bob 1 True qInfo PQSupportOn
     joinConnectionAsync bob "2" False aliceId True qInfo "bob's connInfo" PQSupportOn SMSubscribe
     ("2", aliceId', JOINED sqSecured') <- get bob
     liftIO $ do
@@ -3101,7 +3344,7 @@ testSetConnShortLinkAsync ps = withAgentClients2 $ \alice bob ->
     -- complete connection via contact address
     (aliceId, _) <- joinConnection bob 1 True qInfo "bob's connInfo" SMSubscribe
     ("", _, REQ invId _ "bob's connInfo") <- get alice
-    bobId <- A.prepareConnectionToAccept alice 1 True invId PQSupportOn
+    (bobId, _) <- A.prepareConnectionToAccept alice 1 True invId PQSupportOn
     _ <- acceptContact alice 1 bobId True invId "alice's connInfo" PQSupportOn SMSubscribe
     ("", _, CONF confId _ "alice's connInfo") <- get bob
     allowConnection bob aliceId confId "bob's connInfo"
@@ -3129,7 +3372,7 @@ testGetConnShortLinkAsync ps = withAgentClients2 $ \alice bob ->
     liftIO $ aliceId' `shouldBe` aliceId
     -- complete connection
     ("", _, REQ invId _ "bob's connInfo") <- get alice
-    bobId <- A.prepareConnectionToAccept alice 1 True invId PQSupportOn
+    (bobId, _) <- A.prepareConnectionToAccept alice 1 True invId PQSupportOn
     _ <- acceptContact alice 1 bobId True invId "alice's connInfo" PQSupportOn SMSubscribe
     ("", _, CONF confId _ "alice's connInfo") <- get bob
     allowConnection bob aliceId confId "bob's connInfo"
@@ -3159,7 +3402,7 @@ testAcceptContactAsync alice bob baseId =
     (aliceId, sqSecuredJoin) <- joinConnection bob 1 True qInfo "bob's connInfo" SMSubscribe
     liftIO $ sqSecuredJoin `shouldBe` False -- joining via contact address connection
     ("", _, REQ invId _ "bob's connInfo") <- get alice
-    bobId <- prepareConnectionToAccept alice 1 True invId PQSupportOn
+    (bobId, _) <- prepareConnectionToAccept alice 1 True invId PQSupportOn
     acceptContactAsync alice "1" bobId True invId "alice's connInfo" PQSupportOn SMSubscribe
     get alice =##> \case ("1", c, JOINED sqSecured') -> c == bobId && sqSecured' == True; _ -> False
     ("", _, CONF confId _ "alice's connInfo") <- get bob
@@ -3440,7 +3683,7 @@ testJoinConnectionAsyncReplyError ps@(t, ASType qsType _) = do
         createConnectionAsync a "1" bId True SCMInvitation IKPQOn False SMSubscribe
         ("1", bId', INV (ACR _ qInfo)) <- get a
         liftIO $ bId' `shouldBe` bId
-        aId <- prepareConnectionToJoin b 1 True qInfo PQSupportOn
+        (aId, _) <- prepareConnectionToJoin b 1 True qInfo PQSupportOn
         joinConnectionAsync b "2" False aId True qInfo "bob's connInfo" PQSupportOn SMSubscribe
         liftIO $ threadDelay 500000
         ConnectionStats {rcvQueuesInfo = [], sndQueuesInfo = [SndQueueInfo {}]} <- getConnectionServers b aId
@@ -3580,7 +3823,7 @@ testFastSwitchDeadOldServer :: HasCallStack => (ASrvTransport, AStoreType) -> IO
 testFastSwitchDeadOldServer ps@(t, ASType qsType _) = do
   let bServers = initAgentServers {smp = userServers [testSMPServer2]}
   withSmpServerConfigOn t (cfgJ2QS qsType) testPort2 $ \_ ->
-    withAgent 1 agentCfg initAgentServers testDB $ \a ->
+    withAgent 1 aCfg initAgentServers testDB $ \a ->
       withAgent 2 agentCfg bServers testDB2 $ \b -> do
         (aId, bId) <- withSmpServerStoreLogOn ps testPort $ \_ -> runRight $ do
           (aId, bId) <- makeConnection a b
@@ -3596,7 +3839,10 @@ testFastSwitchDeadOldServer ps@(t, ASType qsType _) = do
           queuedReceived <- drainSwitchCompletedRcvMsg a bId "queued while down"
           liftIO $ queuedReceived `shouldBe` True
           drainSwitchCompleted b aId QDSnd
+          waitOneRcvQueue a bId
           exchangeGreetingsMsgId 7 a bId b aId
+  where
+    aCfg = agentCfg {initialCleanupDelay = 10000, cleanupInterval = 10000, deleteErrorCount = 3}
 
 -- drains switch and network events until the connection reports SPCompleted in the given direction,
 -- tolerating DOWN/UP and intermediate phases (the old server is stopped mid-rotation)
@@ -3616,6 +3862,69 @@ drainSwitchCompletedRcvMsg c connId body = go False
         (_, connId', AEvt SAEConn (Msg' mId _ body')) | connId' == connId && body' == body -> ackMessage c connId' mId Nothing >> go True
         _ -> go seen
 
+testFastSwitchDrainOldQueue :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testFastSwitchDrainOldQueue ps =
+  withSmpServers2 ps $ withAgentClients2 $ \a b -> runRight_ $ do
+    (aId, bId, aM1Id) <- endSwitchOnNewQueue a b
+    ackMessage a bId aM1Id Nothing
+    drainOldQueue a bId b aId
+
+testFastSwitchDrainOldQueueRestart :: HasCallStack => (ASrvTransport, AStoreType) -> IO ()
+testFastSwitchDrainOldQueueRestart ps =
+  withSmpServers2 ps $
+    withAgent 2 agentCfg initAgentServers testDB2 $ \b -> do
+      (aId, bId, _) <- withAgent 1 agentCfg initAgentServers testDB $ \a -> runRight $ endSwitchOnNewQueue a b
+      withAgent 1 agentCfg initAgentServers testDB $ \a -> runRight_ $ do
+        subscribeConnection a bId
+        drainOldQueue a bId b aId
+
+endSwitchOnNewQueue :: HasCallStack => AgentClient -> AgentClient -> ExceptT AgentErrorType IO (ConnId, ConnId, AgentMsgId)
+endSwitchOnNewQueue a b = do
+  (aId, bId) <- makeConnection a b
+  exchangeGreetings a bId b aId
+  liftIO $ setProtocolServers a 1 [noAuthSrvCfg testSMPServer2]
+  m1Id <- sendMessage b aId SMP.noMsgFlags "m1"
+  get b ##> ("", aId, SENT m1Id)
+  ("", c, Msg' aM1Id _ "m1") <- get a
+  liftIO $ c `shouldBe` bId
+  m2Id <- sendMessage b aId SMP.noMsgFlags "m2"
+  get b ##> ("", aId, SENT m2Id)
+  _ <- switchConnectionAsync a "" bId
+  fastSwitchComplete a bId b aId
+  stats <- getConnectionServers a bId
+  liftIO $ rcvSwchStatuses' stats `shouldMatchList` [Nothing, Just RSReceivedQEND]
+  pure (aId, bId, aM1Id)
+
+drainOldQueue :: HasCallStack => AgentClient -> ConnId -> AgentClient -> ConnId -> ExceptT AgentErrorType IO ()
+drainOldQueue a bId b aId = do
+  ("", c, MSG MsgMeta {recipient = (aM2Id, _)} _ "m2") <- get a
+  liftIO $ c `shouldBe` bId
+  ackMessage a bId aM2Id Nothing
+  waitOneRcvQueue a bId
+  bMsgId <- sendMessage b aId SMP.noMsgFlags "hello"
+  get b ##> ("", aId, SENT bMsgId)
+  ("", c', MSG MsgMeta {recipient = (aMsgId, _)} _ "hello") <- get a
+  liftIO $ c' `shouldBe` bId
+  ackMessage a bId aMsgId Nothing
+  aMsgId' <- sendMessage a bId SMP.noMsgFlags "hello too"
+  get a ##> ("", bId, SENT aMsgId')
+  ("", c'', MSG MsgMeta {recipient = (bMsgId', _)} _ "hello too") <- get b
+  liftIO $ c'' `shouldBe` aId
+  ackMessage b aId bMsgId' Nothing
+  liftIO $ noMessages a "nothing else should be delivered to alice"
+  liftIO $ noMessages b "nothing else should be delivered to bob"
+
+waitOneRcvQueue :: HasCallStack => AgentClient -> ConnId -> ExceptT AgentErrorType IO ()
+waitOneRcvQueue c connId = go (300 :: Int)
+  where
+    go n = do
+      stats <- getConnectionServers c connId
+      case rcvSwchStatuses' stats of
+        [Nothing] -> pure ()
+        statuses
+          | n > 0 -> liftIO (threadDelay 100000) >> go (n - 1)
+          | otherwise -> liftIO $ expectationFailure $ "old receive queue was not deleted: " <> show statuses
+
 phaseRcv :: AgentClient -> ByteString -> SwitchPhase -> [Maybe RcvSwitchStatus] -> ExceptT AgentErrorType IO ()
 phaseRcv c connId p swchStatuses = phase c connId QDRcv p (\stats -> rcvSwchStatuses' stats `shouldMatchList` swchStatuses)
 
@@ -4023,13 +4332,23 @@ testServerInformation =
       serverCountry = Nothing
     }
 
-testRatchetAdHash :: HasCallStack => IO ()
-testRatchetAdHash =
+testConnectionVerifyCodes :: HasCallStack => IO ()
+testConnectionVerifyCodes =
   withAgentClients2 $ \a b -> runRight_ $ do
     (aId, bId) <- makeConnection a b
-    ad1 <- getConnectionRatchetAdHash a bId
-    ad2 <- getConnectionRatchetAdHash b aId
-    liftIO $ ad1 `shouldBe` ad2
+    codes1 <- getConnectionVerifyCodes a bId
+    codes2 <- getConnectionVerifyCodes b aId
+    liftIO $ do
+      codes1 `shouldBe` codes2
+      codePQ codes1 `shouldNotBe` Nothing
+    liftIO $ withTransaction (store $ agentEnv a) $ \db ->
+      DB.execute_ db "UPDATE ratchets SET rc_verify_code_ad = NULL, rc_verify_code_pq = NULL"
+    codes1' <- getConnectionVerifyCodes a bId
+    liftIO $ codes1' `shouldBe` codes1
+    liftIO $ withTransaction (store $ agentEnv a) $ \db ->
+      DB.execute_ db "UPDATE ratchets SET ratchet_state = NULL"
+    codes1'' <- getConnectionVerifyCodes a bId
+    liftIO $ codes1'' `shouldBe` codes1
 
 testDeliveryReceipts :: HasCallStack => IO ()
 testDeliveryReceipts =
@@ -4618,8 +4937,7 @@ testServerMultipleIdentities =
         testE2ERatchetParams12
 
 testWaitForUserNetwork :: IO ()
-testWaitForUserNetwork = do
-  a <- getSMPAgentClient' 1 aCfg initAgentServers testDB
+testWaitForUserNetwork = withAgent 1 aCfg initAgentServers testDB $ \a -> do
   noNetworkDelay a
   setUserNetworkInfo a $ UserNetworkInfo UNNone False
   networkDelay a 100000
@@ -4636,8 +4954,7 @@ testWaitForUserNetwork = do
     aCfg = agentCfg {userNetworkInterval = 100000, userOfflineDelay = 0}
 
 testDoNotResetOnlineToOffline :: IO ()
-testDoNotResetOnlineToOffline = do
-  a <- getSMPAgentClient' 1 aCfg initAgentServers testDB
+testDoNotResetOnlineToOffline = withAgent 1 aCfg initAgentServers testDB $ \a -> do
   noNetworkDelay a
   setUserNetworkInfo a $ UserNetworkInfo UNWifi False
   networkDelay a 100000
@@ -4658,8 +4975,7 @@ testDoNotResetOnlineToOffline = do
     aCfg = agentCfg {userNetworkInterval = 100000, userOfflineDelay = 0.1}
 
 testResumeMultipleThreads :: IO ()
-testResumeMultipleThreads = do
-  a <- getSMPAgentClient' 1 aCfg initAgentServers testDB
+testResumeMultipleThreads = withAgent 1 aCfg initAgentServers testDB $ \a -> do
   noNetworkDelay a
   setUserNetworkInfo a $ UserNetworkInfo UNNone False
   vs <-
@@ -4705,7 +5021,7 @@ testServerQueueInfo = do
     Just srvMsgId <- checkMsgQ bob aliceId 1
     get bob =##> \case
       ("", c, MSG MsgMeta {integrity = MsgOk, broker = (smId, _), recipient = (mId, _), pqEncryption = PQEncOn} _ "hello") ->
-        c == aliceId && decodeLatin1 (B64.encode smId) == srvMsgId && mId == msgId
+        c == aliceId && decodeLatin1 (B64.encode $ SMP.unMsgId smId) == srvMsgId && mId == msgId
       _ -> False
     ackMessage bob aliceId msgId Nothing
     liftIO $ threadDelay 200000
@@ -4736,9 +5052,9 @@ testServerQueueInfo = do
     get bob =##> \case ("", c, Msg' mId PQEncOn "hello 4") -> c == aliceId && mId == msgId4; _ -> False
     ackMessage bob aliceId msgId4 Nothing
     liftIO $ threadDelay 200000
-    Just _ <- checkMsgQ bob aliceId 1 -- the one that did not fit now accepted
     get alice ##> ("", bobId, QCONT)
     get alice ##> ("", bobId, SENT msgId5)
+    Just _ <- checkMsgQ bob aliceId 1 -- the one that did not fit now accepted
     liftIO $ threadDelay 200000
     Just _srvMsgId <- checkQ bob aliceId True (Just QNoSub) 1 (Just MTMessage)
     get bob =##> \case ("", c, Msg' mId PQEncOn "hello: quota exceeded") -> c == aliceId && mId == msgId5 + 1; _ -> False

tests/AgentTests/NotificationTests.hs

@@ -230,7 +230,7 @@ testNotificationToken apns = do
     APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData}} <-
       getMockNotification apns tkn
     verification <- ntfData .-> "verification"
-    nonce <- C.cbNonce <$> ntfData .-> "nonce"
+    nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
     verifyNtfToken a tkn nonce verification
     NTActive <- checkNtfToken a tkn
     deleteNtfToken a tkn
@@ -254,7 +254,7 @@ testNtfTokenRepeatRegistration apns = do
     APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData}} <-
       getMockNotification apns tkn
     verification <- ntfData .-> "verification"
-    nonce <- C.cbNonce <$> ntfData .-> "nonce"
+    nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
     NTRegistered <- registerNtfToken a tkn NMPeriodic
     APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData'}} <-
       getMockNotification apns tkn
@@ -273,14 +273,14 @@ testNtfTokenSecondRegistration apns =
     APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData}} <-
       getMockNotification apns tkn
     verification <- ntfData .-> "verification"
-    nonce <- C.cbNonce <$> ntfData .-> "nonce"
+    nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
     verifyNtfToken a tkn nonce verification
 
     NTRegistered <- registerNtfToken a' tkn NMPeriodic
     APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData'}} <-
       getMockNotification apns tkn
     verification' <- ntfData' .-> "verification"
-    nonce' <- C.cbNonce <$> ntfData' .-> "nonce"
+    nonce' <- C.unsafeCbNonce <$> ntfData' .-> "nonce"
 
     -- at this point the first token is still active
     NTActive <- checkNtfToken a tkn
@@ -311,7 +311,7 @@ testNtfTokenServerRestart t apns = do
     -- so that repeat verification happens without restarting the clients, when notification arrives
     withNtfServer t $ runRight_ $ do
       verification <- ntfData .-> "verification"
-      nonce <- C.cbNonce <$> ntfData .-> "nonce"
+      nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
       verifyNtfToken a' tkn nonce verification
       NTActive <- checkNtfToken a' tkn
       pure ()
@@ -327,7 +327,7 @@ testNtfTokenServerRestartReverify t apns = do
       pure ntfData
     runRight_ $ do
       verification <- ntfData .-> "verification"
-      nonce <- C.cbNonce <$> ntfData .-> "nonce"
+      nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
       Left (BROKER _ (NETWORK _)) <- tryE $ verifyNtfToken a tkn nonce verification
       pure ()
   threadDelay 1500000
@@ -348,7 +348,7 @@ testNtfTokenServerRestartReverifyTimeout t apns = do
       APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData}} <-
         getMockNotification apns tkn
       verification <- ntfData .-> "verification"
-      nonce <- C.cbNonce <$> ntfData .-> "nonce"
+      nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
       verifyNtfToken a tkn nonce verification
       pure (nonce, verification)
     -- this emulates the situation when server verified token but the client did not receive the response
@@ -393,7 +393,7 @@ testNtfTokenServerRestartReregister t apns = do
       APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData}} <-
         getMockNotification apns tkn
       verification <- ntfData .-> "verification"
-      nonce <- C.cbNonce <$> ntfData .-> "nonce"
+      nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
       verifyNtfToken a' tkn nonce verification
       NTActive <- checkNtfToken a' tkn
       pure ()
@@ -428,7 +428,7 @@ testNtfTokenServerRestartReregisterTimeout t apns = do
       APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData}} <-
         getMockNotification apns tkn
       verification <- ntfData .-> "verification"
-      nonce <- C.cbNonce <$> ntfData .-> "nonce"
+      nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
       verifyNtfToken a' tkn nonce verification
       NTActive <- checkNtfToken a' tkn
       pure ()
@@ -450,7 +450,7 @@ testNtfTokenMultipleServers t apns = do
         APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData}} <-
           getMockNotification apns tkn
         verification <- ntfData .-> "verification"
-        nonce <- C.cbNonce <$> ntfData .-> "nonce"
+        nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
         verifyNtfToken a tkn nonce verification
         NTActive <- checkNtfToken a tkn
         -- shut down the "other" server
@@ -566,7 +566,7 @@ testNotificationSubscriptionExistingConnection apns baseId alice@AgentClient {ag
     APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData}} <-
       getMockNotification apns tkn
     verification <- ntfData .-> "verification"
-    vNonce <- C.cbNonce <$> ntfData .-> "nonce"
+    vNonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
     verifyNtfToken alice tkn vNonce verification
     NTActive <- checkNtfToken alice tkn
     -- send message
@@ -657,7 +657,7 @@ registerTestToken a token mode apns = do
   Just APNSMockRequest {notification = APNSNotification {aps = APNSBackground _, notificationData = Just ntfData'}} <-
     timeout 1000000 $ getMockNotification apns tkn
   verification' <- ntfData' .-> "verification"
-  nonce' <- C.cbNonce <$> ntfData' .-> "nonce"
+  nonce' <- C.unsafeCbNonce <$> ntfData' .-> "nonce"
   verifyNtfToken a tkn nonce' verification'
   NTActive <- checkNtfToken a tkn
   pure tkn
@@ -1035,7 +1035,7 @@ messageNotification apns tkn = do
   500000 `timeout` getMockNotification apns tkn >>= \case
     Nothing -> error "no notification"
     Just APNSMockRequest {notification = APNSNotification {aps = APNSMutableContent {}, notificationData = Just ntfData}} -> do
-      nonce <- C.cbNonce <$> ntfData .-> "nonce"
+      nonce <- C.unsafeCbNonce <$> ntfData .-> "nonce"
       message <- ntfData .-> "message"
       pure (nonce, message)
     _ -> error "bad notification"

tests/AgentTests/ResolveNameTests.hs

@@ -14,7 +14,6 @@ module AgentTests.ResolveNameTests (resolveNameTests) where
 
 import AgentTests.FunctionalAPITests (withAgent)
 import Control.Monad.Except (runExceptT)
-import qualified Data.Aeson as J
 import qualified Data.ByteString.Lazy as LB
 import Data.List (isInfixOf)
 import Network.HTTP.Types (Status, status200, status404, status502)
@@ -22,7 +21,7 @@ import NamesResolverServer (memCfg, memCfg2, memProxyCfg, withNames)
 import qualified NamesResolverServer as NRS
 import SMPAgentClient
 import SMPClient
-import SMPNamesTests (testNameRecord)
+import SMPNamesTests (availableBody, registeredBody, testNameRecord)
 import Simplex.Messaging.Agent (resolveSimplexName)
 import Simplex.Messaging.Agent.Client (AgentClient)
 import Simplex.Messaging.Agent.Env.SQLite (InitialAgentServers (..), ServerCfg, ServerRoles (..), presetServerCfg)
@@ -71,13 +70,13 @@ withNoNameServers k = withAgent 1 agentCfg (oneSrv (proxySrvCfg testSMPServer))
 resolveNameTests :: Spec
 resolveNameTests = do
   describe "direct path (SPMNever)" $
-    it "404 propagates as SMP host (NAME NOT_FOUND)" testDirectNotFound
+    it "a resolver error propagates as SMP host (NAME RESOLVER)" testDirectResolverErr
   describe "proxy path (SPMAlways)" $
-    it "404 from resolver propagates via proxy as SMP <proxyHost> (NAME NOT_FOUND)" testProxyNotFound
+    it "a resolver error propagates via proxy as SMP <proxyHost> (NAME RESOLVER)" testProxyResolverErr
   describe "TLDTesting path" $
-    it "NAME NOT_FOUND for TLDTesting too" testTestingTldNotFound
+    it "NAME RESOLVER for TLDTesting too" testTestingTldResolverErr
   describe "TLDWeb path" $
-    it "NAME NOT_FOUND for TLDWeb too" testWebTldNotFound
+    it "NAME RESOLVER for TLDWeb too" testWebTldResolverErr
   describe "no resolver configured" $
     it "answers NAME NO_RESOLVER" testNoResolver
   describe "no names servers (names role off everywhere)" $
@@ -86,38 +85,50 @@ resolveNameTests = do
     it "surfaces as SMP host (NAME (RESOLVER ..))" testBackendError
   describe "success path" $
     it "returns NameRecord" testDirectSuccess
+  describe "name availability" $
+    it "an unregistered name answers as available" testAvailSuccess
 
-testDirectNotFound :: HasCallStack => IO ()
-testDirectNotFound =
+testAvailSuccess :: HasCallStack => IO ()
+testAvailSuccess =
+  withDirectResolver (status200, availableBody) $ \c -> do
+    r <- runExceptT $ resolveSimplexName c NRMInteractive 1 (SimplexDomain TLDSimplex "alice" [])
+    case r of
+      Right SMP.NameResponse {registration = SMP.NRAvailable {}} -> pure ()
+      _ -> expectationFailure $ "expected Right NRAvailable, got: " <> show r
+
+-- | 404 is a resolver that predates /v2/resolve: no status from that endpoint
+-- means "not registered", since an unregistered name answers NRAvailable.
+testDirectResolverErr :: HasCallStack => IO ()
+testDirectResolverErr =
   withDirectResolver (status404, "{}") $ \c -> do
     r <- runExceptT $ resolveSimplexName c NRMInteractive 1 (SimplexDomain TLDSimplex "alice" [])
     case r of
-      Left (SMP _ (SMP.NAME SMP.NOT_FOUND)) -> pure ()
-      _ -> expectationFailure $ "expected Left (SMP _ (NAME NOT_FOUND)), got: " <> show r
+      Left (SMP _ (SMP.NAME (SMP.RESOLVER _))) -> pure ()
+      _ -> expectationFailure $ "expected Left (SMP _ (NAME (RESOLVER _))), got: " <> show r
 
-testProxyNotFound :: HasCallStack => IO ()
-testProxyNotFound =
+testProxyResolverErr :: HasCallStack => IO ()
+testProxyResolverErr =
   withProxyAndResolver (status404, "{}") $ \c -> do
     r <- runExceptT $ resolveSimplexName c NRMInteractive 1 (SimplexDomain TLDSimplex "alice" [])
     case r of
-      Left (SMP host (SMP.NAME SMP.NOT_FOUND)) | testPort `isInfixOf` host -> pure ()
-      _ -> expectationFailure $ "expected Left (SMP <proxyHost:" <> testPort <> "> (NAME NOT_FOUND)), got: " <> show r
+      Left (SMP host (SMP.NAME (SMP.RESOLVER _))) | testPort `isInfixOf` host -> pure ()
+      _ -> expectationFailure $ "expected Left (SMP <proxyHost:" <> testPort <> "> (NAME (RESOLVER _))), got: " <> show r
 
-testTestingTldNotFound :: HasCallStack => IO ()
-testTestingTldNotFound =
+testTestingTldResolverErr :: HasCallStack => IO ()
+testTestingTldResolverErr =
   withDirectResolver (status404, "{}") $ \c -> do
     r <- runExceptT $ resolveSimplexName c NRMInteractive 1 (SimplexDomain TLDTesting "bob" [])
     case r of
-      Left (SMP _ (SMP.NAME SMP.NOT_FOUND)) -> pure ()
-      _ -> expectationFailure $ "expected Left (SMP _ (NAME NOT_FOUND)), got: " <> show r
+      Left (SMP _ (SMP.NAME (SMP.RESOLVER _))) -> pure ()
+      _ -> expectationFailure $ "expected Left (SMP _ (NAME (RESOLVER _))), got: " <> show r
 
-testWebTldNotFound :: HasCallStack => IO ()
-testWebTldNotFound =
+testWebTldResolverErr :: HasCallStack => IO ()
+testWebTldResolverErr =
   withDirectResolver (status404, "{}") $ \c -> do
     r <- runExceptT $ resolveSimplexName c NRMInteractive 1 (SimplexDomain TLDWeb "example.com" [])
     case r of
-      Left (SMP _ (SMP.NAME SMP.NOT_FOUND)) -> pure ()
-      _ -> expectationFailure $ "expected Left (SMP _ (NAME NOT_FOUND)), got: " <> show r
+      Left (SMP _ (SMP.NAME (SMP.RESOLVER _))) -> pure ()
+      _ -> expectationFailure $ "expected Left (SMP _ (NAME (RESOLVER _))), got: " <> show r
 
 testNoResolver :: HasCallStack => IO ()
 testNoResolver =
@@ -145,8 +156,8 @@ testBackendError =
 
 testDirectSuccess :: HasCallStack => IO ()
 testDirectSuccess =
-  withDirectResolver (status200, J.encode testNameRecord) $ \c -> do
+  withDirectResolver (status200, registeredBody testNameRecord) $ \c -> do
     r <- runExceptT $ resolveSimplexName c NRMInteractive 1 (SimplexDomain TLDSimplex "alice" [])
     case r of
-      Right nr -> nr `shouldBe` testNameRecord
-      _ -> expectationFailure $ "expected Right NameRecord, got: " <> show r
+      Right SMP.NameResponse {registration = SMP.NRRegistered {nameRecord}} -> nameRecord `shouldBe` testNameRecord
+      _ -> expectationFailure $ "expected Right NRRegistered, got: " <> show r

tests/AgentTests/SQLiteTests.hs

@@ -20,12 +20,13 @@ import Control.Concurrent.Async (concurrently_)
 import Control.Concurrent.MVar
 import Control.Concurrent.STM
 import Control.Exception (SomeException)
-import Control.Monad (replicateM_)
+import Control.Monad (forM_, replicateM_)
 import Control.Monad.Trans.Except
 import Crypto.Random (ChaChaDRG)
 import Data.ByteArray (ScrubbedBytes)
 import Data.ByteString.Char8 (ByteString)
 import Data.List (isInfixOf)
+import qualified Data.Map.Strict as M
 import qualified Data.Text as T
 import Data.Text.Encoding (encodeUtf8)
 import Data.Time
@@ -51,6 +52,7 @@ import qualified Simplex.Messaging.Crypto as C
 import Simplex.Messaging.Crypto.File (CryptoFile (..))
 import Simplex.Messaging.Crypto.Ratchet (pattern IKPQOn)
 import qualified Simplex.Messaging.Crypto.Ratchet as CR
+import Simplex.Messaging.Encoding (Encoding (..))
 import Simplex.Messaging.Encoding.String (StrEncoding (..))
 import Simplex.Messaging.Protocol (EntityId (..), QueueMode (..), SubscriptionMode (..), pattern VersionSMPC)
 import qualified Simplex.Messaging.Protocol as SMP
@@ -135,6 +137,8 @@ storeTests = do
           testCreateRcvMsg
           testCreateSndMsg
           testCreateRcvAndSndMsgs
+      describe "deleteRatchetKeyHashesExpired" testDeleteRatchetKeyHashesExpired
+      it "should keep only the newest skipped message keys" testGetSkippedMsgKeys
       describe "Work items" $ do
         it "should getPendingQueueMsg" testGetPendingQueueMsg
         it "should getPendingServerCommand" testGetPendingServerCommand
@@ -509,7 +513,7 @@ hw = encodeUtf8 "Hello world!"
 ts :: UTCTime
 ts = UTCTime (fromGregorian 2021 02 24) (secondsToDiffTime 0)
 
-mkRcvMsgData :: InternalId -> InternalRcvId -> ExternalSndId -> BrokerId -> MsgHash -> RcvMsgData
+mkRcvMsgData :: InternalId -> InternalRcvId -> ExternalSndId -> ByteString -> MsgHash -> RcvMsgData
 mkRcvMsgData internalId internalRcvId externalSndId brokerId internalHash =
   RcvMsgData
     { internalRcvId,
@@ -518,7 +522,7 @@ mkRcvMsgData internalId internalRcvId externalSndId brokerId internalHash =
           { integrity = MsgOk,
             recipient = (unId internalId, ts),
             sndMsgId = externalSndId,
-            broker = (brokerId, ts),
+            broker = (SMP.unsafeMsgId $ mconcat (replicate 24 brokerId), ts),
             pqEncryption = CR.PQEncOn
           },
       msgType = AM_A_MSG_,
@@ -599,6 +603,42 @@ testCreateRcvAndSndMsgs =
       testCreateSndMsg_ db "snd_hash_1" connId sq $ mkSndMsgData (InternalId 5) (InternalSndId 2) "snd_hash_2"
       testCreateSndMsg_ db "snd_hash_2" connId sq $ mkSndMsgData (InternalId 6) (InternalSndId 3) "snd_hash_3"
 
+testDeleteRatchetKeyHashesExpired :: SpecWith DBStore
+testDeleteRatchetKeyHashesExpired =
+  it "should delete expired ratchet key hashes except the newest in each connection" . withStoreTransaction $ \db -> do
+    g <- C.newRandom
+    Right connId <- createNewConn db g cData1 {connId = ""} SCMInvitation
+    Right connId' <- createNewConn db g cData1 {connId = ""} SCMContact
+    let hashes = ["h1", "h2", "h3", "h4", "h5", "h6"]
+    forM_ hashes $ addProcessedRatchetKeyHash db connId
+    forM_ (take 4 hashes) $ addProcessedRatchetKeyHash db connId'
+    deleteRatchetKeyHashesExpired db 86400 4
+    mapM (checkRatchetKeyHashExists db connId) hashes `shouldReturn` replicate 6 True
+    deleteRatchetKeyHashesExpired db 0 4
+    mapM (checkRatchetKeyHashExists db connId) hashes `shouldReturn` [False, False, True, True, True, True]
+    mapM (checkRatchetKeyHashExists db connId') (take 4 hashes) `shouldReturn` replicate 4 True
+
+testGetSkippedMsgKeys :: DBStore -> Expectation
+testGetSkippedMsgKeys st = do
+  g <- C.newRandom
+  withTransaction st $ \db -> do
+    Right connId <- createNewConn db g cData1 {connId = ""} SCMInvitation
+    Right connId' <- createNewConn db g cData1 {connId = ""} SCMInvitation
+    createSkippedKeys db connId'
+    createSkippedKeys db connId
+    M.map M.keys <$> getSkippedMsgKeys db connId 4
+      `shouldReturn` M.singleton (C.Key "header_key") [1, 2, 3, 4]
+    getMsgNs db connId `shouldReturn` [1 .. 4]
+    getMsgNs db connId' `shouldReturn` [1 .. 10]
+  where
+    createSkippedKeys :: DB.Connection -> ConnId -> IO ()
+    createSkippedKeys db connId = do
+      DB.execute db "INSERT INTO ratchets (conn_id) VALUES (?)" (Only connId)
+      forM_ ([10, 9 .. 1] :: [Int]) $ \msgN ->
+        DB.execute db "INSERT INTO skipped_messages (conn_id, header_key, msg_n, msg_key) VALUES (?, ?, ?, ?)" (connId, "header_key" :: ByteString, msgN, smpEncode ("key" :: ByteString, "iv" :: ByteString))
+    getMsgNs :: DB.Connection -> ConnId -> IO [Int]
+    getMsgNs db connId = map fromOnly <$> DB.query db "SELECT msg_n FROM skipped_messages WHERE conn_id = ? ORDER BY msg_n" (Only connId)
+
 testCloseReopenStore :: IO ()
 testCloseReopenStore = do
   st <- createStore'
@@ -648,7 +688,7 @@ hasMigrations :: DBStore -> Expectation
 hasMigrations st = getMigrations st `shouldReturn` True
 
 errorGettingMigrations :: DBStore -> Expectation
-errorGettingMigrations st = getMigrations st `shouldThrow` \(e :: SomeException) -> "ErrorMisuse" `isInfixOf` show e
+errorGettingMigrations st = getMigrations st `shouldThrow` \(e :: SomeException) -> "withTransaction: database closed" `isInfixOf` show e
 
 testGetPendingQueueMsg :: DBStore -> Expectation
 testGetPendingQueueMsg st = do
@@ -743,9 +783,11 @@ testGetNextRcvChunkToDownload st = do
   withTransaction st $ \db -> do
     Right Nothing <- getNextRcvChunkToDownload db xftpServer1 86400
 
-    Right _ <- createRcvFile db g 1 rcvFileDescr1 "filepath" "filepath" (CryptoFile "filepath" Nothing) True
+    Right fId1 <- createRcvFile db g 1 rcvFileDescr1 "filepath" "filepath" (CryptoFile "filepath" Nothing) True
+    Right _ <- startPreparedRcvFile db fId1
     DB.execute_ db "UPDATE rcv_file_chunk_replicas SET replica_key = cast('bad' as blob) WHERE rcv_file_chunk_replica_id = 1"
     Right fId2 <- createRcvFile db g 1 rcvFileDescr1 "filepath" "filepath" (CryptoFile "filepath" Nothing) True
+    Right _ <- startPreparedRcvFile db fId2
 
     Left e <- getNextRcvChunkToDownload db xftpServer1 86400
     show e `shouldContain` "ConversionFailed"
@@ -783,6 +825,7 @@ testGetNextSndFileToPrepare st = do
     -- Right _ <- createSndFile db g 1 (CryptoFile "filepath" Nothing) 1 "filepath" testFileSbKey testFileCbNonce Nothing
     -- DB.execute_ db "UPDATE snd_files SET status = 'new', num_recipients = 'bad' WHERE snd_file_id = 1"
     Right fId2 <- createSndFile db g 1 (CryptoFile "filepath" Nothing) 1 "filepath" testFileSbKey testFileCbNonce Nothing Nothing
+    Right _ <- startPreparedSndFile db fId2
     DB.execute_ db "UPDATE snd_files SET status = 'new' WHERE snd_file_id = 2"
 
     -- Left e <- getNextSndFileToPrepare db 86400

tests/AgentTests/SchemaDump.hs

@@ -48,6 +48,8 @@ schemaDumpTest = do
   it "verify strict tables" testVerifyStrict
   it "should NOT create user record for new database" testUsersMigrationNew
   it "should create user record for old database" testUsersMigrationOld
+  it "should remove duplicate ratchet key hashes before adding unique index" testRatchetKeyHashesUniqueMigration
+  it "should clear received_qend switch status on downgrade" testRcvSwitchStatusDowngrade
 
 testVerifySchemaDump :: IO ()
 testVerifySchemaDump = do
@@ -114,12 +116,45 @@ testUsersMigrationOld = do
     `shouldReturn` ([Only (1 :: Int)])
   closeDBStore st'
 
+testRatchetKeyHashesUniqueMigration :: IO ()
+testRatchetKeyHashesUniqueMigration = do
+  let beforeUnique = takeWhile (("m20260929_ratchet_indexes" /=) . name) appMigrations
+  Right st <- createDBStore (DBOpts testDB [] "" False True TQOff) beforeUnique (MigrationConfig MCError Nothing)
+  withTransaction' st $ \db -> do
+    SQL.execute_ db "INSERT INTO users (user_id) VALUES (1)"
+    SQL.execute_ db "INSERT INTO connections (conn_id, conn_mode, user_id) VALUES (x'01', 'INV', 1), (x'02', 'INV', 1)"
+    SQL.execute_ db "INSERT INTO processed_ratchet_key_hashes (conn_id, hash) VALUES (x'01', x'aa'), (x'01', x'aa'), (x'01', x'bb'), (x'02', x'aa'), (x'01', x'aa')"
+  closeDBStore st
+  Right st' <- createDBStore (DBOpts testDB [] "" False True TQOff) appMigrations (MigrationConfig MCYesUp Nothing)
+  withTransaction' st' (`SQL.query_` "SELECT processed_ratchet_key_hash_id FROM processed_ratchet_key_hashes ORDER BY processed_ratchet_key_hash_id")
+    `shouldReturn` [Only (1 :: Int), Only 3, Only 4]
+  closeDBStore st'
+
+testRcvSwitchStatusDowngrade :: IO ()
+testRcvSwitchStatusDowngrade = do
+  Right st <- createDBStore (DBOpts testDB [] "" False True TQOff) appMigrations (MigrationConfig MCError Nothing)
+  withTransaction' st $ \db -> do
+    SQL.execute_ db "INSERT INTO users (user_id) VALUES (1)"
+    SQL.execute_ db "INSERT INTO connections (conn_id, conn_mode, user_id) VALUES (x'01', 'INV', 1)"
+    SQL.execute_ db "INSERT INTO servers (host, port, key_hash) VALUES ('localhost', '5001', x'00')"
+    SQL.execute_ db "INSERT INTO rcv_queues (host, port, rcv_id, conn_id, rcv_private_key, rcv_dh_secret, e2e_priv_key, snd_id, status, rcv_queue_id, rcv_primary, switch_status) VALUES ('localhost', '5001', x'01', x'01', x'00', x'00', x'00', x'01', 'active', 1, 0, 'received_qend'), ('localhost', '5001', x'02', x'01', x'00', x'00', x'00', x'02', 'active', 2, 1, 'sending_qadd')"
+  closeDBStore st
+  let beforeRcvSwitchStatus = takeWhile (("m20261008_rcv_switch_status" /=) . name) appMigrations
+  Right st' <- createDBStore (DBOpts testDB [] "" False True TQOff) beforeRcvSwitchStatus (MigrationConfig MCYesUpDown Nothing)
+  withTransaction' st' (`SQL.query_` "SELECT switch_status FROM rcv_queues ORDER BY rcv_queue_id")
+    `shouldReturn` [Only (Nothing :: Maybe Text), Only (Just "sending_qadd")]
+  closeDBStore st'
+
 skipComparisonForDownMigrations :: [String]
 skipComparisonForDownMigrations =
   [ -- on down migration idx_messages_internal_snd_id_ts index moves down to the end of the file
     "m20230814_indexes",
     -- snd_secure and last_broker_ts columns swap order on down migration
-    "m20250322_short_links"
+    "m20250322_short_links",
+    -- on down migration idx_processed_ratchet_key_hashes_hash index moves down to the end of the file
+    "m20260929_ratchet_indexes",
+    -- sqlite_sequence moves after commands table after VACUUM
+    "m20261005_commands_autoincrement"
   ]
 
 getSchema :: FilePath -> FilePath -> IO String

tests/AgentTests/ServerChoice.hs

@@ -71,9 +71,7 @@ initServers =
     }
 
 testChooseDifferentOperator :: IO ()
-testChooseDifferentOperator = do
-  c <- getSMPAgentClient' 1 agentCfg initServers testDB
-  runRight_ $ do
+testChooseDifferentOperator = withAgent 1 agentCfg initServers testDB $ \c -> runRight_ $ do
     -- chooses the only operator with storage role
     srv1 <- withAgentEnv c $ getNextServer c 1 storageSrvs []
     liftIO $ srv1 == testOp1Srv1 || srv1 == testOp1Srv2 `shouldBe` True

tests/CoreTests/BatchingTests.hs

@@ -14,11 +14,13 @@ import Control.Monad
 import Crypto.Random (ChaChaDRG)
 import qualified Data.ByteString as B
 import Data.ByteString.Char8 (ByteString)
+import Data.ByteString.Unsafe (unsafeUseAsCString)
 import qualified Data.List.NonEmpty as L
 import Data.Time.Clock.System (SystemTime, getSystemTime)
 import qualified Data.X509 as X
 import qualified Data.X509.CertificateStore as XS
 import qualified Data.X509.File as XF
+import Foreign.Ptr (plusPtr)
 import Simplex.Messaging.Client
 import qualified Simplex.Messaging.Crypto as C
 import Simplex.Messaging.Encoding
@@ -40,6 +42,8 @@ batchingTests = do
     it "should batch subscription responses with message" testBatchSubResponses
     it "should break on message that does not fit" testClientBatchWithMessage
     it "should break on large message" testClientBatchWithLargeMessage
+  describe "tDecodeServer" $
+    it "should copy IDs out of received block" testDecodeServerCopiesIds
 
 testBatchSubscriptions :: IO ()
 testBatchSubscriptions = do
@@ -166,6 +170,26 @@ testClientBatchWithLargeMessage = do
   (length rs1', length rs2') `shouldBe` (75, 135)
   all lenOk [s1', s2'] `shouldBe` True
 
+testDecodeServerCopiesIds :: IO ()
+testDecodeServerCopiesIds = do
+  sessId <- atomically . C.randomBytes 32 =<< C.newRandom
+  subs <- replicateM 2 $ randomSUB sessId
+  let thParams = testTHandleParams sessId
+  [TBTransmissions s 2 _] <- pure $ batchTransmissions thParams $ L.fromList subs
+  forM_ (tParse thParams s) $ \t -> do
+    let (corrId, entId, _) = tDecodeClient @SMPVersion @ErrorType @Cmd thParams t
+    Right (_, _, (corrId', entId', _)) <- pure $ tDecodeServer @SMPVersion @ErrorType @Cmd (testTHandleParams sessId) t
+    (corrId', entId') `shouldBe` (corrId, entId)
+    sharesBuffer s (bs corrId) `shouldReturn` True
+    sharesBuffer s (unEntityId entId) `shouldReturn` True
+    sharesBuffer s (bs corrId') `shouldReturn` False
+    sharesBuffer s (unEntityId entId') `shouldReturn` False
+
+sharesBuffer :: ByteString -> ByteString -> IO Bool
+sharesBuffer block s =
+  unsafeUseAsCString block $ \blockPtr -> unsafeUseAsCString s $ \ptr ->
+    pure $ ptr >= blockPtr && ptr < blockPtr `plusPtr` B.length block
+
 testClientStub :: IO (ProtocolClient SMPVersion ErrorType BrokerMsg)
 testClientStub = do
   g <- C.newRandom
@@ -183,7 +207,7 @@ randomSUB sessId = do
   thAuth_ <- testTHandleAuth g rKey
   let thParams = testTHandleParams sessId
       TransmissionForAuth {tForAuth, tToSend} = encodeTransmissionForAuth thParams (CorrId corrId, EntityId rId, Cmd SRecipient SUB)
-  pure $ (,tToSend) <$> authTransmission thAuth_ True (Just rpKey) nonce tForAuth
+  pure $ (,tToSend) <$> authTransmission thAuth_ True (Just rpKey) (Just nonce) tForAuth
 
 randomSUBCmd :: ProtocolClient SMPVersion ErrorType BrokerMsg -> IO (PCTransmission ErrorType BrokerMsg)
 randomSUBCmd c = do
@@ -202,7 +226,7 @@ randomNMSGCmd :: SystemTime -> IO (Transmission BrokerMsg)
 randomNMSGCmd ts = do
   g <- C.newRandom
   nId <- atomically $ C.randomBytes 24 g
-  msgId <- atomically $ C.randomBytes 24 g
+  msgId <- atomically $ randomMsgId 24 g
   (k, pk) <- atomically $ C.generateKeyPair g
   nonce <- atomically $ C.randomCbNonce g
   let msgMeta = NMsgMeta {msgId, msgTs = ts}
@@ -221,7 +245,7 @@ randomMSG = do
   g <- C.newRandom
   corrId <- atomically $ C.randomBytes 24 g
   rId <- atomically $ C.randomBytes 24 g
-  msgId <- atomically $ C.randomBytes 24 g
+  msgId <- atomically $ randomMsgId 24 g
   msg <- atomically $ C.randomBytes maxMessageLength g
   pure (CorrId corrId, EntityId rId, MSG RcvMessage {msgId, msgBody = EncRcvMsgBody msg})
 
@@ -235,9 +259,9 @@ randomSEND sessId len = do
   msg <- atomically $ C.randomBytes len g
   let thParams = testTHandleParams sessId
       TransmissionForAuth {tForAuth, tToSend} = encodeTransmissionForAuth thParams (CorrId corrId, EntityId sId, Cmd SSender $ SEND noMsgFlags msg)
-  pure $ (,tToSend) <$> authTransmission thAuth_ False (Just spKey) nonce tForAuth
+  pure $ (,tToSend) <$> authTransmission thAuth_ False (Just spKey) (Just nonce) tForAuth
 
-testTHandleParams :: ByteString -> THandleParams SMPVersion 'TClient
+testTHandleParams :: ByteString -> THandleParams SMPVersion p
 testTHandleParams sessionId =
   THandleParams
     { sessionId,

tests/CoreTests/CryptoTests.hs

@@ -6,8 +6,9 @@
 
 module CoreTests.CryptoTests (cryptoTests) where
 
+import Control.Concurrent (forkIO, newEmptyMVar, putMVar, takeMVar)
 import Control.Concurrent.STM
-import Control.Exception (evaluate)
+import Control.Exception (bracket, evaluate)
 import Control.Monad.Except
 import qualified Data.Aeson as J
 import qualified Data.ByteString.Char8 as B
@@ -22,9 +23,12 @@ import Data.Time.Clock (UTCTime (..))
 import qualified Data.Text.Lazy as LT
 import qualified Data.Text.Lazy.Encoding as LE
 import Data.Type.Equality
+import Data.Word (Word8)
 import qualified Data.X509 as X
 import qualified Data.X509.CertificateStore as XS
 import qualified Data.X509.Validation as XV
+import Foreign (FunPtr, allocaBytes, fillBytes, freeHaskellFunPtr, nullPtr)
+import Foreign.C.Types (CInt, CSize (..))
 import qualified SMPClient
 import qualified Simplex.Messaging.Crypto as C
 import qualified Simplex.Messaging.Crypto.Lazy as LC
@@ -32,9 +36,12 @@ import Simplex.Messaging.Crypto.BBS
 import Simplex.Messaging.Crypto.Entitlement
 import Simplex.Messaging.Crypto.SNTRUP761.Bindings
 import Simplex.Messaging.Crypto.SNTRUP761.Bindings.Defines
+import Simplex.Messaging.Crypto.SNTRUP761.Bindings.FFI (c_sntrup761_keypair)
+import Simplex.Messaging.Crypto.SNTRUP761.Bindings.RNG (RNGFunc)
 import Simplex.Messaging.Encoding (Large (..), smpDecode, smpEncode)
 import Simplex.Messaging.Encoding.String (strDecode, strEncode)
 import Simplex.Messaging.Transport.Client
+import System.Timeout (timeout)
 import Test.Hspec hiding (fit, it)
 import Test.Hspec.QuickCheck (modifyMaxSuccess)
 import Test.QuickCheck hiding (Large)
@@ -113,6 +120,7 @@ cryptoTests = do
   describe "sntrup761" $ do
     it "should enc/dec key" testSNTRUP761
     it "should reject malformed KEM encodings" testSNTRUP761RejectsMalformedEncodings
+    it "should fail key generation with degenerate RNG" testSNTRUP761KeypairDegenerateRNG
   describe "BBS+" $ do
     it "should sign and verify" testBBSSignVerify
     it "should derive public key from secret key" testBBSPublicKeyDerivation
@@ -298,6 +306,26 @@ testSNTRUP761 = do
   KEMSharedKey k' <- sntrup761Dec c sk
   k' `shouldBe` k
 
+foreign import ccall "wrapper"
+  mkRNGFunc :: RNGFunc -> IO (FunPtr RNGFunc)
+
+testSNTRUP761KeypairDegenerateRNG :: IO ()
+testSNTRUP761KeypairDegenerateRNG = do
+  -- constant byte 0 draws invertible g = -(1 + x + ... + x^760), byte 0x20 draws g = 0
+  keypairWithConstantRNG 0 `shouldReturn` Just 0
+  keypairWithConstantRNG 0x20 `shouldReturn` Just (-1)
+  where
+    keypairWithConstantRNG :: Word8 -> IO (Maybe CInt)
+    keypairWithConstantRNG b = do
+      result <- newEmptyMVar
+      -- timeout cannot interrupt a foreign call, so the call runs in another thread
+      _ <- forkIO $
+        bracket (mkRNGFunc $ \_ sz buf -> fillBytes buf b (fromIntegral sz)) freeHaskellFunPtr $ \rng ->
+          allocaBytes c_SNTRUP761_PUBLICKEY_SIZE $ \pkPtr ->
+            allocaBytes c_SNTRUP761_SECRETKEY_SIZE $ \skPtr ->
+              c_sntrup761_keypair pkPtr skPtr nullPtr rng >>= putMVar result
+      timeout 10000000 $ takeMVar result
+
 testSNTRUP761RejectsMalformedEncodings :: IO ()
 testSNTRUP761RejectsMalformedEncodings = do
   smpDecode @KEMPublicKey (smpEncode $ Large shortPublicKey) `shouldSatisfy` isLeft

tests/CoreTests/MsgStoreTests.hs

@@ -34,7 +34,7 @@ import Data.Time.Clock.System (SystemTime (..), getSystemTime)
 import SMPClient (testStoreLogFile, testStoreMsgsDir, testStoreMsgsDir2, testStoreMsgsFile, testStoreMsgsFile2)
 import Simplex.Messaging.Crypto (pattern MaxLenBS)
 import qualified Simplex.Messaging.Crypto as C
-import Simplex.Messaging.Protocol (EntityId (..), ErrorType, LinkId, Message (..), QueueLinkData, RecipientId, SParty (..), noMsgFlags)
+import Simplex.Messaging.Protocol (EncDataBytes (..), EntityId (..), ErrorType (..), LinkId, Message (..), QueueLinkData, RecipientId, SParty (..), noMsgFlags, randomMsgId)
 import Simplex.Messaging.Server (exportMessages, importMessages, printMessageStats)
 import Simplex.Messaging.Server.Env.STM (MsgStore (..), journalMsgStoreDepth, readWriteQueueStore)
 import Simplex.Messaging.Server.Expiration (ExpirationConfig (..), expireBeforeEpoch)
@@ -43,6 +43,7 @@ import Simplex.Messaging.Server.MsgStore.STM
 import Simplex.Messaging.Server.MsgStore.Types
 import Simplex.Messaging.Server.QueueStore
 import Simplex.Messaging.Server.QueueStore.QueueInfo
+import Simplex.Messaging.Server.QueueStore.Types
 import Simplex.Messaging.Server.StoreLog (closeStoreLog, logCreateQueue)
 import System.Directory (copyFile, createDirectoryIfMissing, listDirectory, removeFile, renameFile)
 import System.FilePath ((</>))
@@ -57,7 +58,6 @@ import Simplex.Messaging.Agent.Store.Postgres.Common
 import Simplex.Messaging.Agent.Store.Shared (MigrationConfirmation (..))
 import Simplex.Messaging.Server.MsgStore.Postgres
 import Simplex.Messaging.Server.QueueStore.Postgres
-import Simplex.Messaging.Server.QueueStore.Types
 import SMPClient (postgressBracket, testServerDBConnectInfo, testStoreDBOpts)
 #endif
 
@@ -79,6 +79,7 @@ msgStoreTests = do
         someMsgStoreTests
         it "should correctly update message counts and canWrite flag" testUpdateMessageCounts
         it "tryDelPeekMsg (ACK not from NSE) should reset message counts when queue is empty" testResetMessageCounts
+        it "should expire messages across commit batches" testExpireMessagesInBatches
 #endif
   describe "Journal message store: queue state backup expiration" $ do
     it "should remove old queue state backups" testRemoveQueueStateBackups
@@ -100,6 +101,7 @@ msgStoreTests = do
       it "should get queue and store/read messages" testGetQueue
       it "should write/ack messages" testWriteAckMessages
       it "should not fail on EOF when changing read journal" testChangeReadJournal
+      it "should not add link data to secured messaging queue" testLinkDataSecuredQueue
 
 -- TODO constrain to STM stores?
 withMsgStore :: MsgStoreClass s => MsgStoreConfig s -> (s -> IO ()) -> IO ()
@@ -145,7 +147,7 @@ mkMessage :: MonadIO m => ByteString -> m Message
 mkMessage body = liftIO $ do
   g <- C.newRandom
   msgTs <- getSystemTime
-  msgId <- atomically $ C.randomBytes 24 g
+  msgId <- atomically $ randomMsgId 24 g
   pure Message {msgId, msgTs, msgFlags = noMsgFlags, msgBody = C.unsafeMaxLenBS body}
 
 pattern Msg :: ByteString -> Maybe Message
@@ -266,6 +268,33 @@ testChangeReadJournal ms = do
     (Msg "message 5", Nothing) <- tryDelPeekMsg ms q mId5
     void $ ExceptT $ deleteQueue ms q
 
+testLinkDataSecuredQueue :: MsgStoreClass s => s -> IO ()
+testLinkDataSecuredQueue ms = do
+  g <- C.newRandom
+  (sKey, _) <- atomically $ C.generateAuthKeyPair C.SEd25519 g
+  let st = queueStore ms
+      ld = (EncDataBytes "fixed data", EncDataBytes "user data")
+      rndId = atomically $ EntityId <$> C.randomBytes 24 g
+  (rId, qr) <- testNewQueueRec g QMMessaging
+  (cId, cqr) <- testNewQueueRec g QMContact
+  lnkId <- rndId
+  cLnkId <- rndId
+  runRight_ $ do
+    q <- ExceptT $ addQueue ms rId qr
+    -- the handle is read before SKEY, as in a command that raced with it
+    staleQ <- ExceptT $ getQueue ms SRecipient rId
+    ExceptT $ secureQueue st q sKey
+    liftIO $ addQueueLinkData st staleQ lnkId ld `shouldReturn` Left AUTH
+    freshQ <- ExceptT $ getQueue ms SRecipient rId
+    liftIO $ getQueueLinkData st freshQ lnkId `shouldReturn` Left AUTH
+    cq <- ExceptT $ addQueue ms cId cqr
+    ExceptT $ secureQueue st cq sKey
+    ExceptT $ addQueueLinkData st cq cLnkId ld
+    ld' <- ExceptT $ getQueueLinkData st cq cLnkId
+    liftIO $ ld' `shouldBe` ld
+    void $ ExceptT $ deleteQueue ms q
+    void $ ExceptT $ deleteQueue ms cq
+
 testExportImportStore :: JournalMsgStore 'QSMemory -> IO ()
 testExportImportStore ms = do
   g <- C.newRandom
@@ -327,35 +356,34 @@ testUpdateMessageCounts ms = do
     q <- ExceptT $ addQueue ms rId qr
     let write s = writeMsg ms q True =<< mkMessage s
         hasSize = checkQueueSize ms
-    q `hasSize` (0, True, False)
+    q `hasSize` (0, True)
     Just (Message {msgId = mId1}, True) <- write "message 1"
-    q `hasSize` (1, True, True)
+    q `hasSize` (1, True)
     Just (Message {msgId = mId2}, False) <- write "message 2"
-    q `hasSize` (2, True, True)
+    q `hasSize` (2, True)
     Just (Message {msgId = mId3}, False) <- write "message 3"
-    q `hasSize` (3, True, True)
+    q `hasSize` (3, True)
     Nothing <- write "message 4"
-    q `hasSize` (4, False, True)
+    q `hasSize` (4, False)
     Msg "message 1" <- tryPeekMsg ms q
-    q `hasSize` (4, False, True)
+    q `hasSize` (4, False)
     Msg "message 1" <- tryDelMsg ms q mId1
-    q `hasSize` (3, False, True)
+    q `hasSize` (3, False)
     Msg "message 2" <- tryPeekMsg ms q
     (Msg "message 2", Msg "message 3") <- tryDelPeekMsg ms q mId2
-    q `hasSize` (2, False, True)
+    q `hasSize` (2, False)
     (Msg "message 3", Just MessageQuota {msgId = mId4}) <- tryDelPeekMsg ms q mId3
-    q `hasSize` (1, False, True)
+    q `hasSize` (1, False)
     (Just MessageQuota {}, Nothing) <- tryDelPeekMsg ms q mId4
-    q `hasSize` (0, True, False)
+    q `hasSize` (0, True)
 
-checkQueueSize :: PostgresMsgStore -> PostgresQueue -> (Int64, Bool, Bool) -> ExceptT ErrorType IO ()
-checkQueueSize ms q (size, canWrt, expire) = liftIO $ do
-  [(size', canWrt', expire')] <-
+checkQueueSize :: PostgresMsgStore -> PostgresQueue -> (Int64, Bool) -> ExceptT ErrorType IO ()
+checkQueueSize ms q (size, canWrt) = liftIO $ do
+  [(size', canWrt')] <-
     withTransaction (dbStore $ queueStore ms) $ \db ->
-      DB.query db "SELECT msg_queue_size, msg_can_write, msg_queue_expire FROM msg_queues WHERE recipient_id = ?" (Only (recipientId q))
+      DB.query db "SELECT msg_queue_size, msg_can_write FROM msg_queues WHERE recipient_id = ?" (Only (recipientId q))
   size' `shouldBe` size
   canWrt' `shouldBe` canWrt
-  expire' `shouldBe` expire
 
 testResetMessageCounts :: PostgresMsgStore -> IO ()
 testResetMessageCounts ms = do
@@ -369,25 +397,63 @@ testResetMessageCounts ms = do
     Just (Message {msgId = mId2}, False) <- write "message 2"
     Just (Message {msgId = mId3}, False) <- write "message 3"
     Nothing <- write "message 4"
-    q `hasSize` (4, False, True)
+    q `hasSize` (4, False)
     liftIO $ setIncorrectSize q (10, True)
     Nothing <- write "message 5"
-    q `hasSize` (11, False, True)
+    q `hasSize` (11, False)
     (Msg "message 1", Msg "message 2") <- tryDelPeekMsg ms q mId1
-    q `hasSize` (10, False, True)
+    q `hasSize` (10, False)
     (Msg "message 2", Msg "message 3") <- tryDelPeekMsg ms q mId2
-    q `hasSize` (9, False, True)
+    q `hasSize` (9, False)
     (Msg "message 3", Just MessageQuota {msgId = mId4}) <- tryDelPeekMsg ms q mId3
-    q `hasSize` (8, False, True)
+    q `hasSize` (8, False)
     (Just MessageQuota {}, Just MessageQuota {msgId = mId5}) <- tryDelPeekMsg ms q mId4
-    q `hasSize` (7, False, True)
+    q `hasSize` (7, False)
     (Just MessageQuota {}, Nothing) <- tryDelPeekMsg ms q mId5
-    q `hasSize` (0, True, False) -- reset
+    q `hasSize` (0, True) -- reset
   where
     setIncorrectSize :: PostgresQueue -> (Int64, Bool) -> IO ()
     setIncorrectSize q (size, canWrt) =
       void $ withTransaction (dbStore $ queueStore ms) $ \db ->
         DB.execute db "UPDATE msg_queues SET msg_queue_size = ?, msg_can_write = ? WHERE recipient_id = ?" (size, canWrt, recipientId q)
+
+testExpireMessagesInBatches :: PostgresMsgStore -> IO ()
+testExpireMessagesInBatches ms = do
+  g <- C.newRandom
+  emptiedQs <- replicateM emptiedCount $ newQueue g
+  partialQs <- replicateM partialCount $ newQueue g
+  overQuotaQs <- replicateM overQuotaCount $ newQueue g
+  quotaMsgs <- runRight $ do
+    forM_ (emptiedQs <> partialQs) $ \q -> void $ write q "old 1"
+    forM_ emptiedQs $ \q -> void $ write q "old 2"
+    mapM fillPastQuota overQuotaQs
+  -- msg_ts has second granularity, so the recent messages need a new second to be kept
+  threadDelay 1100000
+  boundary <- systemSeconds <$> getSystemTime
+  runRight_ $ forM_ partialQs $ \q -> void $ write q "recent"
+
+  MessageStats {expiredMsgsCount, storedMsgsCount, storedQueues} <- expireOldMessages False ms boundary 0
+  expiredMsgsCount `shouldBe` (emptiedCount * 2 + partialCount + sum quotaMsgs)
+  storedMsgsCount `shouldBe` (partialCount + overQuotaCount) -- recent messages and quota markers
+  storedQueues `shouldBe` (emptiedCount + partialCount + overQuotaCount)
+  runRight_ $ do
+    forM_ emptiedQs $ \q -> checkQueueSize ms q (0, True)
+    forM_ partialQs $ \q -> checkQueueSize ms q (1, True)
+    -- the quota marker is never expired, and the queue stays blocked until it is acked
+    forM_ overQuotaQs $ \q -> checkQueueSize ms q (1, False)
+  where
+    -- expire_old_messages pages through expired messages and commits per page, so these
+    -- counts put a page boundary inside each group of queues.
+    emptiedCount = 120 :: Int
+    partialCount = 40 :: Int
+    overQuotaCount = 10 :: Int
+    newQueue g = do
+      (rId, qr) <- testNewQueueRec g QMMessaging
+      runRight $ ExceptT $ addQueue ms rId qr
+    write q s = writeMsg ms q True =<< mkMessage s
+    fillPastQuota q = go 0
+      where
+        go n = write q "fill" >>= maybe (pure n) (const $ go (n + 1))
 #endif
 
 testQueueState :: JournalMsgStore s -> IO ()

tests/CoreTests/RetryIntervalTests.hs

@@ -14,8 +14,11 @@ import Util
 retryIntervalTests :: Spec
 retryIntervalTests = do
   describe "Retry interval with 2 modes and lock" $ do
-    testRetryIntervalSameMode
-    testRetryIntervalSwitchMode
+    testRetryIntervalSameMode withRetryNewLock2
+    testRetryIntervalSwitchMode withRetryNewLock2
+  describe "Retry interval with 2 modes" $ do
+    testRetryIntervalSameMode withRetryInterval2
+    testRetryIntervalSwitchMode withRetryInterval2
   describe "Foreground retry interval" $ do
     testRetryForeground
     testRetryToBackground
@@ -41,14 +44,18 @@ testFastRI =
       maxInterval = 40000
     }
 
-testRetryIntervalSameMode :: Spec
-testRetryIntervalSameMode =
+type WithRetry2 = RetryInterval2 -> (RI2State -> (RetryIntervalMode -> IO ()) -> IO ()) -> IO ()
+
+withRetryNewLock2 :: WithRetry2
+withRetryNewLock2 ri action = newEmptyTMVarIO >>= \lock -> withRetryLock2 ri lock action
+
+testRetryIntervalSameMode :: WithRetry2 -> Spec
+testRetryIntervalSameMode withRetry =
   it "should increase elapased time and interval when the mode stays the same" $ do
-    lock <- newEmptyTMVarIO
     intervals <- newTVarIO []
     reportedIntervals <- newTVarIO []
     ts <- newTVarIO =<< getCurrentTime
-    withRetryLock2 testRI lock $ \(RI2State slow fast) loop -> do
+    withRetry testRI $ \(RI2State slow fast) loop -> do
       ints <- addInterval intervals ts
       atomically $ modifyTVar' reportedIntervals ((slow, fast) :)
       when (length ints < 9) $ loop RIFast
@@ -65,14 +72,13 @@ testRetryIntervalSameMode =
                        (20000, 40000)
                      ]
 
-testRetryIntervalSwitchMode :: Spec
-testRetryIntervalSwitchMode =
+testRetryIntervalSwitchMode :: WithRetry2 -> Spec
+testRetryIntervalSwitchMode withRetry =
   it "should increase elapased time and interval when the mode switches" $ do
-    lock <- newEmptyTMVarIO
     intervals <- newTVarIO []
     reportedIntervals <- newTVarIO []
     ts <- newTVarIO =<< getCurrentTime
-    withRetryLock2 testRI lock $ \(RI2State slow fast) loop -> do
+    withRetry testRI $ \(RI2State slow fast) loop -> do
       ints <- addInterval intervals ts
       atomically $ modifyTVar' reportedIntervals ((slow, fast) :)
       when (length ints < 11) $ loop $ if length ints <= 5 then RIFast else RISlow

tests/CoreTests/SOCKSSettings.hs

@@ -2,15 +2,18 @@
 {-# LANGUAGE NamedFieldPuns #-}
 {-# LANGUAGE OverloadedLists #-}
 {-# LANGUAGE OverloadedStrings #-}
+{-# LANGUAGE PatternSynonyms #-}
 {-# LANGUAGE TypeApplications #-}
 {-# OPTIONS_GHC -fno-warn-ambiguous-fields #-}
 
 module CoreTests.SOCKSSettings where
 
 import Network.Socket (SockAddr (..), tupleToHostAddress)
+import Simplex.Messaging.Agent.Client (ipAddressProtected)
 import Simplex.Messaging.Client
+import qualified Simplex.Messaging.Crypto as C
 import Simplex.Messaging.Encoding.String
-import Simplex.Messaging.Protocol (ErrorType)
+import Simplex.Messaging.Protocol (ErrorType, pattern SMPServer)
 import Simplex.Messaging.Transport.Client
 import Test.Hspec hiding (fit, it)
 import Util
@@ -19,6 +22,7 @@ socksSettingsTests :: Spec
 socksSettingsTests = do
   describe "hostMode and requiredHostMode settings" testHostMode
   describe "socksMode setting, independent of hostMode setting" testSocksMode
+  describe "ipAddressProtected, consistent with chosen host and socksMode" testIPAddressProtected
   describe "socks proxy address encoding" testSocksProxyEncoding
 
 testPublicHost :: TransportHost
@@ -94,6 +98,19 @@ testSocksMode = do
       let TransportClientConfig {socksProxy} = transportClientConfig cfg NRMInteractive host False Nothing
        in socksProxy
 
+testIPAddressProtected :: Spec
+testIPAddressProtected = do
+  it "should be protected if SOCKS proxy is used for the chosen host" $ do
+    protected SMAlways HMOnionViaSocks [testPublicHost] `shouldBe` True
+    protected SMOnion HMOnionViaSocks [testPublicHost, testOnionHost] `shouldBe` True
+    protected SMOnion HMPublic [testOnionHost] `shouldBe` True
+  it "should not be protected if SOCKS proxy is not used for the chosen host" $ do
+    protected SMOnion HMOnionViaSocks [testPublicHost] `shouldBe` False
+    protected SMOnion HMPublic [testPublicHost, testOnionHost] `shouldBe` False
+  where
+    protected socksMode hostMode hosts =
+      ipAddressProtected defaultNetworkConfig {socksProxy = Just defaultSocksProxyWithAuth, socksMode, hostMode} (SMPServer hosts "" (C.KeyHash ""))
+
 testSocksProxyEncoding :: Spec
 testSocksProxyEncoding = do
   it "should decode SOCKS proxy with isolate-by-auth mode" $ do

tests/CoreTests/StoreLogTests.hs

@@ -22,7 +22,7 @@ import qualified Data.X509.Validation as XV
 import SMPClient
 import Simplex.FileTransfer.Protocol (FileInfo (..))
 import Simplex.FileTransfer.Server.Store (FileRec (..), FileRecipient (..), FileStoreClass (..), RoundedFileTime, STMFileStore (..))
-import Simplex.FileTransfer.Server.StoreLog (FileStoreLogRecord (..), readWriteFileStore)
+import Simplex.FileTransfer.Server.StoreLog (FileStoreLogRecord (..), readWriteFileStore, serverFileName)
 import qualified Simplex.Messaging.Crypto as C
 import Simplex.Messaging.Encoding.String
 import Simplex.Messaging.Protocol
@@ -37,6 +37,7 @@ import Simplex.Messaging.Server.StoreLog
 import Simplex.Messaging.SystemTime
 import Simplex.Messaging.Transport (SMPServiceRole (..))
 import Simplex.Messaging.Transport.Credentials (genCredentials)
+import System.FilePath ((</>))
 import Test.Hspec hiding (fit, it)
 import Util
 
@@ -269,6 +270,18 @@ fileStoreLogTests = do
           saved = [AddFile sId file createdAt Nothing EntityActive, BlockFile sId blockedWithNotice],
           compacted = [AddFile sId file createdAt Nothing (EntityBlocked blockedWithNotice)],
           state = M.fromList [(sId, (file, createdAt, Nothing, EntityBlocked blockedWithNotice))]
+        },
+      SLTC
+        { name = "put file with wrong file name",
+          saved = [AddFile sId file createdAt (Just expiresAt) EntityActive, PutFile sId "/etc/passwd"],
+          compacted = [AddFile sId file createdAt (Just expiresAt) EntityActive],
+          state = M.fromList [(sId, (file, createdAt, Just expiresAt, EntityActive))]
+        },
+      SLTC
+        { name = "put file in another directory",
+          saved = [AddFile sId file createdAt (Just expiresAt) EntityActive, PutFile sId ("/srv/xftp" </> serverFileName sId)],
+          compacted = [AddFile sId file createdAt (Just expiresAt) EntityActive, PutFile sId ("/srv/xftp" </> serverFileName sId)],
+          state = M.fromList [(sId, (file, createdAt, Just expiresAt, EntityActive))]
         }
     ]
 

tests/CoreTests/XFTPStoreTests.hs

@@ -9,7 +9,7 @@ import Data.Word (Word32)
 import Simplex.FileTransfer.Protocol (FileInfo (..), SFileParty (..))
 import Simplex.FileTransfer.Server.Store
 import Simplex.FileTransfer.Server.Store.Postgres (PostgresFileStore, importFileStore, exportFileStore)
-import Simplex.FileTransfer.Server.StoreLog (closeStoreLog, readWriteFileStore, writeFileStore)
+import Simplex.FileTransfer.Server.StoreLog (closeStoreLog, readWriteFileStore, serverFileName, writeFileStore)
 import Simplex.FileTransfer.Transport (XFTPErrorType (..))
 import qualified Simplex.Messaging.Crypto as C
 import Simplex.Messaging.Protocol (BlockingInfo (..), BlockingReason (..), EntityId (..))
@@ -17,6 +17,7 @@ import Simplex.Messaging.Server.QueueStore (ServerEntityStatus (..))
 import Simplex.Messaging.Server.StoreLog (openWriteStoreLog)
 import Simplex.Messaging.SystemTime (RoundedSystemTime (..))
 import System.Directory (doesFileExist, removeFile)
+import System.FilePath ((</>))
 import Test.Hspec hiding (fit, it)
 import UnliftIO.STM
 import Util
@@ -265,7 +266,7 @@ testMigrationRoundTrip = do
       sId2 = EntityId "migration_file_2"
       rId1 = EntityId "migration_rcp_1_"
   addFile stmStore sId1 fileInfo1 testCreatedAt (Just testExpiresAt) EntityActive `shouldReturn` Right ()
-  void $ setFilePath stmStore sId1 "/tmp/file1"
+  void $ setFilePath stmStore sId1 ("/tmp" </> serverFileName sId1)
   addRecipient stmStore sId1 (FileRecipient rId1 rcpKey1) `shouldReturn` Right ()
   let testBlockInfo = BlockingInfo {reason = BRSpam, notice = Nothing}
   addFile stmStore sId2 fileInfo2 testCreatedAt Nothing (EntityBlocked testBlockInfo) `shouldReturn` Right ()
@@ -289,7 +290,7 @@ testMigrationRoundTrip = do
   case result1 of
     Right (FileRec {fileInfo = fi, filePath, expiresAt, fileStatus}, _) -> do
       size fi `shouldBe` 128000
-      readTVarIO filePath `shouldReturn` Just "/tmp/file1"
+      readTVarIO filePath `shouldReturn` Just ("/tmp" </> serverFileName sId1)
       expiresAt `shouldBe` Just testExpiresAt
       readTVarIO fileStatus `shouldReturn` EntityActive
     Left e -> expectationFailure $ "getFile sId1 failed: " <> show e