🤖 Agent: merged after verify and CI parity passed, built merged with current main (launcher-ship-1). #1521: generated docs keep LF line endings on Windows, so the launcher builds there again (dry run 40035909909 green on all three OSes). #1522: Prepare no longer requires Netty's optional native transport jar, which Prism never downloads (fixes#1487; test plus a real Prism-layout prepare).
launcher.yml's Windows build has failed since the reference test landed: Git for Windows checks
text out with CRLF (core.autocrlf), so docs/launcher-commands.md never equalled what markdown()
generates. .gitattributes now pins LF for the generated files tests compare byte for byte, and
the launcher's check reads the file whatever line endings a clone gave it, with a test of a
CRLF checkout.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed, built merged with current main (reports-gate-2: each claim checked against vanilla 26.3, in-frame and --server-thread; evidence on the issue and in the PR body).
🤖 Agent: merged after verify and CI parity passed, built merged with current main (st2: entity use with either hand and server chat components; lan_e2e checks pass in both modes. The hands check's only flake was a vanilla crit arrow killing the cow, fixed in c89695dd9: 40 of 40 runs pass).
🤖 Agent: merged after verify and CI parity passed. A stopgap while in-frame is the default: an overloaded world slows down instead of dropping to 2 FPS (2,500 pigs: 1.7 → 16 FPS). It is removed with the in-frame path (FEATURES 4zg) once the server thread is the default again (#1463–#1466 and following).
🤖 Agent: merged after verify and CI parity passed, built merged with current main (st-default-1: server-thread gaps checked against vanilla, each with a lan_e2e check that fails on main and passes here, host and guest, both modes; tables in the PR body).
🤖 Agent: merged after verify passed, with local proof: all 100 client harnesses ran on main and on the branch and their transcripts were diffed (89 byte-identical, the rest only timings or stamps). 81 client examples become one probe binary.
🤖 Agent: merged after verify passed, with local proof: all 217 world harnesses ran on main and on the branch and their transcripts were diffed (203 byte-identical, the rest only timings or known run-to-run counts). 168 world examples become one probe binary, with 1,097 duplicated helper lines and 11 dead examples gone.
A fully drawn bow fires a crit arrow: AbstractArrow.onHitEntity adds
random.nextInt(damage / 2 + 2) to the 6 damage, so one shot in five does 10
and kills the 10 health cow. The off hand bucket then had nothing to milk
("the host's off hand holds minecraft:bucket"), on either side, in any mode.
Under 20 ticks of draw the arrow is no crit and does at most 6. The check
now also asserts the cow is still there after the shot.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed, built merged with current main (st1: /particle and explosion particles from a server reach its players; lan_e2e evidence in the PR body).
BundleItem.broadcastChangesOnContainerMenu calls slotsChanged(player
inventory) after a bundle click. The menus that don't check which container
changed recompute from their inputs (crafting grids, loom, stonecutter,
cartography, merchant, smithing's hasRecipeError). The menu click harness
found it on a loom with a fixture result: 8000/8000 cases now match (seeds 7
and 11).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed (shaders stack: Java packs draw their own terrain, water and shadow map, as in Iris). BSL shows real shadows on Vulkan and Metal; frame times and screenshots are on #1326. Without a pack, nothing changes.
🤖 Agent: merged after verify passed; main builds again. Local proof on the Mac mini: cargo check -p horizon-client --bins --tests fails on 8a4ebcc55 with the two E0599 errors and passes here.
#1432 deleted both as unused; #1424 (level events) and #1425 (scoreboard)
merged after it and call them, so horizon-client's binary failed with
E0599 at level_particles.rs:54 and scoreboard.rs:338. Both functions come
back unchanged from 4d7f30f14^1.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A carried bundle clicked on a stack took its place instead of taking it in
(#1456): Horizon's menus had no BundleItem click overrides. The PICKUP click
now runs them first, as AbstractContainerMenu.tryItemClickBehaviourOverride:
- a carried bundle takes as much of a clicked stack as its weight allows
(tryTransfer), and a right click on an empty slot puts its first stack
there (removeOne, the rest going back in);
- a stack clicked onto a bundle in a slot goes in as far as it fits, and a
right click with an empty hand takes the bundle's first stack out.
BundleContents.Mutable is in horizon_client::bundle (exact Fraction
weights: 1/max stack, 1 for a beehive with bees, a nested bundle's contents
plus 1/16; shulker boxes don't fit), used by the menu engine and the
LAN/server-thread container engine. The insert, insert-fail and remove-one
sounds play at the player.
In single player the inventory screen's empty-hand press went to an older
click path instead of the menu engine; it now goes to the engine, as on a
server, so the right click on a bundle works there too.
The menu click harness now has bundles (empty, loaded, nested; often in
hand) and compares the player's sounds: 3000/3000 bundle-heavy cases match,
2796/3000 without the overrides.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed; each fix has a lan_e2e check (host and guest) passing on the server thread and in-frame (bugs-st-1).
🤖 Agent: merged after verify and CI parity passed; each fix has a lan_e2e check (host and guest) passing on the server thread and in-frame (bugs-st-1).
🤖 Agent: merged after verify and CI parity passed; each fix has a lan_e2e check (host and guest) passing on the server thread and in-frame (bugs-st-1).
🤖 Agent: merged after verify and CI parity passed (review-0-survey quick win: dead pub items; the generated version.rs is kept as the generator writes it).
A server's players (LAN guests, and single player's window on its server
thread) never had a totem ready, and nothing applied one: lethal damage
killed them with the totem still in hand (#1292). Each guest turn now
readies the totem from the guest's hands and, when its hurt set it off,
uses it up at the end of the turn (LivingEntity.checkTotemDeathProtection):
1 health, the item_used stat, the used_totem trigger, the use vibration,
the totem's effects.
Entity event 35 now goes to every client around, the holder's own too,
and a client plays it as ClientPacketListener.handleEntityEvent does: the
totem particles, the sound heard locally (not sent on as a sound packet,
so no client hears it twice) and the holder's rumble.
lan_e2e totem: host and guest, off hand and main hand, both modes.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Single player's world ticks in the frame by default (#1428). When its ticks
take longer than 50 ms, every frame fell behind and ran the client timer's
ten, so the picture stalled at 1-2 FPS (a player's report on an AMD 680M,
in-frame tick at 36.7 ms). Vanilla's world ticks on the integrated server's
thread, which falls behind without costing frames and logs "Can't keep up!".
A frame now stops running world ticks once they took a tick's length and
drops the rest, logging vanilla's overload warning (1 s + 20 ticks behind,
at most every 10 s + 100 ticks). Ticks of a few milliseconds still get the
full ten per frame; sprints and server clients are unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed. The parity probe prebuild repeats each harness's own -p and features selection, so harness builds after the prebuild compile nothing (0 of 10, was 10 of 10). The full harness set ran on it locally in #1453's before/after diff.
🤖 Agent: merged after verify passed, with local parity proof (103/103 on this head). GUI sprites are measured in GUI units whatever the pack's resolution, as in vanilla: inventories, tooltips and toasts look right with Faithful 32x and 64x.
🤖 Agent: merged after verify passed, with local parity proof: parity --changed passed 103/103 on #1458's head, which contains this PR. Faithful 64x reproduced a startup panic on main (a 32768 px shared atlas over the 16384 GPU limit); with the fix it reaches the world. A pack that can't fit now falls back to the default resources, as vanilla does.
🤖 Agent: merged after verify and CI parity passed; release dry run 39935111815 was green (the smoke test drew a world, no blocker, it would publish as the release and latest). One channel, as Ludovic decided on 2026-10-23.
🤖 Agent: merged after verify and CI parity passed; lan_e2e (hands, styles, feedback, chat, frames, riding, mounts, beds, contacts, worldspawn) passes on the server thread and in-frame (bugs-st-2).
block_survival and state_provider hashed the name as an artifact instead of
passing it; decorated_region runs chunk_dump through chunk_generation's binary.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed. No runtime change for players: #1445 deletes 10 tools/ scripts that nothing references (no harness depends on them); #1446 exports the mod-SDK symbols only on wasm32, so the guest crates link on macOS, and the mod sandbox tests pass.
🤖 Agent: merged after verify passed. No runtime change for players: #1445 deletes 10 tools/ scripts that nothing references (no harness depends on them); #1446 exports the mod-SDK symbols only on wasm32, so the guest crates link on macOS, and the mod sandbox tests pass.
🤖 Agent: merged after verify passed; docs only (AGENTS.md, CLAUDE.md symlink, docs/architecture.md). Every path and harness name was checked against the repo.
81 client examples that harnesses run become modules of examples/probe:
harnesses run `probe <name> ...` (or `cargo run --example probe -- <name>`),
and the shared helpers (audio_common, hud_fixture, chat_screen_fixture)
move with them. The benchmarks and manual tools stay standalone examples.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
main.rs now only lists the probes; dispatch.rs runs them. The runner leaves
main.rs out of a probe-binary harness's dependencies and keeps dispatch.rs,
so adding a probe no longer reruns (or flake-checks) every harness of its
crate. It also follows crate::<probe> references between probes.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Eight benches (block_facts, chunk, entity_geometry, json_readers, load,
save, streaming, tick_chunk_membership), lan_flat, save_parity and
vanilla_diff: no harness, tool, doc or workflow names them (review plan,
section 1, dead code). -2,222 lines; Git history keeps them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Now that the probes are modules of one binary, the helpers copied between
them live once in examples/probe/support: the collision probes' shape
printer (emit, values, box_values, bounds, 8 copies each), the canonical
NaN bits of f64 and f32 (50 copies), the registry JSON-to-NBT reader (15)
and the JSON block position (4). -1,097 lines of copies.
f64::NAN.to_bits() is 0x7ff8000000000000, the literal some copies used,
so every probe prints the same bytes.
tools/parity.py follows include!: three probes include another probe's
source, and stay dependent on it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The 168 horizon-world examples the parity harnesses run become modules of
one example, examples/probe, that dispatches on its first argument;
crate::args() gives each the command line its example saw. Linking ~170
binaries, each instantiating horizon-world's generics again, was most of
the parity job's probe build (review plan, section 1).
The harnesses build `--example probe` and run `probe <name> ...`.
tools/parity.py counts only the probes a harness ran, and the support
modules they name, as its dependencies, so editing one probe still reruns
only its own harnesses (and PARITY_RUST_SCOPE keeps working).
The six manual tools (benches, loot_fill, structure_blocks, vanilla_diff,
attribute_memory, behaviour_load_profile) stay separate examples.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A record naming an example since renamed or deleted made the whole Cargo
run fail ("no example target named ..."), so no probe was prebuilt and
every harness built its own: the probe binaries rename ~250 of them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
With Faithful 32x the inventory showed the top-left quarter of its
background stretched over the screen, tooltips lost their background and
toasts doubled in size (#1442): the GUI code drew every atlas sprite in
texels, where vanilla draws in GUI units whatever the pack's resolution.
- Atlas sprites carry their GUI size: the size of the default resources'
texture of that name (what vanilla's `blit(..., 256, 256)` calls and
sprite sizes assume), their texels only for what copies pixels (font
sheets, fire frames). Textures only a pack has keep their own size.
The texture size filter reads the GUI size too, so Faithful 64x's
1024x1024 container backgrounds are no longer dropped.
- Nine-slice sprites follow `GuiGraphicsExtractor.blitNineSlicedSprite`:
their `gui.scaling` width, height and borders (Faithful's tooltip
background is 50x50 with a 10-pixel border), `stretch_inner`, the
single-quad and three-part cases, and borders capped at half the
drawn size by integer division.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The shared GUI atlas was always 4096 texels wide, so Faithful 64x needed
4096x32768: past the 16384 texels a side GPUs allow, and wgpu panicked
in create_texture at every launch until the pack was removed (#1442).
- The atlas widens (4096, 8192, 16384) until it fits within 16384 a
side, so Faithful 64x loads. Its height rounds up to a multiple of
1024 instead of a power of two: the default resources' atlas drops
from 4096x8192 to 4096x5120, Faithful 32x's from 4096x16384 to
4096x12288.
- Textures that still don't fit are an error, not a GPU panic. At
startup it rolls the packs back as Minecraft.rollbackResourcePacks
does: the game starts with its own resources, and the selection is
cleared and saved (clearResourcePacksOnError), so the next launch
doesn't fail again. The existing "Resource reload failed" screen
shows the reason.
- The renderer refuses an atlas larger than the device allows with an
error instead of a wgpu validation panic.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed; release dry run 39935111815 was green (the smoke test drew a world, no blocker, it would publish as the release and latest). One channel, as Ludovic decided on 2026-10-23.
AGENTS.md (CLAUDE.md links to it): the crates and what each owns, how to
build and test fast, where each subsystem of horizon-client lives, how Game
is organised, the two single-player paths and how to test both, the vanilla
reference, and the rules that bite. docs/architecture.md: feature -> files ->
checks for the 40 areas changed most in the last 30 days.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed. Metal GPU occlusion culling: on the M4 Pro at RD 32, FPS is +8% to +48% (interleaved A/B against main); the hidden sections match Vulkan's; parity --changed passed 104/104.
The 120-line hand-written central-directory and local-header parser goes:
the zip crate (already the client's dependency for world backups) indexes
the archive, zip64 and odd offsets included. Entries are still read raw and
inflated here, so as with Java's ZipFile (vanilla's FilePackResources)
their CRC isn't checked. New tests: folder entries, a nested pack root, a
Windows-style name, one broken entry among good ones, and files that aren't
zips.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
free_port (7 copies), offline_uuid (4), varint and parse_varint (3 and 2),
dbits (6), fbits (4) and settled (2) move into lab.py; every caller uses
lab's. The other server harnesses keep importing free_port from
test_dedicated_server_parity, which now takes lab's. Imports the moved
code left unused go too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The file is generated from the pinned version lock, and test_gen_version_consts
checks JAVA_VERSION against it; deleting constants by hand breaks the next
regeneration and that test.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A host build of a mod (cargo build or test on the workspace) made a cdylib
whose component exports macOS's ld couldn't take: its exported-symbols list
cuts `horizon:mods/[email protected]#init` at the `#`, so coords-hud and
horizon-mods-test-guest failed to link. The exports now exist only for
wasm32; on the host the macro still checks the mod and keeps it used.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after the contribution gate. The claims were checked against Iris and OptiFine, and the weak parts were rewritten on the branch: the stage kinds, the Java shadow feed and the fake uniforms were removed; settings now apply on Done; packs are imported by drag and drop; memory is only allocated while a pack reads it; the MakeUp target is fixed. Tested on Vulkan and Metal with 6 packs, profiles, imports and broken packs; local parity passed 96/96. Frame times before and after are in the PR comment. Thanks @kerneldroid.
command_parse_summary, structure_start_oracle, fetch_client_assets,
compare_structure_starts, profile_behaviour_cache, command_tree_diff,
compare_loot_fill, compare_structure_blocks, test_lan_lifecycle and
PROFILE_SHARED_JSON.md: no workflow, doc, harness, test or other tool
names them (review-plan.md, Dead code). History keeps them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
world_spawn starts a vanilla and a Horizon server for each of 60 seeds,
one at a time: 1,144-1,548 s on the CI runner, the parity job's longest
harness and its floor whenever a PR's changes reach it. With
PARITY_SHARDS = 4 each shard creates every fourth seed's worlds, in its
own build directory.
The shard helper datafix had moves to tools/parity_shard.py, which both
use. It stays out of lab.py, which every harness imports.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
As in Iris's ShaderPackScreen: at the end of the list it scrolled out
of sight with a few packs. It's greyed out while the picked pack has no
options.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The probe prebuild ran `cargo build --workspace --examples --bins` (or
named targets with --workspace). The harnesses build with `-p <crate>`, and
Cargo unifies features over the selected packages: under --workspace,
flate2, libc, memchr, syn, thiserror, postcard and digest get other
features for horizon-world, and horizon-mods (pack) and horizon-telemetry
(upload) for horizon-client. So every probe the prebuild made was a
variant no harness used, and each harness rebuilt its crate and its probe
again while the others waited on Cargo's lock.
Record each probe with its command's package and feature flags and
whether that command builds examples (which turns dev-dependencies on),
and prebuild one Cargo run per such selection. --bins and --examples are
recorded too, so the fast-profile horizon-client build that
save_roundtrip and world_reopen run is prebuilt as well. A harness that
never ran here takes its probes from the shared dependency lists instead
of a build of all ~275 examples.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
cargo test --workspace built and linked all ~275 examples, 59% of the
test build (review-plan section 1), though none of their tests run.
Test the libraries, binaries and integration tests, then the doc tests,
as tools/gate.py and tools/pr_ready.py already do. Clippy --all-targets
still type-checks every example, and the parity job builds and runs them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The world target made the opaque depth, normal and material textures
for every scaled or anti-aliased frame, packs or not: 12 bytes a pixel
nobody read. They now follow what the pack reads, and Post::generation
tells the stages to bind again when its targets change (a recreated
target of the same size, or a shadow map made after the stages bound).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
MakeUp's gbuffers write colortex1, and its composites read the world
from there: with the world in colortex0, it showed nothing but fog.
Iris's fallbacks for terrain pick the program (gbuffers_terrain, then
textured_lit, textured, basic); its first draw buffer takes the world.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Review fixes for #1414 (contribution gate):
- Java packs no longer get Horizon's shadow map: it was mirrored in y and
reversed in depth against the GL conventions the packs sample with, had
no distortion and only the camera's sections, so BSL drew no shadows
while the pass cost 0.65 ms (MakeUp, M4 Pro). shadowtex* stays a far
depth, as on main; the packs' own shadow programs are #1326.
- Invented uniforms removed (centerDepthSmooth fixed at 0.5,
ambientOcclusionLevel and shadowMapResolution as uniforms, constant
temp/rainfall/biome/held items); an unknown uniform stays zero as
before. wetness keeps its Iris smoothing. Unused directives removed.
- texture.<stage>.<sampler> is scoped to Iris's stages (final reads the
composite ones) instead of overriding the sampler in every pass;
texture.noise replaces noisetex. One noise generator and one RGBA
upload for both pack kinds.
- The G-buffer pass, the opaque-depth copy and the shadow pass run only
for a pack that reads them (naga's per-entry-point binding use, not a
text search): every pack paid all three before.
- Api 2's prelude additions go in their own file after api 1's, so api 1
packs keep the names they declare (aces_tonemap, material_at...).
- Stage kinds shadow, gbuffer, deferred and final ran as composite: only
composite is accepted; api 2 adds formats (validated), profiles,
option descriptions and a noise texture.
- Shader Pack Settings applies once when the page closes, as Iris's Done
(each click or slider step rebuilt the pack: 2.7 s per click, 12.7 s
for one drag with BSL). Labels follow Iris (.lang, else the name).
A pack whose options can't be read isn't read again every frame.
- Saved option values an option doesn't list are ignored, as OptiFine's
ShaderOption.setValue (a stale file broke BSL's compile).
- Import: the blocking file dialog (osascript/zenity/kdialog/PowerShell
on the event loop) is gone; a pack dropped on the Shader Packs page is
copied and picked, as Iris's onFilesDrop. Drops elsewhere do nothing,
an existing pack isn't overwritten, links aren't followed.
- Docs and the example say what the code does (SSAO, not GTAO).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
On Apple's GPUs the passes the test adds cost a fixed share of the frame
(it is stored and reloaded at each), which the hidden sections repay only
at high render distances: at 12 culling cost an M4 Pro 0.05 ms of GPU time
a frame, at 32 it saved 0.25 to 1.6 ms. Other backends keep culling every
frame; HORIZON_GPU_CULL=all culls every frame on Metal too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed; lan_e2e (hands, styles, feedback, chat, frames, riding, mounts, beds, contacts, worldspawn) passes on the server thread and in-frame with all four merged on main (bugs-st-2).
🤖 Agent: merged after verify and CI parity passed; each fix has a lan_e2e check (host and guest) passing on the server thread and in-frame (bugs-st-1).
horizon-client's dev-dependency on naga added the hlsl-out feature for the
gfx.rs test that translates every shader for DirectX 12. With resolver 2 a
dev-dependency's features apply only to tests and examples, so Cargo kept
two builds of naga, wgpu-hal, wgpu-core, wgpu, horizon-shaderpack and the
horizon-client lib, and rebuilt them on the first example or test after a
game build.
Measured on the Mac mini (cargo build of the game, then one example):
the example took 37.3 s and recompiled that whole stack before, 1.2 s
and only its own crate after; the target dir is 1 GB smaller (4.9 to
3.9 GB). The game binary is the same size: nothing calls the HLSL writer,
so the linker drops it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed. In-frame single player is the default again until the server thread reaches parity (FEATURES 4zg); release builds were compared with v0.2.0 and v0.2.11 on the Mac mini. The server thread stays available with --server-thread or HORIZON_SERVER_THREAD=1.
93 pub items (78 functions and methods, 14 constants, 1 struct) in horizon-client,
horizon-world, horizon-core, horizon-friends and horizon-server-files had
no caller anywhere in the repository: not in the crate, the other crates,
the examples, the tests or the binaries. rustc can't flag them because
horizon-client and the others are libraries, where a pub item counts as
used. The imports and the private helpers only they used go with them
(title.rs date_string, goals.rs flag_index).
Also drops the #[allow(dead_code)]/#[allow(unused)] items that only hid
dead code: chat_debug_fill, merchant_title and Goals::locked_by had no
caller, and four stubs (_keep, _unused, _rule_kind, const _) existed only
to keep an import used. LightTables::from_parts is used by its tests, so
it is #[cfg(test)] instead of allowed.
Protocol id tables and entity data serializer ids stay: they mirror
vanilla's registries in full. The launcher's three unused items are left
for a launcher PR.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Flat draws (one index list per frame, the default on Apple GPUs at high
render distances) turned occlusion culling off. Now phase 2's draws are
flattened after the test, by a compute pass that copies only the sections
the test saw, and each page run is drawn indirectly with the count it
reached.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
GPU occlusion culling ran on Vulkan only: its test needs early fragment
tests, which naga's Metal backend does not emit. On Metal the boxes are
tested in a pass of their own: phase 1's depth is bound read-only as the
attachment and as a texture, and the fragment shader compares its own
depth with it (reversed Z, greater or equal), as the fixed-function test
would. HORIZON_GPU_CULL=shader takes this path on Vulkan too, for checks.
Frames drawn flat (the default on Apple GPUs at high render distances)
still skip culling here.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Every release is now published as the release (one channel, Ludovic's
decision of 2026-10-23), so the launcher drops its Stable/Alpha choice:
the Updates row in Settings, `horizon-launcher update-channel` and its MCP
tool, the channel on the Versions page, in `versions` and in the footer.
It updates to the newest game release that is not a pre-release (GitHub's
latest) and looks every 30 minutes. The pre-releases of the old Alpha
channel stay listed and installable by hand, never as an update. Old
`update_channel` / `beta_builds` keys in settings.json are ignored.
On Linux a release build that starts no longer rewrites the
applications-menu entry to point at itself. It writes the entry only when
there is none, or when the program it starts (TryExec, else Exec) no longer
exists; turning the Settings switch on or `desktop-entry` still points it
at this launcher, and HORIZON_LAUNCHER_DESKTOP_EXEC is still honored. A
test build started with the real HOME took over a working entry and then
left it broken when its target/ was deleted (2026-10-23).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Ludovic's decision (2026-10-23): no more Alpha and Stable. release.yml
publishes every game release as the release and GitHub's latest instead of
a pre-release that promote.yml made stable once a day. The checks that
gated a promotion run in the release itself: `decide` holds the train while
an issue labelled release-blocker is open (forced runs too), and the publish
job runs `release_train.py gate` right before publishing: the Mac mini's
smoke-test.json must be there and show a world was drawn, and no blocker may
be open. Otherwise nothing is published.
Notes cover everything since the last release that is not a pre-release, so
the first single-channel release lists all of it since v0.2.0. promote.yml,
tools/release_channels.py and its tests are deleted. The cadence is
unchanged (5 minutes of settling, 20 at most, release-now at once).
Docs and site: one channel in release-flow, versioning, security, the FAQ,
install, the home page and the Releases page (Game / Launcher filter); no
player-facing Alpha wording left.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify and CI parity passed; each fix has a lan_e2e check (host and guest) passing on the server thread and in-frame (bugs-st-1).
Players on the Alpha channel report single player as broken since v0.2.1
(#1265 made the integrated server thread the default again). On v0.2.11's
server thread, placed signs open no editor (#1380), held books don't open
(#1366), a totem doesn't save its holder (#1292) and death keeps the
inventory with keep_inventory off (#1284); the in-frame world and Stable
v0.2.0 do all four right. About 45 more [server-thread] issues are open.
`--server-thread` or HORIZON_SERVER_THREAD=1 turns it on, as before #1114;
`--no-server-thread` and HORIZON_SERVER_THREAD=0 still work. LAN guests and
dedicated servers are untouched. Temporary until the server thread reaches
parity (FEATURES 4zg).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A server never sent particles: /particle on single player's server thread
drew into the server's own engine, and TNT exploded without its burst for
anyone but an old-mode host. The protocol gains
ClientboundLevelParticlesPacket and ClientboundExplodePacket; /particle
sends the particles to each viewer in range (every player without
viewers, as PlayerList.getPlayers) and an explosion sends its packet to
the players within 64 blocks; their clients draw them as
handleParticleEvent and handleExplosion do.
/particle's spread was measured from the command's source as if it were a
position; it is the spread as given (Vec3Argument.getVec3).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The server never sent its scoreboard, so a sidebar shown on single
player's server thread or to a LAN guest stayed empty. Each tick a placed
player is now sent what changed since it was last sent the board, as
ServerScoreboard's callbacks do: every team, and the objectives in a
display slot with their slots and scores (all of it on joining).
A server's client also finds its team's sidebar by its profile name rather
than single player's "Player".
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A client of a server (single player on its server thread, a LAN guest, a
dedicated server's player) played only three of the level events the server
sent (a block broken, the dragon egg's and a consumable's teleport): a
spawner's burst, waxing and scraping copper, a composter filling and the rest
showed nothing. Every one now goes through the handler the in-frame game uses
(LevelEventHandler).
Bone meal's level event (1505) played its sound but never its growth
particles, in either mode: it now spawns them as BoneMealItem.
addGrowthParticles does (spread over a grass block's, nylium's, netherrack's
or moss's neighbourhood and water, inside a growing block, below rooted dirt
and mangrove leaves). The server also sends it to the player who used the
bone meal, as vanilla's levelEvent names no player.
Waxing copper with honeycomb and an axe or hoe's transforms (wax off, scrape,
strip, till) make their particles and sound on the clicker's own client, as
vanilla's client runs Item.useOn; the server sends them to the others only.
The transform's preview is shared with the in-frame use.
`horizon-ctl status` counts the particles made by type. lan_e2e's new
`sparks` check: the host and the guest each use bone meal on wheat, wax
copper and take the wax off, and a spawner spawns a zombie; both sides make
the particles.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A splash or lingering potion, an experience bottle or an eye of ender a
player threw was never an entity a server tracked: on single player's server
thread, for LAN guests and on a dedicated server the item was used up and
nothing flew.
The server now numbers them like its other projectiles and tracks them as
vanilla's entities: splash_potion, lingering_potion and experience_bottle
with ThrowableItemProjectile.DATA_ITEM_STACK (the potion, for its colour),
and eye_of_ender. A client draws them where the server puts them and drops
them when the server removes them.
lan_e2e's new `thrown` check: the host and the guest each throw a splash
potion and an eye of ender, and both sides see them fly.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A server never told its clients who holds a mob's lead, so on single
player's server thread, a LAN guest's or a dedicated server's client the
lead was used up and the mob tied, but no rope showed.
The tracker now sends ClientboundSetEntityLinkPacket as vanilla does: in
ServerEntity.sendPairingData for a leashed entity, and when the lead is tied
or comes off (Leashable.setLeashedTo / dropLeash, the holder 0). The holder
is the player the mob sees (a guest, else the host) or a mob; a fence knot
isn't an entity on Horizon's server yet. The client sets the mob's holder
from it: itself, another player or a mob.
lan_e2e's new `leads` check ties and unties a cow on both sides.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A client of a server (single player on its server thread, a LAN guest, a
dedicated server's player) only applied a rabbit's baby flag and a creeper's
swell from SetEntityData. It now reads the synched data vanilla's entities
read in onSyncedDataUpdated: Entity's on-fire flag, LivingEntity's health
(the wither's boss bar follows it), the baby flag of every ageable mob,
Sheep.DATA_WOOL_ID, TamableAnimal's sitting and tame flags, the wolf's and
the cat's collar colour, and the cube mobs' size. A remote wolf or cat starts
with vanilla's red collar.
The server sends the wolf's and cat's collar colour. Pairing still carries
only the values that differ from the defaults; a mob's state now lists the
defaults of the entries that can go back to theirs (health, the baby flag,
the wool, the collar), and the tracker sends that default when one does:
a calf growing up or a sheep's white wool growing back reach the clients.
Sheep saves wrote the colour and shearing they were loaded with (0 unless
the save had them) and loads ignored both: they now save and load the wool
as it is (Sheep.add/readAdditionalSaveData).
lan_e2e's new `looks` check summons the mobs and changes them on both sides.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The server told joining players of a default border and never of a
change, so on single player's server thread and for LAN guests the wall
neither showed nor stopped anyone. Players now get the real border on
entering a level and each change as BorderChangeListener sends it.
The player's movement collides with the border's walls
(CollisionGetter.borderCollision) instead of being clamped after moving,
which a sprint-jump got past. A guest (and single player's own player on
its server thread) outside the border takes its damage on the server, as
the host's player did.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The client of a server (single player on its server thread, LAN guests)
ignored ClientboundUpdateAttributesPacket for its own player, so a new
maximum health never reached it: its hearts stayed at ten and the health it
was sent was clamped to 20. It now applies each synced instance's base and
modifiers (handleUpdateAttributes) and re-applies the received health.
BossBar::new started hidden; ServerBossEvent starts visible.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
save_roundtrip found an allay saved with InLove 0: Horizon's `animal`
flag is broader than vanilla's Animal class.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed; #1190 has interleaved GPU A/B against main on the HD 620 and M4 (within noise) and 103 parity harnesses passing; #1413 was proven with a 1080p run on the desktop.
🤖 Agent: merged after verify passed; #1190 has interleaved GPU A/B against main on the HD 620 and M4 (within noise) and 103 parity harnesses passing; #1413 was proven with a 1080p run on the desktop.
🤖 Agent: merged after verify passed; lan_e2e (absorption breath lava contacts shield deathmining worldspawn drops underwater beds minecarts levels) passes in both single-player modes on the top of the chain, plus clippy and the client tests.
🤖 Agent: merged after verify passed; lan_e2e (absorption breath lava contacts shield deathmining worldspawn drops underwater beds minecarts levels) passes in both single-player modes on the top of the chain, plus clippy and the client tests.
🤖 Agent: merged after verify passed; lan_e2e (absorption breath lava contacts shield deathmining worldspawn drops underwater beds minecarts levels) passes in both single-player modes on the top of the chain, plus clippy and the client tests.
🤖 Agent: merged after verify passed; lan_e2e (absorption breath lava contacts shield deathmining worldspawn drops underwater beds minecarts levels) passes in both single-player modes on the top of the chain, plus clippy and the client tests.
🤖 Agent: merged after verify passed; lan_e2e (absorption breath lava contacts shield deathmining worldspawn drops underwater beds minecarts levels) passes in both single-player modes on the top of the chain, plus clippy and the client tests.
🤖 Agent: merged after verify passed; lan_e2e (absorption breath lava contacts shield deathmining worldspawn drops underwater beds minecarts levels) passes in both single-player modes on the top of the chain, plus clippy and the client tests.
🤖 Agent: merged after verify passed; lan_e2e (absorption breath lava contacts shield deathmining worldspawn drops underwater beds minecarts levels) passes in both single-player modes on the top of the chain, plus clippy and the client tests.
- Critical hits also need the attacker out of water, not blind and not
riding (Player.canCriticalAttack) (#1393).
- Mending repairs the off hand too: candidates in EquipmentSlot order,
main hand, off hand, then the armour (#1349).
- Phantoms spawn only in the Overworld with spawn_mobs, spawn_monsters and
spawn_phantoms on, near every player (the local one and each guest in
the level, with their own time since rest), with vanilla's draw order
(#1388, #1387).
- A level entered by a portal or a dimension change keeps the world's
natural_health_regeneration (#1339).
- /item replace ... with air empties the slot, and counts it as changed
as vanilla's SlotAccess.set does (#1332).
- Thrown potions stay where they are on frozen ticks (/tick freeze) and
land once the game runs again (#1364).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- The remote client sends Interact for a targeted mob only for a hand
vanilla's client-side Mob.interact would accept with it (leads, name
tags, food, buckets, shears, saddled mounts, tamed pets, ...), so a bow
drawn or food eaten while aiming at a cow goes on to useItem (#1350).
- Both hands: the main hand's item, then the off hand's entity, block and
item (Minecraft.startUseItem). The server runs Interact with the packet's
hand (GuestUse::Entity), and mobs, armour stands, leads and mounts use
the held hand's stack, so an off-hand bucket milks a cow and off-hand
wheat feeds it (#1362).
- The entity tracker sends an item frame's item and rotation when they go
back to empty and 0, so a punched frame shows empty on clients (#1355).
- lan_e2e `hands`: the bow, the golden apple, the off-hand bucket and wheat
and the emptied frame, for the host and the guest.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- Guests' player data saves Fire, Air and fall_distance and puts them back
on their body when they come back (single player's players are guests
of the integrated server), so a burning player still burns after a
reopen (#1323); the old single player's level.dat keeps fall_distance
too (#1327).
- Every Animal saves and loads InLove, so a fed animal stays in love
across a reopen and is not fed again (#1396).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The sneaking case read the health after releasing Shift and waiting 10
ticks on the magma, so a hit from standing up counted (and one could land
between the first reading and the key press).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A bed sets the respawn point as vanilla's ServerPlayer.startSleepInBed does
(once in reach, day or night): a RespawnConfig at the bed's head with the
player's yaw, not forced; straw beds never set it (straw_bed_rule).
Respawning resolves it through findRespawnAndUseSpawnBlock, which now
handles beds: AbstractBedBlock.findStandUpPosition (beside the bed on the
side away from the yaw it was set from, around it, then on it; bunk beds
too) through DismountHelper, facing the bed. A bed that is gone or boxed
in sends the player to the world spawn with "You have no home bed or
charged respawn anchor, or it was obstructed", and the point is forgotten:
in the window's world, on the server thread and for LAN guests. Leaving
the End goes to the bed too (not using it up).
Saves keep a bed as the player's `respawn` (resolved through its block
when loaded), and a `spawnpoint` in the Overworld as forced. A fallback
that is the world spawn follows `setworldspawn` with a bed or anchor set.
tools/test_use_item_on_parity.py: respawn cases at a bed's head set from
some yaw (intact, bunk, only the foot left or gone, crowded), against
vanilla's real findRespawnAndUseSpawnBlock; 4000/4000 and 12000/12000 on
two seeds. lan_e2e `beds`: intact, gone and boxed in, host and guest.
Fixes#1308
Co-Authored-By: Claude Opus 5.5 <[email protected]>
perf_bench ran every game in a 1280x720 window only. It now runs the
standard set at 1280x720 and 1920x1080 (--sizes picks others), each size
its own reports: <label>-rd<N>.json at 720p as before, <label>-1080p-rd<N>.json
at 1080p.
Horizon takes vanilla's --width and --height for its window's starting size
(1280x720 by default); perf_bench passes them to every game, and to Horizon
only off 720p, so builds before this one still run the 720p set.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The vignette, drawn over every frame, went from one texture sample to up to
four texel loads and per-pixel address arithmetic: +0.85 ms of GPU a frame on
an HD 620 (integer modulo, removed in the commit before), still +0.4 ms after
it.
- A blurred sprite gets a one-texel border in the atlas, copied from its
opposite edges: the hardware's bilinear sample of the atlas with the
entity group's linear sampler is then GL_LINEAR with repeat, edges
included. Clamp-to-edge keeps the sample point half a texel inside the
sprite, which is what the hardware's clamp does for both filters.
- Each sampling has its own pipeline, the shader specialised by override
constants: a repeat sprite runs main's arithmetic and one sample, nearest
or linear; only clamp adds a clamp. The sampling moves from the vertices
to the quad, so the vertex layout is main's again.
- A black vignette changes no pixel (it darkens by texture x colour), and is
no longer drawn: in daylight its brightness fades to 0 within half a
minute, so most frames outdoors skip it, which main never could.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and a proof run on the Mac mini: the self-check passed (parity 22/22), and six TNT minecarts now chain (6 → 0; 5 survive on main). Thanks @levzzz5154!
🤖 Agent: merged after verify passed and a proof run on the Mac mini: the self-check passed (parity 22/22), and lan_e2e minecarts passes in both single-player modes. Thanks @levzzz5154!
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
The server sends a knocked-back, pushed or jumping mob's move at once
(Entity.needsSync), and the remote client walks every living entity along
the server's positions with vanilla's SteppedInterpolationHandler: each
position takes the type's update interval to reach, sped up while steps
pile up. Before, the client waited for the next due update and then
lerped there in fewer ticks, so a hit pig jumped away fast.
tools/test_stepped_interpolation_parity.py drives vanilla's unmodified
InterpolationData and the Rust port with the same packets and ticks
(random cases and a punched pig's knockback); 135236 positions and
rotations match bit for bit.
Fixes#1374
Co-Authored-By: Claude Opus 5.5 <[email protected]>
ItemEntity.tick: an item stuck inside a full block (one set on it) has no
physics on the server and moves toward the closest free side
(Entity.moveTowardsClosestSpace: north, south, west, east, up; 0.1 to 0.3 a
tick), instead of staying in the block (#1252).
A server's client (LAN guests, and the window on the server thread) digs at
the same speed as the server: getDestroyProgress with the submerged mining
speed, haste, fatigue and efficiency, not the plain tool speed, so it no
longer finishes early under water and gets the block put back (#1254).
lan_e2e: `drops` (a diamond under a block set on it comes out, in both
views) and `underwater` (dirt by hand at eye height under water: about
halfway at 35 ticks on client and server, broken by 105).
Fixes#1252, #1254.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
SetWorldSpawnCommand sets the level's spawn and its angle; a player without
a bed or anchor respawns there (it used to set the player's own spawn
point, so other players and later deaths kept the old one). The angle is
saved in level.dat's spawn (yaw, pitch), and a guest whose saved respawn
point is the world spawn follows it when it moves.
lan_e2e `worldspawn`: /setworldspawn 24 blocks east, then the host and the
guest die and respawn there.
Fixes#1310.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
ItemCooldowns ticks once a tick for each player: a server's players count
their shield cooldown down in tick_guests, and the host's world in tick_world
(it was in tick_using, which ran for the host and again for every guest's
swap, so a guest's shield came back early on the server thread).
A guest that dies stops digging (its mining, the delay and the crack go), so
the block it was breaking doesn't break once it respawns.
lan_e2e: `shield` (a vindicator's axe disables each player's shield; the
server's cooldown then falls 40 in 40 ticks) and `deathmining` (die while
breaking obsidian, respawn, the obsidian stays). `lan status` and the agent
state report the shield cooldown.
Fixes#1315, #1316.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Entity.applyEffectsFromBlocks: MagmaBlock.stepOn burns a player standing on
it who is not sneaking (hot_floor, 1), CactusBlock and CampfireBlock.entityInside
hurt what touches them (cactus 1, campfire 1, soul campfire 2, lit only).
A server's players (LAN guests, and the host on the server thread) now get
these and the grown sweet berry bush prick along the moves their client sent
that tick: the server rebuilds the tick's movement history and runs the
damage half of the block effects.
lan_e2e `contacts`: magma, sneaking on magma, a lit campfire, a cactus, grown
and young berry bushes, for the host and the guest.
Fixes#1305, #1311, #1312, #1313.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The server sends a player's own air supply and absorption amount
(Entity.DATA_AIR_SUPPLY_ID, Player.DATA_PLAYER_ABSORPTION_ID) when they
change, and the client applies them: on the server thread and as a LAN
guest the breath meter follows the server's Respiration-extended air
(#1304) and the golden hearts show (#1306). The connected client's tick
sets the helmet's Respiration and the Nether's FAST_LAVA as the window's
own tick does (#1309).
AbsorptionAmount is written from the live body and read back on load,
in the window's world and for a server's players (#1307).
lan_e2e: absorption, breath and lava checks.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The camera overlays (the vignette every frame) read up to four texels a
pixel, each wrapped with two integer modulos. Intel GPUs emulate integer
division. Reducing the coordinate to one period first (fract, as the shader
did before) leaves every read at most one texel outside the sprite, so a
select wraps it. Same texels and weights; clamp-to-edge is unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the agent-control-2 integration of all six PRs on the Mac mini. agent_control_e2e passes in CLI and MCP (one known survival flake), launcher_mcp_e2e passes, No AI is refused, and the options, screen_layout, menu_click and world_options parity harnesses pass. Ludovic asked for this on 2026-10-23.
🤖 Agent: merged after verify passed and the agent-control-2 integration of all six PRs on the Mac mini. agent_control_e2e passes in CLI and MCP (one known survival flake), launcher_mcp_e2e passes, No AI is refused, and the options, screen_layout, menu_click and world_options parity harnesses pass. Ludovic asked for this on 2026-10-23.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
On Apple GPUs timestamps go only at pass boundaries, and the frame was
bracketed by two one-thread compute passes. Two things made it read
about 0.01 ms:
- Metal resolves counter samples without waiting for the passes that
write them. Resolved in the frame's own command buffer, the end
timestamp was mostly the one an earlier frame left in the shared
query. Each readback slot now has its own queries, and with markers a
slot is resolved in a later frame, once on_submitted_work_done says the
GPU finished the frame it timed.
- A compute pass depends on nothing, so the end marker ran alongside the
frame's render passes. The end is now a render pass that loads the
frame's target, shading one pixel with every write masked: it runs
after every pass that drew into the target. The start stays a compute
pass (cheaper than a load and store of the target).
M4 Pro, render distance 32, offscreen: GPU p50 0.002 ms before, now
1.95 ms against a 1.95 ms frame (village) and 4.46 ms against 4.51 ms
(mountain), where the frame is GPU-bound. Vulkan and DirectX 12, which
write timestamps between commands, are unchanged but for the per-slot
queries.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the wave-3 integration on the desktop by integ-w3: main + 17 PRs, fmt, clippy 0, 1929 workspace tests, parity --changed (only known flakes failed, and those passed on rerun), lan_e2e in both modes, and the perf numbers back to back against main posted on each PR.
🤖 Agent: merged after verify passed and the agent-control-2 integration of all six PRs on the Mac mini. agent_control_e2e passes in CLI and MCP (one known survival flake), launcher_mcp_e2e passes, No AI is refused, and the options, screen_layout, menu_click and world_options parity harnesses pass. Ludovic asked for this on 2026-10-23.
🤖 Agent: merged after verify passed and the agent-control-2 integration of all six PRs on the Mac mini. agent_control_e2e passes in CLI and MCP (one known survival flake), launcher_mcp_e2e passes, No AI is refused, and the options, screen_layout, menu_click and world_options parity harnesses pass. Ludovic asked for this on 2026-10-23.
🤖 Agent: merged after verify passed and the agent-control-2 integration of all six PRs on the Mac mini. agent_control_e2e passes in CLI and MCP (one known survival flake), launcher_mcp_e2e passes, No AI is refused, and the options, screen_layout, menu_click and world_options parity harnesses pass. Ludovic asked for this on 2026-10-23.
🤖 Agent: merged after verify passed and the agent-control-2 integration of all six PRs on the Mac mini. agent_control_e2e passes in CLI and MCP (one known survival flake), launcher_mcp_e2e passes, No AI is refused, and the options, screen_layout, menu_click and world_options parity harnesses pass. Ludovic asked for this on 2026-10-23.
🤖 Agent: merged after verify passed. Ludovic asked on 2026-10-23 for every hardware fact players send (opt-in). The end-to-end run against a local server is in the description: breakdowns need ≥3 installs, with no install ID or exact value. The consent version goes up to 2, so players who opted in are asked again, as the privacy page promises.
🤖 Agent: merged after verify passed and a local run on the Mac mini. lan_e2e on the server thread passes for host and guest: boats, minecarts, frames, ladder, lightning, riding and mounts. Parity --changed passed 25 of 25, and the server-thread default is confirmed. Single player runs on its server thread again. Thanks @rdbtCVS for pushing for the root fix.
Under Wayland the title screen can come before the window is on a screen, so the
report's hardware.display stayed empty (seen end to end in a headless sway).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- The desktop measured: Horizon v2026.10.22.13 against vanilla 26.3 at render
distance 12 and 32, 3 runs each, 1280x720, VSync off, in a headless sway on
the RTX 4080 (never on its screen). Replaces the "not yet measured" note.
- /benchmarks shows each players' group's hardware split (RAM, VRAM, drive,
system, GPU kind) and, once the telemetry server has some, a "Players'
hardware" table from /v1/aggregate's breakdowns. Hidden without data.
- The picker takes the memory too. Players' groups whose installs have neither
the picked RAM nor the GPU's VRAM class count as further away, and the result
says what they had and when a game's peak memory is a large share of the RAM.
- Each GPU in the spec table has its memory (0: shared).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
website/docs/agents.md, in the nav under "Going further", is the player's page.
It covers what an agent can do, connecting it from the launcher, copy-paste
config for Claude Code, Codex, Cursor and any MCP client, the game's AI Agent
page, a CLI quick reference, safety and No AI.
docs/agent-control.md is rewritten for developers:
- the surfaces, and why one MCP server covers the launcher and the game;
- the inventory, before and after, for the launcher and the game;
- No AI on every surface, and the launcher commands that stay;
- connecting from the launcher and the game;
- the design rules for a new command.
docs/launcher-commands.md lists every launcher command, its arguments and its
MCP tool. It is generated from the clap definition (cli_tools::markdown). The
launcher test the_launcher_reference_is_up_to_date fails when it is stale or a
tool is undocumented, like horizon-ctl's generated reference.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Horizon Settings gains a row: "AI Agent..." and "No AI: Off". No AI hides the
AI Agent page and closes the control socket at once.
The AI Agent page has an Agent Control switch (horizon-extras.txt
agentControl, off by default): on, the game the player starts opens the
default control endpoint, so horizon-ctl and MCP clients reach it without
--control. Turning it off closes the endpoint at once; dropping a control
server now also stops its listener thread. The page shows whether an agent
is connected, when it last sent a command and the socket, and Copy
Connection Details copies the socket path, a horizon-ctl line to try and the
MCP config for this game.
agent_control_e2e.py --scenario setting covers it: Agent Control on opens the
endpoint without --control, the pages list their widgets, Copy works, and
turning the switch off removes the endpoint while the game keeps running.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Right after a window is made, Wayland doesn't say which screen it is on, so
current_monitor() is None and display was missing (seen end to end in a headless
sway). Fall back to the primary, else the only, monitor, and read it again when a
session ends; agent control's status (the benchmark report's display) uses the same.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Settings gets an AI group with the No AI switch (horizon_core::no_ai, the
same switch the game reads) and, while it is off, a Connect an AI agent
section that replaces the old command-line note: a picker for Claude Code,
Codex, Cursor or any MCP client, the ready-made config with a Copy button,
the client's own add command (claude mcp add / codex mcp add) shown and run
only when the player clicks Add, and Test the connection, which starts the
MCP server as a client would and says how many tools answered. Client CLIs
are found on the PATH or in the folders installers use, since a launcher
started from the desktop gets a short PATH on macOS.
horizon-launcher agent-config --test runs the same test; the launcher
harness checks it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The MCP tools are generated from the CLI's clap definition (cli_tools.rs):
mods_search is horizon-launcher mods search, with the same arguments, and
a call runs that command with --json. So the mod browser, modpacks, hosted
servers, settings, game arguments, data collection, desktop entry and the
benchmark are all reachable, and new CLI commands become tools by
themselves. Destructive ones need confirm: true and carry destructiveHint.
The session keeps play, server (with server_command/log/stop),
account_list and account_sign_in in-process, renamed after the CLI
(launch_client is play; play takes join for an invite code).
New CLI commands: settings [KEY VALUE] (selected version, folders,
install-new-versions, opening animation) and agent-config, the MCP config
for Claude Code, Codex, Cursor and any client. play takes --world and
--creative. Under No AI, mcp and agent-config are hidden and refused.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The table gains what only the raw socket could do (options pages, connect
and invite codes, disconnect, LAN, Play with Friends, books, signs, mount
inventories, effects, xp, enchant, potion, fill, containers) and new
commands: worlds, world-rename, world-backup, world-delete --confirm (the
Select and Edit World screens, from the title screen only), servers (the
Multiplayer list) and interact-entity (right-click an entity). status,
mouse, pad, release, bench and stats become MCP tools; summon takes a
position. Destructive and network tools carry MCP annotations.
docs/horizon-ctl-commands.md is generated from the table
(horizon-ctl help --markdown) and a test fails when it drifts. The
agent_control_e2e manage scenario exercises the new commands.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
horizon_core::no_ai reads and writes the noAi line of the game's
horizon-extras.txt, so the launcher and the game share one switch. While
it is on the game never opens its control socket, horizon-ctl refuses every
command with exit 4 and kind denied, horizon-ctl mcp and horizon-launcher
mcp refuse to start (and refuse calls if it is turned on mid-session), and
horizon-launcher no-ai [on|off] shows or sets it. The game's extras keep
the key when they rewrite the file. Both e2e harnesses check it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
main (#1271) pins Horizon Settings and Mods as the grid's last buttons;
this branch adds Credits & Attribution after them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed. Go-day step 5 (docs/versioning.md): Ludovic said go on 2026-10-23 ("tu peux envoyer la 0.2"). The bridges are out (v2026.10.22 and launcher-v0.17.0), and the latest stable is v2026.10.22.07, built after the freeze. The launcher redesign (#1269) and the mod browser are merged.
🤖 Agent: merged after verify passed. Ludovic approved the redesign on 2026-10-23. Integration run on the desktop by integ-w3 (main + the mod browser stack + this), then main merged and re-checked on the Mac mini (launcher clippy, tests).
What players send (opt-in, unchanged switches) now describes the whole machine:
CPU cores and base/boost clock, the GPU's maker, integrated or discrete and its
VRAM, the drive under the game folder (NVMe/SSD/HDD), the screen (size, refresh,
scale), on battery or mains, laptop or desktop. Perf, crash and bug items carry
them in `system`; benchmark reports in `hardware`. No serials, host or user
names, or paths. CONSENT_VERSION goes to 2, so players are asked again.
The server keeps the new fields in `systems.extra` (the table is rebuilt once
on old databases, ids kept). /v1/aggregate and /v1/benchmarks/summary add
breakdowns by RAM class, VRAM class, drive, OS and integrated/discrete GPU;
a class is published only with at least 3 different installs, never a raw value.
website/scripts/benchmarks.mjs mirrors the classes on the shared fixtures.
The summary's "not Horizon or plain vanilla" check returned None (accepted)
instead of the rejection; it now matches the site's script.
The launcher's "View what would be sent" and `data show` list every field.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
tools/lan_e2e.py: main added the mounts check; levels stays last (it moves
the players between dimensions, and skips itself without --server-thread).
CHECKS and CHECK_FNS: 67 each.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
#1271 removed the title's live world backdrop (backdrop_live), so the
locator bar hides on the title alone.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the integration run on the desktop (integ-w3: main + the mod browser stack + #1269; fmt, clippy, 1224 + 131 tests, mod_browser_harness ok).
🤖 Agent: merged after verify passed and the integration run on the desktop (integ-w3: main + the mod browser stack + #1269; fmt, clippy, 1224 + 131 tests, mod_browser_harness ok).
🤖 Agent: merged after verify passed and the integration run on the desktop (integ-w3: main + the mod browser stack + #1269; fmt, clippy, 1224 + 131 tests, mod_browser_harness ok).
🤖 Agent: merged after verify passed and the integration run on the desktop (integ-w3: main + the mod browser stack + #1269; fmt, clippy, 1224 + 131 tests, mod_browser_harness ok).
🤖 Agent: merged after verify passed and the integration run on the desktop (integ-w3: main + the mod browser stack + #1269; fmt, clippy, 1224 + 131 tests, mod_browser_harness ok).
🤖 Agent: merged after verify passed and the integration run on the desktop (integ-w3: main + the mod browser stack + #1269; fmt, clippy, 1224 + 131 tests, mod_browser_harness ok).
tools/lan_e2e.py: main added the minecarts, frames, lightning and riding
checks; levels stays last (it moves the players between dimensions, and
skips itself without --server-thread). CHECKS and CHECK_FNS: 66 each.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- game.rs: main (#1271) dropped the Horizon-theme checks from menu_blur and
the title scroll; kept without them, with this branch's WinScreen parts
(the credits alone blur, the title doesn't scroll under the Win screen).
Screen keeps main's Friends and this branch's Win.
- control.rs: horizon-ctl options pages keep credits, horizon and
shader-packs.
- options_ui.rs: Page keeps Credits and main's Horizon. The Options grid is
main's (Horizon Settings and Mods in vanilla's Telemetry and Credits slots;
Horizon Extras now sits inside Horizon Settings) with Credits & Attribution
after them, laid out by this branch's row count.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
hud.rs: HudAtlas::slots keeps this branch's trim size parameter, with
main's doc line (the on-screen keyboard's text replaced the Horizon menu
text, #1271).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
graphics.rs: main (#1271) dropped the level key from horizon-graphics.txt;
farDetail stays after lodDistance in main's format (21 keys, 21 values).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A server's players (single player on its server thread, LAN guests,
dedicated-server players) get the horse inventory: shift + use on a tamed
horse, donkey, mule, camel or llama, or the inventory key while riding one
(ServerboundPlayerCommandPacket OPEN_INVENTORY), opens HorseInventoryMenu
over the mount with ClientboundMountScreenOpenPacket. Clicks run on the
menu engine's new Kind::Mount (saddle and body ArmorSlots that take what
the mount may wear, the chest, AbstractMountInventoryMenu.quickMoveStack),
and the slots go back to the mount with the item's equip sound.
The client shows a server's mount screen, predicts its clicks with the
same engine and sends OPEN_INVENTORY when the inventory key is pressed
while riding. Llamas now have their inventory in the window too: the
carpet slot and a chest of `strength` columns, no saddle slot.
lan_e2e `mounts` checks it for the host and a guest, with and without
--server-thread.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The home page's "At a glance" list gets a fifth row with how many play right now, read from
https://telemetry.playhorizon.dev/v1/live (3 s timeout, again every minute while the tab is
visible). It stays hidden while nobody plays or the server doesn't answer. The FAQ no longer
says the "I'm playing" ping is off: the server is live since #1276.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed. telemetry.playhorizon.dev is live on Railway (service telemetry in horizon-site, volume /data): /healthz, /v1/live (a test ping counted) and /v1/benchmarks/summary (CORS for playhorizon.dev) answer. The PR's local end-to-end runs are in the description.
launcher/src/main.rs: main (#1271) removed the Game menus setting (one menu
style now) where this branch restyled Settings. Kept the Game and About
group headings, dropped the Game menus row as main does, and renumbered the
rows after it (game arguments n+4, Horizon n+5, Launcher n+6).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed. Local proof on the merged head bf274d92: fmt, clippy -D warnings, client and launcher tests, parity --changed origin/main, and test_friends.py --server-thread. CI parity passed on the earlier head. Every Horizon setting is reachable from vanilla-style screens (list in the PR body).
ServerExplosion.hurtEntities skipped minecarts: a TNT minecart's blast
never reached the carts beside it. Carts now take the explosion's damage
through VehicleEntity.hurtServer (destroyed past 40, where
MinecartTNT.destroy primes a TNT cart with a 0-39 tick fuse, as
damageSourceIgnitesTnt counts explosions) and the knockback from their
eyes. The exploding cart is the explosion's source, so it is removed
before the blast and can't hit itself. Cart ticks now look carts up by id,
since a blast can remove or prime other carts mid-loop.
lan_e2e takes main's version: #1280 kept `boats` (passing with and
without --server-thread) and added `minecarts`, `frames` and `lightning`
next to `ladder`, and main's run() already sets HORIZON_SERVER_THREAD
explicitly as this branch does. The code keeps this branch's change:
single player runs on its server thread by default again.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Single player on its server thread, LAN guests and dedicated-server
players could not ride: a server player's turn had no mount input, and
its client had nothing to sit on.
The server now drives a server player's animal from that player's
PlayerInput (forward, back, strafe, jump, sneak, sprint) and rotation,
through the same tick_mount_riding the window's player uses:
- the jump bar charges per player (MountState swapped with the turn);
- sneaking gets off, with SetPassengers and the teleport beside it;
- a mount ridden by another player can't be taken or fed (isVehicle);
- an animal whose rider left stops taking its input.
The client sits where positionRider puts it on the animal it was sent,
and charges its own jump bar for the HUD.
Ridden movement (pre-existing, also in the window): Mob.setSpeed sets zza
too, so the ridden input set before the speed was lost and pigs and
striders crawled while horses walked on with no key held; the input is
now set after the speed (travelRidden). The ridden speed uses the
effective movement speed (a cold strider's modifier, potions), and a
shivering strider rides at 0.35 of it (Strider.getRiddenSpeed).
lan_e2e riding: the host and a guest saddle and ride a horse (walk and
jump), a pig and a strider with their sticks, a camel and a donkey, sit
on a llama, harness and fly a happy ghast, and get off, with and
without --server-thread. `mobs` reports ridden, controlled, saddled.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
OldMinecartBehavior.tick ends with pushAndPickupEntities: after a cart
moves, every other cart in its box (0.2 wider) gets push(cart). Horizon
pushed all pairs once before any cart moved, so carts placed inside each
other stayed exactly overlapping (push skips distances under 1e-4) and
never split apart. Pushes now run per cart, after its move, with
pushOtherMinecart's Mth sin table facing and float thresholds.
With #1285 Play with Friends works on the server thread: friends_view
offers sharing there, so the pause menu's Friends icon is active on both
paths and inactive only on another server.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged on verify plus local proof. tools/test_friends.py passes with the host on its server thread and in the frame (code, self-hosted relay and short code, stop). Parity --changed: 22/0. Whichever of #1271 and this lands second drops #1271's server-thread greying of the Friends icon.
🤖 Agent: merged on verify plus local proof, since CI parity is queued behind the Mac mini runners. Parity --changed origin/main on this head: 102/0. lan_e2e passes with and without --server-thread, for the host and for a LAN guest: boats, minecarts, frames, ladder, lightning. The default is unchanged here; #1265 follows once riding animals work for a server's players too.
Conflict: #1237's weather now returns its own vertices; the title draws
none (the live title backdrop is gone).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the contributor integration on the desktop (v5: workspace tests 1887/0, naga 0 failures, parity --changed 304/0). Rendering checked with before/after scenes: the screenshots match, and per-frame CPU and GPU time goes down.
Main's Screen::blurs_background (in-world menu blur) has no arm for the
new Screen::Win. WinScreen.extractBackground fills the screen with the End
portal while the poem rolls and falls back to Screen.extractBackground
(blurred) for the credits alone, so Game::menu_blur blurs only for the
credits without the poem.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the contributor integration on the desktop (v5: workspace tests 1887/0, naga 0 failures, parity --changed 304/0). Rendering checked with before/after scenes: the screenshots match, and per-frame CPU and GPU time goes down.
Stacks #1197 on #1194 so it merges cleanly after it. game/packs.rs: both
PRs add one line to the resource reload; both are kept (waypoint styles,
then armour-stand trims).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
tools/lan_e2e.py: main added the ladder and boats checks where this branch
added levels. Both are kept, with levels still last: it moves the players
between dimensions (and skips itself without --server-thread).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
hud.rs: HudBatch keeps both new fields, end_portal (this branch) and
blur_at (main's menu blur), in the struct, its constructor and the test
batch. draw_range already draws EndPortal runs on either side of blur_at.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the contributor integration on the desktop (v7: core 99 and worldgen 81 tests, 23 worldgen/density/noise parity harnesses 23/0; bit-exact by construction, with a test against the scalar path).
Sharing a world with an invite code refused with the integrated server
on its own thread ("open a singleplayer world first"): the window is a
client there. It now shares that server as Open to LAN publishes it:
friends_start publishes it when it isn't, the invite tunnel points at
its port, stopping sharing unpublishes it again, and the pause menu
shows the Friends entry with the players the server lists.
test_friends.py --server-thread runs the host's world on its server
thread; the host's own move and blocks go through /tp, /fill and
/setblock, which the server answers in both modes.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
"Horizon 2026.10.22.07: #1255, mob pushes, fix, #1222" now shows "Single player runs in the frame again
until server players have vehicles and 3 more".
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- ThinkPad (i7-7500U, HD Graphics 620, Mesa): Horizon v2026.10.22.04 (two flights per scene) against vanilla
26.3 (two games) at render distance 12. Java loads chunks faster there and flies the mountain faster; the
page says so. Render distance 32 isn't published for it: neither game loaded the spawn in 5 minutes, and
Horizon kept only 900 to 1,400 columns resident when given 25.
- "Where Java is ahead" also lists scenes where vanilla's average beats Horizon's by more than 5%.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Conflicts: the menu blur (#1249) now runs without the Horizon style check,
chat and the tab list draw under every screen as before the blur split,
and the HUD's second range draws without the removed colour grade. Play
with Friends blurs the world behind it like vanilla's other menus.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
#1234 replaced column_size with a LazyLock table and #1238 added the
lightning quad count to gfx.rs's max chain; this branch rewrote both
places. weather.rs keeps this branch's code with #1234's table and tests;
gfx.rs takes the max over cloud, lightning and weather quads.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
With single player on its server thread the window drew no lightning,
and LAN guests and dedicated-server players never did: the server never
sent the bolt (it lived outside the tracked entities) and the client had
nothing to draw for a lightning_bolt AddEntity.
The server gives each bolt an entity id and tracks it like the other
entities; the client adds a visual-only bolt for it (LightningBolt.tick
off the server: sounds, the sky flash, the drawn bolt) and ticks bolts
and the flash in its level tick. `status` reports bolts and flashes, and
lan_e2e's `lightning` check summons one beside each player and asserts
both sides get the bolt and the flash.
lan_e2e: the vehicle checks put the players apart first, and run()
sets HORIZON_SERVER_THREAD explicitly either way, as #1265 does.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Single player on its server thread, LAN guests and dedicated-server
players could not board, drive or break boats and minecarts, nor use
item frames and armor stands: the client sent Interact/Attack only for
mobs, the server aimed only at mobs, and nothing carried riding over
the network.
Client (remote.rs, remote_vehicles.rs):
- Interact/Attack for boats, minecarts, item frames, armor stands and
paintings, with the hit location (Minecraft.hitResult).
- SetPassengers seats the player; a boat ticks on the rider's client and
goes to the server as MoveVehicle and PaddleBoat every tick; a
MoveVehicle back puts it where the server says; a minecart moves on the
server and the rider sits in it.
Server (lan_vehicles.rs, lan_actions.rs, hub.rs):
- look_at_entity aims at every pickable object, at the client's hit point.
- The guest's vehicle is its Entity.vehicle during its turn; tick_boats
and tick_carts leave guest-ridden vehicles to their riders.
- handleMoveVehicle with its checks (invalid values kick, too quickly,
moved wrongly, colliding with anything new), handlePaddleBoat.
- Player.rideTick: sneaking gets out, then SetPassengers and the
teleport beside the vehicle (dismountTo); a minecart moves with the
rider's input; a passenger's moves only turn it.
- SetPassengers for guests through hub.riding; one rider per vehicle.
lan_e2e: boats (host and guest board, paddle, the other side sees the
boat move, sneak out, break), minecarts and frames (item frame and armor
stand) for both, passing with and without --server-thread.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the contributor integration on the desktop (v5: workspace tests 1887/0, naga 0 failures, parity --changed 304/0). Rendering checked with before/after scenes: the screenshots match, and per-frame CPU and GPU time goes down.
🤖 Agent: merged after verify passed and the contributor integration on the desktop (v5: workspace tests 1887/0, naga 0 failures, parity --changed 304/0). Rendering checked with before/after scenes: the screenshots match, and per-frame CPU and GPU time goes down.
🤖 Agent: merged after verify passed and the contributor integration on the desktop (v5: workspace tests 1887/0, naga 0 failures, parity --changed 304/0). Rendering checked with before/after scenes: the screenshots match, and per-frame CPU and GPU time goes down.
#1264 wrapped build_surface in `if !self.identity_end_surface(pos)` where
this branch adds the uniform_biomes field to the same SurfaceQueries. Kept
the condition with the new field inside it; both tests are kept.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the wave 2 integration run on the desktop (main + 16 contributor PRs, tree 71afc562: workspace tests 1887/0, launcher 113, naga translations 0 failures, parity --changed 304/0).
🤖 Agent: merged after verify passed and the wave 2 integration run on the desktop (main + 16 contributor PRs, tree 71afc562: workspace tests 1887/0, launcher 113, naga translations 0 failures, parity --changed 304/0).
🤖 Agent: merged after verify passed and the wave 2 integration run on the desktop (main + 16 contributor PRs, tree 71afc562: workspace tests 1887/0, launcher 113, naga translations 0 failures, parity --changed 304/0).
🤖 Agent: merged after verify passed and the wave 2 integration run on the desktop (main + 16 contributor PRs, tree 71afc562: workspace tests 1887/0, launcher 113, naga translations 0 failures, parity --changed 304/0).
🤖 Agent: merged after verify passed and the wave 2 integration run on the desktop (main + 16 contributor PRs, tree 71afc562: workspace tests 1887/0, launcher 113, naga translations 0 failures, parity --changed 304/0).
🤖 Agent: merged after verify passed and the wave 2 integration run on the desktop (main + 16 contributor PRs, tree 71afc562: workspace tests 1887/0, launcher 113, naga translations 0 failures, parity --changed 304/0).
🤖 Agent: merged after verify passed and the wave 2 integration run on the desktop (main + 16 contributor PRs, tree 71afc562: workspace tests 1887/0, launcher 113, naga translations 0 failures, parity --changed 304/0).
🤖 Agent: merged after verify passed and the wave 2 integration run on the desktop (main + 16 contributor PRs, tree 71afc562: workspace tests 1887/0, launcher 113, naga translations 0 failures, parity --changed 304/0).
Without min_chunks (the old call) wait_loaded behaves as before, which tests/test_perf_bench.py's
old-builds case relies on.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- DEFAULT_ENDPOINT is https://telemetry.playhorizon.dev. The heartbeat, the launcher's live count and the
outbox upload all go there; HORIZON_TELEMETRY_URL still names another server, and HORIZON_TELEMETRY_URL=off
sends nothing at all. CI workflows and the harnesses that start the game or the launcher set off. Consent
is unchanged: the four switches stay opt-in, only the anonymous ping is always on.
- GET /v1/benchmarks/summary: players' benchmark uploads grouped by GPU, CPU, render distance and game, groups
of 3+ installs only, medians per install then across installs, no install ID, day, OS version, driver,
memory or screen. Cached 10 minutes, CORS for the website. Same rules as website/scripts/benchmarks.mjs,
tested on the same fixtures; a deleted install leaves its groups at once.
- /v1/aggregate no longer lists the newest 50 raw benchmark reports (hardware, driver, day, one install each):
uploads are published only through the summary.
- Behind Railway's proxy the rate limit counts X-Real-IP, else the last X-Forwarded-For entry (the one the
proxy adds; a client could choose the first). Still in memory only, never stored or logged.
- crates/horizon-telemetry-server/railway.json: Dockerfile build from the repository root, health check
/healthz, one replica in europe-west4 (one SQLite file on the /data volume).
- Docs: docs/telemetry.md, the server README (the summary's shape, Railway setup), the privacy page (the ping
is on, where the data lives, how benchmark uploads are shown).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Right after a teleport, and at spawn, Horizon v2026.10.22.04 reports nothing queued and steady chunk counts
for over a second before it starts loading. wait_loaded() took that for "loaded" and stopped the clock with
0 to 3 sections meshed (load 0.1 s), so the flights ran over empty terrain and the frame rates meant nothing.
Now the counts must change once since the wait began (or stay at the full view for 5 s), and the resident
chunks must reach 80% of the view circle (pi x RD^2), before a steady second counts as loaded.
Measured with it on the Mac mini and the ThinkPad for the website's Benchmarks page (#1274). At RD 32 a 4-thread
laptop needs --timeout above the default 300 s.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Three runs of each game at render distance 12 and 32 (tools/perf_bench.py, seed 1234, 1280x720, VSync off).
Horizon's average sits at the 144 Hz display (macOS presents in step with it even with VSync off): marked as
capped on the page.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
- /releases: every game and launcher release on one page, newest first, Stable / Alpha / Launcher marked and
filterable. The newest three are open; each shows what it added and fixed, the full notes, its downloads and
a link to GitHub. Baked at build time, refreshed from GitHub when a newer release is out. /releases#release-<tag>
opens one.
- Home: "Recently added" replaces the patch notes block: the last five releases on one line each and three
highlights from what they added.
- /benchmarks: Horizon against vanilla Java Edition 26.3 measured on our machines (average FPS, 1% lows, p95
frame time, start to world, chunk loads, memory, at render distance 12 and 32), saying where Java is ahead;
a GPU + CPU picker that shows a measurement when there is one and otherwise a marked estimate with a range,
from our own spec table (method explained on the page); players' results read live from the telemetry
server's GET /v1/benchmarks/summary, falling back to the baked public/benchmarks.json, groups of 3+ installs
only.
- scripts/benchmarks.mjs builds public/benchmarks.json from our reports (and, for tests, telemetry items) with
the same rules as the server; fixtures shared with the server's test.
- Releases and Benchmarks in the nav and footer; docs/performance.md links the page.
- Pixelify Sans' "fi" ligature drew "fixes" as "Axes": common ligatures off.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Captured on the Mac mini with the game's offscreen capture (1280x720),
replacing the Horizon menu style's Video page.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The pause menu's Friends icon is recorded as a widget, so a controller
reaches it and the control socket names it. The sharing screen's long help
URL ran off the screen: a Troubleshooting... button beside Done opens it,
and the hint drops out when it would reach the footer. The graphics docs
say what the two Horizon screens hold (vanilla's own settings stay on Video
Settings).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Drop the Horizon menu style (options_horizon.rs, front.rs, the live title
backdrop and its post grade, the launcher's menu-style setting, CLI and MCP
tool). Every screen is vanilla's. Horizon's own settings hang off vanilla's
Options the way mods add theirs: Horizon Settings... and Mods... take the
Telemetry/Credits slots, and Horizon Settings opens Graphics, Advanced
Graphics, Shader Packs, Controller Settings, Extras and Mods, each a vanilla
options list with tooltips. Play with friends moves behind vanilla's Friends
buttons: the pause menu's opens Play with Friends (share, copy, stop), the
title screen's opens Join a Friend, and Direct Connection takes a full code.
Old files load as before: horizon.txt's menuStyle line and
horizon-graphics.txt's level line are ignored.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The comma form of modf (`v.y = trunc(x), x - trunc(x)`) printed without
brackets in a declaration's initialiser, which naga's GLSL front end then
read as a second declarator. A unary + around it has glsl-lang print the
brackets.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The game keeps the offline name as "name:Player12345" in horizon-player.txt;
the top bar and the friends-only server list showed the whole line. Both
now read the name after the key.
With no version installed, Home has no landscape: its title starts below
the top bar like the other pages, and the sidebar says "No version
installed" instead of "Needs preparing".
Co-Authored-By: Claude Opus 5.5 <[email protected]>
main's warmed-timing change and this branch both edited the usage comment
at the top of examples/chunkgen_bench.rs. Kept this branch's line, which
adds [--snapshot PATH]; no code change.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The translator now legalises what these packs write and naga refuses:
lookups without a level outside fragment shaders become textureLod at
level 0; struct varyings travel as one varying per field; textureGather
is emulated with texelFetch; a bool given to a numeric constructor
becomes cond ? 1 : 0; `const in` parameters lose the const; array fields
of private structs get their natural stride (naga lays every struct out
as std140); modf into a vector component assigns it instead of passing a
pointer. shaders.properties keeps the empty lines the preprocessor
leaves, so Photon's `) \` before `#else` ends its value.
real_packs_load now also parses every translated program again as WGSL,
which is what wgpu is given.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Revert commit cdaf1d79a0, preserving the ladder regression check added afterward.
Server-player bugs also affect dedicated servers and Open to LAN. Running single player in the frame only hides those bugs there; keep the shared server-thread path as the default and fix the underlying issues.
🤖 Agent: merged on verify plus local harness proof (CI parity is queued behind about 15 runs). The new lan_e2e ladder check: on main, ten blocks of ladder took the guest from 20 to 13 health; with the fix it lands at full health. Main was merged in, keeping both the ladder and boats checks.
Versions and Patch notes list a hundred releases, and every frame laid out
and painted all of them. Rows more than a screen away from the viewport are
now empty blocks as tall as when they were last drawn (cull.rs), so the
page keeps its length; a row being installed or showing an error is always
drawn. Scrolling Versions on the Mac mini: median frame 8.2 -> 7.0 ms.
HORIZON_LAUNCHER_PROJECT=<slug> opens a project's page on the Mods page,
for screenshots, as HORIZON_LAUNCHER_PAGE opens a page.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged on verify plus local harness proof. It is a release-blocker fix, and CI parity is queued behind about 15 runs on the Mac mini. lan_e2e boats fails on main with --server-thread (riding None) and passes with this PR's default (in the frame, the path that was the default until #1114). fmt and clippy -D warnings pass. Server players' vehicles are tracked in FEATURES.md 4x.
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
Blocks palettes (dark default, light, or the system's, chosen in Settings), Pixelify Sans for
headings and controls, Pixelarticons icons, square panels with an outline, a bevel and a lip,
inventory tabs, and pixel art drawn in code: a stone texture behind the pages and a landscape
behind Home. A top bar with the pages and the account chip, and a sidebar with the instances
(Horizon shows the grass block the mod browser draws from the player's prepared textures), the
hosted servers, and the launcher's version and channel. The mod browser's pages (kinds as tabs,
project icons in slots, verdicts as badges, the project page) are restyled, not rebuilt. The
benchmark also measures scrolling the Versions list.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
#1200 and the teams change appended test modules at the same place as this
branch's push_tests. Main's modules are kept as they are and push_tests
follows them unchanged.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
🤖 Agent: merged after verify passed and the integration run on the desktop (main 34c3c1cf + the 14 contributor PRs: workspace tests 1,869/0, launcher 113, parity --changed 320/0).
Player's MoveSimulationType is AUTHORITATIVE_SIDE_AND_SERVER, so a
ServerPlayer travels on the server too, and every tick on a ladder or
vine LivingEntity.handleOnClimbable resets its fall distance. Horizon's
server only replayed the client's moves (ServerPlayer.doCheckFallDamage),
so the fall distance grew all the way down a ladder and hurt the player
where it landed: LAN guests, dedicated-server players and, while single
player ran on the server thread (#1114), the single player.
lan_e2e ladder: the guest slides down ten blocks of ladder and lands with
full health (before: 20 -> 13).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A column whose box lies entirely inside the frustum holds no section the
frustum would reject (plane distances only grow from its nearest corner
to a section's farthest, and rounding keeps the order), so its sections
skip the test. At render distance 32 most columns are past the distance
where the vertical field of view takes in their whole height.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
horizon-theme gains the Blocks palettes (light and dark, WCAG AA on every panel and on the stone),
Pixelify Sans (OFL) and Pixelarticons (MIT); tokens.css now carries the Blocks colours, the pixel
font and the block edges. The site sets headings, navigation and buttons in Pixelify Sans, turns
the index lists into stacks of panels with an outline, a bevel and a lip, puts a stone texture
behind the page and Horizon's pixel landscape (a dawn or a sunset) behind the hero, both drawn in
code by scripts/art.mjs. Light and dark still follow the system, with the toggle.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Since #1114 single player's window is a player of its own server thread,
and a server's players can't use boats or minecarts: the client sends
Attack and Interact only for mobs (remote_start_attack, remote_use_entity),
the server aims a player's attack or use only at mobs and dragon parts
(look_at_entity), and neither side handles riding (no SetPassengers on the
client, no MoveVehicle or PaddleBoat packets). So a boat on water could not
be boarded or broken in single player.
The server thread is opt-in again (--server-thread or
HORIZON_SERVER_THREAD=1); --no-server-thread still works.
lan_e2e boats: the host right-clicks a boat on water and rides it, sneaks
out, and breaks it. Without --server-thread the host now runs the game's
default instead of forcing the frame, so the check covers what players get.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The pack lists them with the mods left off, the port's log and fetch label say
the same, and docs/modpacks.md no longer asks for a key.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Importing a modpack from the Mods page (or `mods install`) now ports it right away
for the selected version, so Play works straight after. port-pack leaves a
namespace two carried mods both claim to none of them (YUNG's mods all ship
`yungsapi`, which made the game drop three of them), and the launcher applies
the game's first-wins rule before it starts a pack, listing what it left off.
Measured on Vanilla Perfected, Slimes Adventure and All the Mods 10 (docs/modpacks.md):
2, 20 and 42 mods ported, every world loads, no crash.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
mod_browser_harness.py searches, checks and installs a mod through the launcher's
mods CLI in a scratch game folder, opens a world with it offscreen, places one of
its blocks and crafts one of its recipes on a crafting table, and times the world
load with and without it. Macaw's Fences: check 4.0 s, install 3.1 s, world load
5.4 s -> 7.3 s, hedge placed, 4 crafted.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A result's title opens its page in the launcher: icon, authors, downloads,
followers, licence, verdict (for a Java mod, what ports and what doesn't) and
Install, links, then Description (Markdown), Gallery and Versions tabs
(browser::details, Source::details on Modrinth's project and members).
Filters: sort (relevance, downloads, follows, newest, updated), loader, game
version, environment and categories, from the source's tag lists.
Icons and gallery images are fetched at run time over HTTPS from the sources'
image CDNs only, size-capped, decoded and scaled to PNGs in a 128 MB disk cache
(browser::images); a placeholder shows until they arrive. Vanilla's mark is a
grass block drawn from the player's prepared game files. Modrinth's logo isn't
shown: its branding is all rights reserved without written permission.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Ludovic dropped CurseForge for now (its API needs a paid key). The Mods page shows
"CurseForge (coming soon)" beside the Modrinth switch, disabled and inert;
browser::sources() is Modrinth only, so no search or install goes to CurseForge,
and `--source curseforge` says it is coming soon. Text that asked for a key or
an environment variable now says CurseForge support is coming soon. Importing a
CurseForge zip or instance still works; jars only CurseForge serves are listed
as "CurseForge downloads coming soon".
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Rust 1.99's clippy denies chunks_exact with a constant size
(chunks_exact_to_as_chunks), so `cargo clippy --all-targets -D warnings`
failed on this test. Same check, written as as_chunks::<4>().
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The game and its integrated server map items and blocks to network ids through
the registries report (registries.json), which the content root left vanilla:
a mod's item in a slot, given or picked up, reached the client as air. The
content root now rewrites registries.json too, appending the new items and
blocks after vanilla's ids in plan order.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The Mods page gets a Mods tab: Fabric, NeoForge, Forge and Quilt mods for 26.3
(1.21 and later with Show everything). Each result's jar is downloaded (allowed
CDNs, checksum), checked with horizon-port-pack (Works, Partly, Doesn't run, with
what ports and what doesn't) and the answer cached per file. Install copies the
carry into horizon-mods/<id>/ and turns it on in horizon-mods.txt, with required
mods that port; libraries with nothing to port are listed as not ported; id or
namespace clashes are refused. Remove deletes it and turns it off.
The Modpacks page becomes the Mods page's Modpacks tab: imported packs first,
then the search. CLI: mods search --kind mod, mods check <project>.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A recipe whose result is one of the mod's own items that the carry didn't declare
(its block had no vanilla template) loaded and crafted air. Vanilla refuses a
recipe for an unknown item; the carry now moves such recipes to outside/ and lists
them (recipes_unregistered), and check says "N recipes for left-out blocks".
Found by the mod browser's end-to-end run on Macaw's Fences.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
`horizon-port-pack check <jar> <out>` carries a single jar into a folder content
mod and reads its class files (constant pool, superclass, declared methods) for
what the carry can't take: block entities, menus, entities, blocks and items with
behaviour of their own, recipe types, worldgen code, Mixins, game events. The
verdict and the lists of what ports and what is skipped go to check.json, for the
launcher's mod browser to show before install.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Every animation frame's mip chain goes to video memory once, at the
sprite's own size, as vanilla keeps its animated sprites' frames on the
GPU. A tick that changes sprites writes one 20-byte job per changed
sprite, and a compute dispatch per mip level repeats the chain into the
array's larger layers and mixes interpolated frames
(animate_sprite_interpolate), instead of blending on the CPU at the
array's size and copying each sprite level by level.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
spawn_category_for_position looked up the picked entry's type twice per
attempt and asked mobsAt again at the position it had just picked the entry
from. The type is now resolved when the entry is picked, and the recheck is
skipped at that position (the list there holds the entry). The nearest player
distance only feeds comparisons with 24 and the despawn distances (128 at
most), so an attempt near its start measures only the players that can be
within 128 blocks of it; farther ones read as infinitely far, which every check
answers the same. The water-ambient biome tag is read only for that category.
Same lists, same random draws.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Review fixes on top of the particle serializers:
- PARTICLES (17) is List<ParticleOptions>: a VarInt count, then each
particle. Reading it as one particle only worked for empty lists, so a
living entity with effect particles (LivingEntity.DATA_EFFECT_PARTICLES)
misread everything after it. Value::Particles holds the list.
- Item particles use ItemStackTemplate.STREAM_CODEC (item, count, patch),
not the optional slot codec.
- Vibration keeps its position source (block position or entity id and
y offset) so it writes back the bytes it read; geyser data keeps its
Int exactly instead of going through f32.
- Enchantment names come from the server's synced registry, not the
builtin report (enchantment is data-driven).
- Armor stands: 26.3's DEFAULT_DEPTH_CLEAR_VALUE is 0.0 (bytecode
dconst_0), so a marker's empty box falls back to a range of 4 and a
visible marker stand draws; restore that. Holograms stay hidden through
the invisible flag.
Test: particles_round_trip covers every data shape, a PARTICLES list and
the entry after it.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged. The fast channel is now "Alpha" (Ludovic, 2026-10-22); old "nightly" / beta_builds settings still read as Alpha. verify passes; launcher tests 96+14+1, release tools' Python tests 27, website build OK.
The whole of Horizon is in beta, so the channel that installs every build as
soon as it is built is called Alpha; Stable keeps its name. The launcher's
Settings, the update-channel command, the MCP tool, the release notes footer
and the docs say Alpha. update_channel is saved as "alpha"; "nightly" and the
old beta_builds switch still read as Alpha.
Co-Authored-By: Claude Sonnet 5.5 <[email protected]>
The Mods page (launcher 0.17) replaces "a mod browser is planned"; the Horizon
shader format is no longer called planned; Nightly-only things say Nightly.
The "I'm playing" ping has no server (DEFAULT_ENDPOINT is None), so the FAQ and
the privacy page say it is off instead of sent. The channel switch says it
needs the launcher release after 0.17. The README says what deploys the site.
Co-Authored-By: Claude Sonnet 5.5 <[email protected]>
With a JDK installed, the launcher passed the preparation the home of the
JDK horizon-prepare finds (javac/../..) as JAVA_HOME and on the PATH.
Found through the PATH, that is often a stub: on macOS /usr/bin/javac
gives JAVA_HOME=/usr, and `JAVA_HOME=/usr /usr/bin/javac -version` never
returns (the stub calls itself), so the preparer subprocess hung on a Mac
with Temurin installed. When horizon-prepare finds a JDK by itself, run
the preparation as before, in an unchanged environment; only Minecraft's
own Java (the official launcher's, or one downloaded from Mojang) is
passed on.
Also skip the Mojang download on x86-64 Linux without glibc's loader
(NixOS without nix-ld, musl distributions): Mojang's Java can't start
there, so say so instead of downloading 115 MB first. And keep each
runtime home once (dedup only dropped neighbours).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Serializer 17 (particles) and 16 hit a catch-all that returned Ok(None),
which stopped EntityData::read from walking the rest of the metadata list.
On armor stands that carry a level-particle effect slot, slots 2 and 3
(CustomName text and visible flag) were silently discarded, so serverside
holograms (Gamster's social-media panels and leaderboards) rendered as
nothing.
Implement both serializers as value-preserving decodes: a ParticleData
enum covering every 26.3 particle shape (state varints for ids 1/2/36/121/125,
colours, dust, dust_color_transition, effect/instant, scalar, shriek, trail,
item, vibration and geyser variants) backed by a static minecraft:particle_type
id table, with write_value arms for round-trips.
Co-Authored-By: cairn-code (GLM-5.3-Flash Free) <[email protected]>
to_stack kept only damage, dyed colour, map id and a head profile, so a server item's custom name, lore, enchantments and custom data never reached the inventory tooltip. Build the vanilla `components` compound from the decoded patch and run it through the same decoder saved item NBT uses, then show lore lines between the name and the details.
Co-Authored-By: cairn-code (DeepSeek V4.1 Flash) <[email protected]>
On a Mac without a JDK, preparing failed with "Java 25 is needed ...
/usr/bin/javac -version printed no version: Unable to locate a Java
Runtime": /usr/bin/javac is macOS's placeholder. Mojang's Java runtimes
are whole JDKs, javac included (java-runtime-epsilon is 25.0.1 on every
platform), and the official launcher has already downloaded one for
anyone who played 26.3.
- horizon-prepare: java::find also tries Minecraft's runtimes
(<game dir>/runtime/<component>/<platform>/<component>, jre.bundle on
macOS) and the launcher's pick (Options::java_home), and names the
macOS placeholder for what it is.
- launcher: when no JDK is found, downloads the runtime the version JSON
names (javaVersion.component) from Mojang into its own game folder,
every file checked against its sha1, executables and links as the
official launcher lays them out. A version's own preparer binary or
Python script gets it through JAVA_HOME and the PATH.
Checked for real on Linux: the download is 147 files, its javac 25.0.1
compiles and runs a class with --release 25.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01AxbYUJt4ShJFCdyLNCxF21
A marker armor stand has an empty box, so vanilla's shouldRenderAtSqrDistance computes a zero range and the model never draws; Horizon substituted a 4-block range, which turned the stacked marker stands a lobby uses for holograms into a visible tower. Follow vanilla and let the zero edge cull the model. Server NPCs and mobs also carry CustomName, CustomNameVisible and the shared invisible flag in entity data; Horizon only read those for armor stands. Apply them to every mob, keep the styled component for the name tag, and hide the body when the flag is set.
Co-Authored-By: cairn-code (DeepSeek V4.1 Flash) <[email protected]>
A server lobby draws its signs as walls of framed maps and puts player heads on its NPCs. Horizon dropped every map_item_data packet, so framed maps stayed blank, and the item component decoder rejected attack_range (30) and profile (72), which made the whole equipment and container packets that carried them undecodable. Decode both components, carry a profile into the item data the head model reads, and apply server map patches to the map texture. The map pool now holds a wall of maps so the atlas does not re-upload every map every frame.
Co-Authored-By: cairn-code (DeepSeek V4.1 Flash) <[email protected]>
Holograms are invisible armor stands carrying a CustomName. Armor stands parsed neither CustomName nor CustomNameVisible, and name tags came only from mobs, so a hologram drew nothing. Read both fields from NBT and entity data, keep them across a save, and put armor stands in the name tag pass at their bounding-box height even when their model is skipped.
Co-Authored-By: cairn-code (DeepSeek V4.1 Flash) <[email protected]>
The tab list resolved skins only for listed players and never handed them to the level's player models, so every player and NPC in a server world kept the default Steve or Alex texture. Resolve every profile's textures each frame and give each player entity its skin, including entities that are never in the tab list. Player heads read the minecraft:profile component with the insecure lookup vanilla uses, so unsigned textures from offline servers draw too.
Co-Authored-By: cairn-code (DeepSeek V4.1 Flash) <[email protected]>
winit's macOS set_cursor_position re-associates the mouse with the
pointer (CGAssociateMouseAndMouseCursorPosition(true)), so warping to
the window centre right after the Locked grab released it again and the
pointer moved freely in first person. Warp before grabbing on macOS;
Wayland still warps after, since it moves a pointer only while locked.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_0188LkUPntcZdfMF3wAqR2tK
TerrainRegion::new refused a region whose chunks bounding box passed
2^20 columns, where it took any set of distinct chunks before: the
bulk_section parity fixture puts a chunk at x = 4194304 beside the
origin and failed with InvalidChunks. Keep the grid for the usual small
box and fall back to a map past it; both refuse a position twice. The
lookup now subtracts in i64, so no coordinate can wrap into the grid.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Vanilla 26.3 Mth.floor(double) is (int) Math.floor(v), so a value below
i32::MIN floors to i32::MIN. The truncate-and-step version wrapped to
i32::MAX there (block_collision_scan parity failed at the chunk of
x = -2147483648.8: Rust 134217727, vanilla -134217728). Step down with
saturating_sub; the fast path is unchanged, and the test now checks
the ends of the i32 range, infinities and f64::MIN/MAX against
f64::floor as i32.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A server's player (a LAN guest, and since the server thread the single
player too) always respawned in the Overworld. Its respawn anchor was
dropped when its action ended, and its save kept only an Overworld point.
- The guest keeps a respawn point outside the Overworld (respawn_elsewhere,
saved as the player's `respawn`) and has it in place while it acts, so an
anchor it uses is its own.
- guest_respawn respawns it in that point's level, waits there for the
chunks, then resolves the point through its block
(findRespawnAndUseSpawnBlock): beside the anchor, using a charge, or the
world spawn and block.minecraft.spawn.not_valid when the block is gone.
- `spawnpoint` in another level than the Overworld sets a forced point there
(SetSpawnCommand) instead of Overworld coordinates, and a forced point that
is no anchor holds where a player fits (isPossibleToRespawnInThis), in the
window's own game too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
handle_server_portal ignored End gateways, so on a server (dedicated, or
single player's server thread) no player could use one: the in-frame game's
gateway code moved only its own player.
A gateway that is not cooling down now takes a server player at once
(EndGatewayBlock.entityInside, triggerCooldown, setAsInsidePortal), within
the End, keeping its motion and look (Relative.DELTA and ROTATION). An exact
exit is reached straight away; otherwise the player waits at the destination
while it loads, and the exit is found or built there (the in-frame
resolve_gateway, shared through gateway_job).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
On a server (dedicated, or single player's server thread) each dimension is
its own level, but a command ran in whichever level was in place. So a guest
in the Nether running `execute in minecraft:overworld run setblock ...` got
"That position is not loaded", and `execute at <a player in another level>`
used the wrong level's blocks.
with_source now puts the source's level in place for the command
(CommandSourceStack.getLevel), as as_target already does for a player in
another level, and puts the old one back after. `execute at` takes the
entity's level (withLevel(entity.level())), and a player's level is the
guest's own (target_level) even while another level is in place.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
On a server the guest is held where it entered the Nether until the
chunks around the destination are lit, then placed in its portal. The
check read the guest's position 40 ticks after the dimension changed,
which on a busy machine was still the held spot, and then stepped the
guest out of a portal that wasn't there.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A new `levels` check (--server-thread) runs the host and a guest in different
dimensions: execute in ... tp, mobs ticking in two levels at once with unique
ids, rejoin and save/reload with the guest in the Nether, death there, the
host through a Nether portal alone, the guest through an End portal and the
exit portal alone. All of it already works on the server thread.
docs/server-levels.md says how the server's levels map onto vanilla's, what
the dedicated server and the server thread share, the cost of an extra level
and what's left. docs/server.md no longer says players can't change dimension.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
tick_chunk_random asks for the lowest randomly ticking section from a height
up, instead of asking about every section: a chunk costs one question per
section that ticks, not 24. PalettedContainer::has tests whole words for the
few matching palette indices (a field-wise zero test) instead of reading
every entry, so a section read again after an edit costs a few hundred word
operations.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15): fmt, clippy, workspace + launcher tests, release build and every affected parity harness pass. main merged into the branch with the same resolution as the tested tree; verify passes on that head.
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15): fmt, clippy, workspace + launcher tests, release build and every affected parity harness pass. main merged into the branch with the same resolution as the tested tree; verify passes on that head.
ExtraCover::new checked only x0/x1 against the i32 range, so two areas
near i32::MAX or i32::MIN in z built a grid whose row arithmetic
(cz + dz) overflows: a panic in debug builds, wrong cells in release.
Check z the same way, and test both ends.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
🤖 Agent: merged after integration wave 2 on the desktop (integ/w2 @ 38307d15 = main c16b41ab + this wave): fmt, clippy -D warnings, workspace + launcher + telemetry-server tests, release build and every affected parity harness pass; the 3 harness failures and the panel test passed on rerun (known flakes).
The daily promotion moves to 20:00 UTC (21:00 in Paris in winter time):
PROMOTE_AT and promote.yml's cron (20:07, retries until 21:37), with the docs
that give the time. A launcher left open on Nightly looks for new releases
every 30 minutes; Stable keeps 4 hours. The launcher README's Settings
screenshot shows the Updates choice.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The launcher's "Beta builds" switch becomes an Updates choice of two channels,
stored as "update_channel": "stable" | "nightly" (an old "beta_builds": true
reads as nightly). The CLI (update-channel) and MCP (update_channel) show and
set it, and list each release's channel as stable or nightly.
Nightly ships constantly: the release train waits 5 minutes for a merge wave to
settle (20 at most) instead of 15 (an hour). Stable moves once a day:
promote.yml runs at 19:00 UTC (retries until 20:37) and promotes the newest beta
that had been out 6 hours by 19:00, so late runs and retries pick the same one.
The stalled-promotion issue waits 30 hours. Release notes, docs, the website's
install and FAQ pages say Stable and Nightly.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Every font definition of the resources (and packs) is now loaded, not
just font/default.json, and a text component's "font" field picks one:
"minecraft:alt" draws the enchanting table's Galactic alphabet,
"minecraft:illageralt" the illager runes, a pack's own font its glyphs,
and a font that doesn't exist draws every character as the missing
glyph (FontManager.getFontSetRaw).
- font.rs: one FontSet per font id with its providers in order (space,
bitmap, unihex, ttf) and their filters ("uniform", "jp"). Bitmap
sheets outside the atlas (another namespace or folder) are copied into
the glyph region when first drawn.
- TrueTypeGlyphProvider: "ttf" providers (file under font/, size,
oversample, shift, skip; CFF fonts refused) rasterized with ab_glyph
into the glyph region as rrrr coverage (text_intensity), metrics as
GlyphBitmap (bearings over the oversampling, whole-pixel advances).
- text::Style gains the font id; it is parsed from JSON and SNBT
components, pack descriptions and scoreboard number formats, inherited
by children and written back.
- Obfuscated text picks its random glyph in the style's font.
Fixes#336
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Leaving the End through the exit portal for the first time now rolls the
end poem, the credits and the post-credits text (WinScreen), as does a
server's WIN_GAME event; Options -> Credits & Attribution -> Credits rolls
the credits alone. The texts come from the resource root
(texts/end.txt, credits.json, postcredits.txt), so packs can replace them.
- win_screen.rs: vanilla's line building (PLAYERNAME, the obfuscated
names from LegacyRandom(8124371), wrapping at 256), scroll speeds and
keys (Space, Ctrl, Up, Escape), the last line pinned mid-screen, the
logo, the vignette and the backgrounds.
- HUD: Blend::Vignette (RenderPipelines.VIGNETTE) and Blend::EndPortal
(the End portal effect as a GUI fill, the poem's background).
- Obfuscated text draws a random glyph of the same width each frame.
- seenCredits is read and saved; the player stays out of the world while
the poem rolls and arrives in the Overworld after it.
- horizon-ctl `credits [poem] [SCROLL]` and `options credits`.
Fixes#338
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Vanilla random-ticks a section while its count of randomly ticking blocks
and fluids is above zero (LevelChunkSection.isRandomlyTicking), and asks as
it reaches each section. Horizon ticked every section whose palette listed
such a state, even after the last one was gone, and decided before the tick's
scheduled ticks, which drew positions vanilla doesn't and shifted its
position sequence.
RandomTickSections now reads each section's blocks (PalettedContainer::has),
again only for the sections that changed, and tick_chunk_random asks for
each section when it reaches it; the game reads a chunk's sections again only
after a block changed. New harness random_tick_sections compares the
sections and positions with vanilla's real section counts.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
🤖 Agent: merged after integration wave 1 on the desktop (main e730dc14 + #1108#1114#1131#1157#1158#1162 = ae482a72): fmt, clippy -D warnings, workspace tests (one port-race flake in horizon-server panel, passes on rerun), release build, tools/parity.py --changed origin/main 299/302 then the 3 rerun: save_roundtrip pass, world_reopen pass x2 (one early-count flake), dedicated_server = known flake also failing on main (fixed by #1183/#1184).
🤖 Agent: merged after integration wave 1 on the desktop (main e730dc14 + #1108#1114#1131#1157#1158#1162 = ae482a72): fmt, clippy -D warnings, workspace tests (one port-race flake in horizon-server panel, passes on rerun), release build, tools/parity.py --changed origin/main 299/302 then the 3 rerun: save_roundtrip pass, world_reopen pass x2 (one early-count flake), dedicated_server = known flake also failing on main (fixed by #1183/#1184).
🤖 Agent: merged after integration wave 1 on the desktop (main e730dc14 + #1108#1114#1131#1157#1158#1162 = ae482a72): fmt, clippy -D warnings, workspace tests (one port-race flake in horizon-server panel, passes on rerun), release build, tools/parity.py --changed origin/main 299/302 then the 3 rerun: save_roundtrip pass, world_reopen pass x2 (one early-count flake), dedicated_server = known flake also failing on main (fixed by #1183/#1184).
🤖 Agent: merged after integration wave 1 on the desktop (main e730dc14 + #1108#1114#1131#1157#1158#1162 = ae482a72): fmt, clippy -D warnings, workspace tests (one port-race flake in horizon-server panel, passes on rerun), release build, tools/parity.py --changed origin/main 299/302 then the 3 rerun: save_roundtrip pass, world_reopen pass x2 (one early-count flake), dedicated_server = known flake also failing on main (fixed by #1183/#1184).
🤖 Agent: merged after integration wave 1 on the desktop (main e730dc14 + #1108#1114#1131#1157#1158#1162 = ae482a72): fmt, clippy -D warnings, workspace tests (one port-race flake in horizon-server panel, passes on rerun), release build, tools/parity.py --changed origin/main 299/302 then the 3 rerun: save_roundtrip pass, world_reopen pass x2 (one early-count flake), dedicated_server = known flake also failing on main (fixed by #1183/#1184).
🤖 Agent: merged after integration wave 1 on the desktop (main e730dc14 + #1108#1114#1131#1157#1158#1162 = ae482a72): fmt, clippy -D warnings, workspace tests (one port-race flake in horizon-server panel, passes on rerun), release build, tools/parity.py --changed origin/main 299/302 then the 3 rerun: save_roundtrip pass, world_reopen pass x2 (one early-count flake), dedicated_server = known flake also failing on main (fixed by #1183/#1184).
LocatorBar: when the server tracks waypoints for the player, the bar
above the hotbar shows a dot per waypoint within 60 degrees of the
view (farthest first), coloured from the waypoint's colour or its id's
hash, with the sprite its style picks by distance
(assets/*/waypoint_style/*.json, WaypointStyleManager) and the blinking
up or down arrow when the waypoint is above or below the screen. The
background is nine-sliced. Hud.nextContextualInfoState picks between
the jump bar, the experience bar (100 ticks after an experience change)
and the locator bar.
Waypoints come from ClientboundTrackedWaypointPacket on remote servers
and, on a served world, from the host's own connection in
lan/waypoints.rs; entity waypoints in view follow the entity's eyes.
Fixes#339
Co-Authored-By: Claude Opus 5.5 <[email protected]>
EquipmentLayerRenderer's trim layer: the pattern's
textures/trims/entity/<layer>/<asset>.png recoloured from its base
palette (palette metadata section) to the material's palette, or to the
equipment asset's trim_overrides palette (iron armour with an iron trim
uses trim/iron_darker), as PalettedTextureManager does. Each texture is
made once, on first use, into one of 32 atlas trim slots (least recently
drawn reused), and drawn over the armour layer on the player, on the
guests a LAN host sees and on armour stands.
Fixes#337
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Pack-wide numbers after the modpacks-12 PRs (#1175, #1176, #1177, #1191):
structures that start, processors carried and dropped per mod, content,
load time and the play-through, and the gaps ranked by how much a player
notices them.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
TextureMetadataSection's blur and clamp choose how vanilla samples a
standalone texture (ReloadableTexture.setSampler). The atlas now keeps
each texture's flags and the screen-effect quads (vignette, pumpkin blur,
powder snow outline, underwater, equippable camera overlays) sample
their sprite rectangle linearly and with clamp or repeat as those flags
say, instead of always nearest and repeat.
Fixes#335
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A content pack's first start rewrote 19 per-state tables one after another
and reopened each zip mod's archive for every file it copied (its whole
index each time). Every start then decoded the atlases' mod textures and
parsed the mods' block models one file at a time on the main thread: with
ATM10's 42 data mods that was about 70 % of the sampled startup.
- mod_content: the tables are rewritten on up to three threads, largest
first; a zip mod's archive is opened once for all its files
(horizon-mods' Content::reader).
- hud::atlas_sprites and ParsedModels::load list their files first, then
decode and parse them on every core (par::par_map, results in the input
order). The sprites come out in the order the one-by-one loop gave, and
the content root is byte-identical.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The last two workstation menus guests couldn't use, both now on the
shared menu engine, which single player runs too:
- Lectern: LecternMenu on the engine (Kind::Lectern): the page buttons
(1, 2, 100+) and Take Book (3, mayBuild) as ContainerButtonClick, the
page as data slot 0 clamped to the book, the book in its one slot
(removeItem, setItem a no-op, max 1). The host opens minecraft:lectern
for a guest, applies the menu to the block entity (powered pulse, page
turn level event 1043, has_book), resends changes and closes it on
stillValid. Guests drive it from the book screen.
- Cartography table: CartographyTableMenu on the engine (Kind::Cartography,
setupResultSlot, onTake, onCraftedBy's new map ids through a Maps env),
replacing the old station code; guests get the same through
guest_menu_op.
- LAN item stacks carry map_post_processing and writable/written book
contents (they were blank for guests); map_post_processing is no longer
saved (ItemStack.CODEC never saves it).
- A guest who enchants gets the enchant_item stat.
- Click parity: VanillaMenuOracle and menu_fixture gain both menus.
- lan_e2e: new lectern and cartography checks; enchanting checks the
stat; server_count/server_reply ignore the guest's own operator
broadcast echo on --server-thread; the grindstone check counts levels
as well as points.
Fixes#868
Co-Authored-By: Claude Opus 5.5 <[email protected]>
On Linux an unprivileged thread can't raise itself, so #1179's render
thread hint does nothing there. The dedicated server already lowered its
helper threads instead (world generation, lighting, connection writers:
10 nice levels, or the idle class). The client now does the same, so the
world-generation and meshing workers (two pools with single player's
server thread) no longer hold the render thread back while terrain loads.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Trees of neighbouring chunks overlap on chunk corners, and the one
decorated first wins. With several worker threads vanilla decorates in a
different order each run: seed -2080727421395798566 spawned at Y 64, 66
or 72 over 10 runs. The harness retried vanilla up to 3 times and passed
if any run matched, which still failed when none of them gave 72.
With -Dmax.bg.threads=1 (Util.maxAllowedExecutorThreads) vanilla
decorates in its scheduler's order, the one Horizon follows: that seed
gave 72 in 10 of 10 runs, as Horizon does. The other seeds that varied
are fixed too (10 of 10 each: -8171837717659076940 at 0 67 0,
-5296387317982839690 at -16 73 16, -211473008665507967 at 0 66 112).
The retries go, and vanilla's one spawn is compared exactly.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Three races made the two sides differ with nothing wrong in Horizon:
- Steps ended on a fixed 1.5 s sleep. On a slow run a chat or its
broadcast (vanilla queues the chat, then the broadcast as a later
task) landed in the next step. Each step now also ends on a barrier:
every connected client sends `random value 1..2` twice, and waits for
the answer, which only the sender gets and which is left out of the
transcript.
- The client closed every socket at once, and the server logs
"<name> lost connection" in whatever order its network threads see
the closed connections. The players still connected now leave one at
a time: each waits until it has been told the previous one left
(PlayerList.remove broadcasts that after logging). The harness waits
for Carol's line before stopping the server.
- A player the server kicks (Bob, Mallory) is logged as
"lost connection: <reason>", or as "Disconnected" when vanilla's
network thread closes the connection before disconnect(reason) runs.
The reason sent is still compared in the transcript, so the log line
no longer compares it for kicked players.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Natural spawning was a quarter of a 40-player server thread. Every spawn
attempt cloned the position's whole spawn list (NaturalSpawner.mobsAt) to
check the picked entry is still in it, and every group cloned it again to
pick from; mobs_at now returns the biome's or the structure's list as it is
stored (the fortress list lives in a static). The nearest player's distance,
asked on every attempt, is now the minimum of the distances instead of a
search that kept the player too. Same lists, same draws, same distances.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Single player on the server thread publishes inside the integrated
server, so self.lan alone missed it; and the fps block runs only when a
frame is drawn.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Reduced starts each far tile level at half the distance Full does (about a
quarter of the far quads beyond the full-detail edge, whose finest tiles stay).
Full stays the default everywhere and no preset changes it. On integrated
GPUs other than Apple's, the setting's description suggests Reduced.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
floor_i32 is now Java's Mth.floor (truncate, step down for negative
fractions): f64::floor is a libm call on baseline x86-64, made for every
Perlin sample. Perlin's gradient table is stored as floats (every
component is -1, 0 or 1, exact either way), so the corner dot products
skip three conversions each. About 10% fewer instructions generating
terrain; values unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01M3pNrSHA2ifNrH7SAMHgSW
The vanilla protocol client moved on after a fixed time per step. On a
busy machine, either side's server could be slower than that: the
summoned villager was not seen yet when the trade step ran, the furnace
had not smelted the ingot when it was taken, a snowball was not thrown
yet, or the pearl had not landed. Each mode failed now and then on one
side only.
Those steps now keep the same minimum time and then also wait, for up to
30 s, until it happened: the villager is tracked, the furnace's result
slot has the ingot (plus 1 s), the snowball entity was added (plus 1 s),
the pearl teleport and the health update arrived. The respawn line also
leaves out block updates: a lava fall keeps flowing while the game runs,
and how many updates arrive before the respawn depends on timing.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A list sorted only by distance switched vertex buffer every few sections:
about 1200 indirect multi-draws a frame at render distance 32, most of a
few entries, which Mesa's Intel driver unrolls on the CPU. Grouping each
band's sections by page keeps the list front to back within the band, with
one call per page and band.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The client asks for its render thread what it already asks for the
server thread (vanilla's MinecraftServer.spin): user-interactive QoS on
macOS, above-normal priority on Windows, on machines with more than four
processors. While terrain loads at 32 chunks, the worker pools (two with
single player's server thread), the chunk encoders and other programs
outnumber the cores. A frame that gave its core up in queue.submit then
waited 10 to 200 ms to get one back: the long frames of #678.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
tick_lan described every tracked entity afresh each tick: a new String for
its type, new vectors for its entity data, equipment and attributes, all
freed again at the end of the tick. The states now live in the LAN host
between ticks and each tick writes over the last tick's in place
(lan::entity::States), so those allocations are kept. The states, and so
every packet built from them, are the same as before.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The game sends an anonymous ping once a minute while it runs (a random
session ID kept only in memory, the version, the platform, whether a LAN
or friends world is open). The telemetry server counts sessions seen in
the last 2 minutes, in memory only, with its own per-address rate limit
and a session cap, and serves rounded counts at GET /v1/live. The
launcher shows "N playing now" on the Play page when the count is
reachable, and tells players what the ping holds on the first-run
question, in Settings, and once on the Play page for players who
answered before. docs/telemetry.md and the privacy page say the same.
Like the rest of telemetry, nothing is sent until HORIZON_TELEMETRY_URL
is set or DEFAULT_ENDPOINT gets a server.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Three races made the comparison of the servers' lines about the player
fail at random:
- a login's lost connection is logged at the server's next tick, and the
next login could start before it: each login now ends once the server
logged its lost connection;
- vanilla's authenticator disconnects a player, then logs why, while the
server thread may already log the lost connection: each login's lines
are compared thread by thread;
- a player the server kicks closes the connection once it reads why; when
that comes first vanilla logs the lost connection as "Disconnected"
(seen once in 12 desktop runs), so a login's kick reason stands for both.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
ThrownEnderpearl.onHit teleports the owner to oldPosition() (where the
pearl started its last tick) with ServerPlayer.teleport, which sends the
position packet whatever the distance. A guest's turn sent a teleport only
when the player moved more than 0.01 blocks, and the pearl took the hit
point: a pearl thrown at one's feet often lands within that, and the
dedicated_server harness saw "teleported 0 time(s)" where vanilla's server
sends 1 (about one run in four on the desktop).
The pearl now moves the player to its previous position and marks the
teleport (PlayerMotion.teleported), which the end of the guest's turn
sends even when the player stays put.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A client of a server (single player's server thread, LAN, multiplayer)
handled every received packet in the frame it arrived. Crossing a chunk
border at 32 chunks brings 25 to 60 columns at once, 0.1 ms each to
decode on an M4 Pro: a 3 to 8 ms frame each time. Once a frame has spent
1 ms on chunk packets, the events from the next chunk packet on wait for
the next frame, in order; events that have waited a tick are all handled
at once, so a slow machine's backlog never holds packets longer than
that. The status's pending_chunks counts the waiting events.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Formations' PiecedStructure starts like its SimpleStructure, then lays out
the element's jigsaw children with vanilla's jigsaw. The game's jigsaw
assembler now exposes its child layout (add_children) for a start piece
placed by another structure type; horizon:formations reads horizon:pieced
(max_depth, max_distance_from_center), which carry writes for
formations:pieced.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
ctov:modular_compat runs another processor list only when the mod it names
is loaded: carry writes that list in when the mod is ported along and takes
the hook out otherwise (142 copies in ATM10, all for Nature's Spirit).
Repurposed Structures' mineshaft_sky_view_processor (71 lists) becomes the
game's horizon:mineshaft_sky_view: nothing at or above the world's
OCEAN_FLOOR_WG height, except plain rails over a block that can occlude,
waterlogged where the world holds water. Placement contexts gain
can_occlude.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A guest's use of a crafter now opens CrafterMenu (minecraft:crafter_3x3)
over the block entity. The host runs its clicks on the menu engine's
verified Kind::CrafterBlock (a disabled slot refuses items, filling one
enables it), sends the disabled slots and triggered as data slots
(CrafterBlockEntity.containerData), and handles
ServerboundContainerSlotStateChangedPacket (new in horizon-protocol) as
CrafterBlockEntity.setSlotState. The guest shows the server's crafter,
predicts its clicks and sends a slot toggle before the click as
CrafterScreen.slotClicked does.
Part of #868.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
BeaconMenu now runs on horizon_client::menu (Kind::Beacon): the payment
slot (beacon_payment_items, one item), quickMoveStack and updateEffects
with BeaconBlockEntity.validateEffects, verified against the vanilla class
by test_menu_click_parity with a new SET_BEACON op (valid and invalid
powers for each pyramid level).
ServerboundSetBeaconPacket is now in the protocol. The LAN host opens
minecraft:beacon for guests with the block entity's levels and powers as
data slots (kept current while the menu is open), applies a guest's
powers to the block entity (power_select while the beam is up) and drops
the payment when the menu closes, as BeaconMenu.removed does. Guests show
the server's levels and powers and their Done button sends the packet.
Single player's beacon clicks go through the engine too.
Part of #868.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
EnchantmentMenu now runs on horizon_client::menu: the item slot (one
item) and the lapis slot, its quickMoveStack, slotsChanged clearing the
offers without an enchantable item, and clickMenuButton's checks against
the costs, lapis and levels, verified against the vanilla class by
test_menu_click_parity (with its ten data slots and the offer buttons;
the oracle's player now has a profile for the invalid-button log line).
The offers and the enchanting itself, which need a level, stay with the
caller.
Single player routes the table's clicks, drags and double clicks through
the engine. The LAN host opens minecraft:enchantment for guests with
their own enchantment seed (saved with their player data), sends the
costs, seed and clues as data slots, and on ContainerButtonClick
enchants the item, takes the guest's level and lapis and gives a new
seed. Guests show the server's offers and send the button the client
menu accepts.
Part of #868.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
AnvilMenu and SmithingMenu (ItemCombinerMenu) now run on
horizon_client::menu: their slot rules, createResult (the anvil's cost
data slot, repair count and rename, the smithing table's hasRecipeError),
mayPickup with the player's levels, the results' onTake and
quickMoveStack, verified against the vanilla classes by
test_menu_click_parity with repairs, merges, renames (a new RENAME op,
AnvilMenu.setItemName) and the player's levels. The harness found the
anvil treating only enchanted books as books (vanilla: anything with
stored enchantments), now fixed, and the grindstone's and anvil's result
slots now count for a double click's pick-all as in vanilla.
Single player routes both screens' clicks, drags and double clicks
through the engine; the anvil's wear and level events 1030/1029, and the
smithing table's 1044, play with vanilla's random pitch.
The LAN host opens minecraft:anvil and minecraft:smithing for guests,
handles ServerboundRenameItemPacket (new in horizon-protocol), charges the
guest's levels and wears the anvil; guests send the rename box as
AnvilScreen.onNameChanged does and show the server's cost. The control
socket's `type` now reaches the anvil's rename box.
Part of #868.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
GrindstoneMenu now runs on horizon_client::menu (Kind::Grindstone): its
input filter, computeResult, the result's onTake (both inputs cleared,
the experience sum handed to the caller) and quickMoveStack, verified
against the vanilla class by test_menu_click_parity with enchanted,
damaged and cursed tools and books.
Single player routes the grindstone's clicks, drags and double clicks
through the engine and rolls the orbs and level event 1042 (now playing
block.grindstone.use) where the result is taken. The LAN host opens
minecraft:grindstone for guests on the same engine and awards the orbs at
the block; guests read enchantments, stored enchantments, repair cost and
custom names from the server's stacks, and the host now sends repair
cost.
Part of #868.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
StonecutterMenu joins horizon_client::menu (Kind::Stonecutter): the input's
recipe list (slotsChanged / setupRecipeList), selectedRecipeIndex as its data
slot, clickMenuButton (an invalid index is taken too), the result's onTake
(one input used up, awardUsedRecipes, the take sound) and quickMoveStack with
the leftover of a result move dropped. test_menu_click_parity compares it
with vanilla's StonecutterMenu, recipe buttons included.
Single player's stonecutter screen now clicks through the engine (drags and
double clicks included), and its recipe grid calls clickMenuButton as
StonecutterScreen does. A LAN guest's stonecutter opens minecraft:stonecutter
on the host's engine, as the loom does (#868).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A guest's use of a loom opens LoomMenu (minecraft:loom) for the guest. The
host runs the guest's menu on the shared menu engine (horizon_client::menu,
Kind::Loom, as single player does): clicks, the pattern buttons
(ServerboundContainerButtonClickPacket), the content and data slot sync after
each, the take-result sound at the loom, stillValid (the block gone or out of
reach) and clearContainer when it closes or the guest leaves. The guest opens
the loom screen for the server's menu, predicts its clicks and buttons on the
engine and sends them, runs the menu's client-side slotsChanged on the
server's slots and reads the selected pattern from the data slot.
The framework (game/lan_stations.rs, game/remote_stations.rs) takes the other
workstations one by one (#868).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Carry maps mostructures:generic to vanilla's jigsaw with horizon:start_checks
(terrain_radius 2, terrain_range = heightRange) and mostructures:type to a
random spread with the config's default spacing and separation and its sets
to avoid as horizon:exclusion_zone. The game's exclusion zone gains
"their_zones": false, which tests the avoided sets by spread and frequency
alone, as ModStructurePlacement does.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
#1007's white overlay on Hyprland is #1075's washed-out swapchain: a
Wayland compositor with colour management makes Mesa list Rgba16Float
first, and the first non-sRGB format was taken. Reproduced on a real
GPU (Mesa 26.2 ANV, Intel HD 620) under headless weston 16: the old
choice washes the frame out, main's Bgra8Unorm doesn't. This test keeps
that list, as wgpu reports it, choosing an 8-bit format.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
TerrainRegion indexes its chunks through a grid over their bounding box.
OreFeature asks the world whether any column under its blob reaches the
floor heightmap in one call, and tracks tested positions in a bitset.
BiomeTemperature no longer preallocates its map and queue. Same output.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01M3pNrSHA2ifNrH7SAMHgSW
A version prepared while the sound copy failed (Mojang's republished 26.3
asset index, #1076, fixed in #1091), or one that reused such a version's
resources when installed, stays silent and showed as plain Prepared. The
Play page now says it is prepared without its sounds and makes Prepare
again the visible fix; versions --installed and the MCP list say so too.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
The client asks for its render thread what it already asks for the
server thread (vanilla's MinecraftServer.spin): user-interactive QoS on
macOS, above-normal priority on Windows, on machines with more than four
processors. While terrain loads at 32 chunks, the worker pools (two with
single player's server thread), the chunk encoders and other programs
outnumber the cores. A frame that gave its core up in queue.submit then
waited 10 to 200 ms to get one back: the long frames of #678.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
HORIZON_SPIKES spans keep the thread's CPU time beside their wall time,
and the slow-frame line gives the render thread's CPU time in the frame.
A span that mostly waited (for the GPU, a lock or a core) prints both
(`submit 30.2/0.1`), so a frame that waited reads apart from one that
worked. Found the cause of #678's long frames: the render thread losing
its core in queue.submit.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Releasing a Riptide trident now starts LivingEntity.startAutoSpinAttack(20,
8.0): for 20 ticks the player takes the SPIN_ATTACK pose (0.6x0.6, eye
height 0.4), hits the first living entity its swept box touches for the
spin's damage and bounces back (velocity x -0.2), and stops on a horizontal
collision. The player model spins (setupRotations: rotX(-90 - xRot),
rotY(age * -75)), SpinAttackEffectLayer draws the riptide boxes, and the
first-person hand uses ItemInHandRenderer's spin transform. LAN guests see
the host's spin through living flag bit 4.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
A new Horizon Extras option, Lighting > Dynamic lights (`dynamicLights`,
off by default): a torch, lantern or other light-emitting item held by a
player or mob, dropped as an item, or any entity on fire lights the blocks,
entities and hand around it, as LambDynamicLights and OptiFine do.
Client-only and at render time: the light engine, chunk meshes and saves
never see it. Each frame the game picks the 16 sources nearest the camera
(dynamic_light::select) and passes them in Globals; terrain.wgsl and
entity.wgsl raise each vertex's block light to `level - distance` when that
is brighter. Off, the count in globals.time.z is 0, nothing is gathered and
the shaders' loop runs zero times, so no pipeline is rebuilt either way.
An item glows like its block's default state (torch 14, lantern and
glowstone 15, lit redstone torch 7); a lava bucket 15; burning entities 15.
Fixes#814
Co-Authored-By: Claude Opus 5.5 <[email protected]>
On single player's server thread the broken log's drop and its pickup
both come from the server, a few ticks after `break` answers: the
pickup could land before the scenario looked for the drop, which it
then never found (StopIteration). Step 3 now waits up to 20 ticks for
the logs in the inventory or the drop to walk to.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Step 2 clones a stack in the creative tab and closes the screen. The
clone lived only on the client: ItemPickerMenu's carried stack is the
inventoryMenu's, and AbstractContainerMenu.removed clears the carried
stack only for a ServerPlayer, so vanilla's client still holds it when
the survival inventory opens. With a stack in hand the middle button's
release sends CLONE (AbstractContainerScreen.mouseReleased); the server
finds the carried stack it was told about wrong and sends the empty one
(broadcastChanges). On single player's server thread that answer takes
a round trip, as in vanilla, so step 3 now waits up to 20 ticks for it.
The in-frame world puts the clone away when the menu closes.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
From b9f3821f (fix/server-thread-default): the host window draws what
it was sent on the server thread, and agent_control_e2e waits two ticks
where the server answers a click.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
On a client of a server (single player's server thread, a LAN guest) the
placed block comes with the server's answer: `place` waits up to a second
for it before reporting no change. agent_control_e2e tries the next free
spot when the tree's trunk above the broken log hides a spot's faces.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
(cherry picked from commit 9821bb7ef5)
- The command harness (HORIZON_COMMAND_SCRIPT) drives the world itself,
without the event loop, so it does not start the server thread.
- `status` lists the minecarts a client of a server was sent.
- save_roundtrip: single player's server thread finishes the chunks its
player tracks, as vanilla's integrated server does, including some that
vanilla's server (view distance 4) left before their terrain or light.
Those had no blocks, light or heightmaps to compare yet.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
(cherry picked from commit 2818c89f24)
Single player's world now runs on the integrated server thread unless
asked otherwise, as vanilla's does, so a world tick no longer holds up
a frame. `--no-server-thread` or `HORIZON_SERVER_THREAD=0` keeps the
world in the frame as an escape hatch; `--server-thread` stays.
lan_e2e.py sets the host's mode explicitly either way.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
(cherry picked from commit 4ff23b686d)
PerlinNoise::add_volume worked out each Y coordinate's cell, offset and
fade once per X/Z column. Every column walks the same Y coordinates, so
they are now computed once per volume. The arithmetic is unchanged, so
the output is bit for bit the same.
ClimateIndex kept its tree as nodes with a Vec of child ids each. The
built tree is now renumbered breadth first into flat arrays, so a
branch's children are one contiguous run. The search visits the same
nodes in the same order, so ties resolve exactly as before.
terrain_bench (seed 12345, 8x8 chunks, 1 thread, medians of 3 runs):
terrain 14.5 -> 12.4 ms/chunk, biomes 1.20 -> 1.06 ms/chunk.
Checked against vanilla 26.3 in-process: 2304 chunks on seeds 12345, -1
and 987654321987 identical, and the biome-source and biome-generation
parity suites (2.1M lines) match.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01M3pNrSHA2ifNrH7SAMHgSW
On a server every player is an extra area, and streaming asks of every
loaded and wanted column, several times a tick (unloading, dropping
protos, finishing, decoration, the generation scan), whether one of
those squares reaches it: a scan of every player's square per question.
The squares are now one grid of distances, made again when a player
changes chunk, and in_extra reads it. Differential test against the scan
over 60 random layouts, every margin.
At 50 players, 6081 columns, 3 questions each: 0.45 ms -> 0.04 ms, plus
0.02 ms to make the grid.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
HopperBlockEntity.suckInItems scanned every item entity for each hopper
that ticked without a container above. During the block-entity pass the
game now keeps the set of block columns item boxes reach; a hopper whose
column is not in it has nothing to suck, and the scan (same order, same
item taken) runs only for the rest. Items do not move within the pass:
removals leave a superset, and additions (pushed at the end with a new
id) or removing the last item rebuild the set. It is dropped at the end
of the pass.
500 hoppers, 200 items: 1.36 -> 0.05 ms per tick
2000 hoppers, 500 items: 13.52 -> 0.26 ms per tick
2000 hoppers, 2000 items: 60.6 -> 3.3 ms per tick
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
The Shader Pack setting cycled through what shaderpacks/ held, so with the
folder empty (or the packs in the launcher's version folder, as brabus_ had
them) it silently stayed Off. The row now opens a Shader Packs page: a
sidebar section in the Horizon menus, Video Settings > Shader Packs... in
the Classic ones. It shows the game home's shaderpacks/ path and an Open
Folder button (made if missing), says when there is no pack, and lists one
switch per pack. A pack that fails to load says why on its row. The list is
read again every second while the page shows.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
GL fills the channels a fragment shader leaves out; wgpu refuses a pipeline
whose output has fewer channels than its target. Steadfast's final pass
writes `out vec3` and failed to build in the game.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
Preprocessor (GLSL and shaders.properties), as JCPP does:
- a #define replaces an earlier macro (Bloop, Steadfast: DefineRedefined);
- in #if, a float counts as its integer part (Steadfast's
shaders.properties float token), and a macro always defined empty counts
as 0 (Iris's MC_GL_VENDOR_OTHER, Bloop's TAA_USE_YCOCG);
- an #if may run to the end of the input (Bloop's taa.fsh);
- #error "..." lexes in skipped blocks (Photon);
- shaders.properties keeps its empty lines and hides backslashes from the
preprocessor (IRIS_PASSTHROUGHBACKSLASH), so a continued value no longer
swallows the next directive or line (Photon, Bliss, Steadfast's custom
uniforms; Complementary's isCold uniform is now complete).
Legaliser: % through overloads (naga won't take ivec2 % int: Noble),
sqrt of a negative constant left to the GPU (Bloop's NaN), constants naga
can't fold (?:, mix: Bliss) become globals set in main, and 16/8-bit types
(u16vec2 & co., Noble on NVIDIA) become their 32-bit ones.
horizon-client: real_packs_load runs the game's own load over the packs in
HORIZON_TEST_SHADERPACKS (the compatibility table's source).
Co-Authored-By: Claude Opus 5.5 <[email protected]>
sculk_sync_index collected every loaded chunk key into a set and diffed
it against the known chunks every ten ticks, thousands of keys on a busy
server. It now remembers the chunk map's membership revision and skips
the diff while it is unchanged; a replaced chunk keeps its key and was
not rescanned before either.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
tick_allays scanned every block entity of the level for each allay,
every tick, to find a playing jukebox within 10 blocks. The playing
jukeboxes are now collected once, at the first allay (no allay starts
or stops one).
piglin_pickups looked up the piglin_loved tag and hashed every item
entity's id once per piglin. Each item's flag is now computed once and
kept until an item is added (always pushed at the end under a new id)
or removed.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
GameLevel's isUnobstructed, noCollision (living shulkers) and the
spawners' entity count each scanned every mob, computing each 1.5 KB
mob's box. A spawn pass holds the game immutably, so after eight
queries it builds a grid of the mobs' boxes once and reads candidates
from it; the first queries still scan, as the build costs a few scans.
100 box tests, 2000 mobs: 3.48 -> 0.31 ms (grid build included)
Also gives the mob_ai_probe example the new others_index field.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
pois_in_range (AcquirePoi and the job, bed and meeting point searches)
asked, for every candidate POI in range, how many tickets it had, and
each answer scanned every other mob for villagers holding it: POIs times
mobs per search. The others' tickets are now counted into a map at the
first candidate and read from it; this villager's own memories are
still checked per POI, as before.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
FollowParentGoal (every baby, every tick), the pufferfish's puff check,
the enderman's block placing and the iron golem's DefendVillage scanned
every mob. They now read candidates from the position grid around the
area they test (one block wider where they test positions, as a box
whose bottom face holds the position meets it), in snapshot order, and
test exactly as before.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
On a server every ticked mob is bound to its nearest player, and each tick
the bindings of mobs that are gone were dropped by scanning every mob for
each binding. Mob is 1.5 KB, so the scan misses the cache on every mob:
with 1000 loaded mobs that was 0.44 ms a tick, with 2000 1.86 ms; a set of
the live ids makes it 0.02 and 0.04 ms (stand-in structs of Mob's size).
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
tick_creature_entities called mob_world_base for every mob. That rebuilt
the player view (its allies set scans every mob UUID) and, with any
scoreboard team, an owned copy of every mob's team, each a clone of the
team's player set: cubic in the mobs on a team. The base is now built
once and cloned per mob, with the teams borrowed behind an Arc.
100 mobs, half on a team: 37.7 ms -> 0.06 ms per tick
500 mobs, half on a team: 4297 ms -> 0.31 ms per tick
MobWorld::seen looked mobs up by scanning the snapshot; the mob ticks
now pass an id index (first slot wins, as the scan did): 2000 lookups
over 2000 mobs 1.09 -> 0.01 ms. The nearest-living-entity sensors, the
look-at goal, the ravager's roar and MobWorld::nearby read candidates
from the tick's position grid instead of every mob, in the same order:
2000 mobs 0.39 -> 0.11 ms per tick.
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
Pushes between mobs wait in a list until their target ticks; each mob's
tick scanned the whole list, and take_mob_pushes removed each push from
the events in place, shifting the rest. In a pen where every mob pushes
its neighbours both grow with the square of the crowd per mob.
The pending pushes stay a list while short (64) and become one list per
mob past that, each in push order, so every velocity adds up the same
values in the same order. take_mob_pushes splits the events in one pass,
keeping the order of the others. A test checks velocities bit for bit
and the remaining events against the old code.
30 mobs pushing each other: 0.11 -> 0.04 ms per tick
100 mobs pushing each other: 1.73 -> 0.37 ms
200 mobs pushing each other: 9.07 -> 2.61 ms
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
Every lit furnace, smoker and blast furnace called Recipes::cook each
tick, which scanned all cooking recipes (string compares and hash
lookups) and the furnace then cloned the recipe found, its ingredient's
item set included. Recipes now keeps, built at the first lookup, the
first recipe of each kind for each item, which is what the scan
returned; the furnace borrows it.
1000 lookups over 320 recipes: 2.07 -> 0.08 ms
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
Two costs that grow with players on the server thread:
- Hub::update_view rebuilt a player's chunk queue on every chunk crossing
by scanning the queue (a Vec) once for each chunk of the view. It now
checks a set. Same queue, same order (differential test against the old
scan). 111 -> 16 us per crossing at view 10, 6.5 ms -> 0.13 ms at 32.
- ChunkTickets rebuilt a hash map of every ticketed chunk whenever any
player changed chunk, and every block entity, mob and ticking chunk
hashed into it. Levels are now a grid over the tickets' bounding box,
with the map kept for tickets spread too far apart. At 50 players:
build 0.23 -> 0.08 ms, lookups about 2x faster (differential test over
200 random ticket layouts).
Co-Authored-By: Claude Opus 5.5 (1M context) <[email protected]>
Claude-Session: https://claude.ai/code/session_01UxXpf7d4RgpespL2G2UsGE
.github/release-line.txt = 0.2: the release train's next runs publish v0.2.0 (a beta, as always)
and [email protected] at once, "the first release of 0.2" (docs/versioning.md). Merge it only after
the bridges (the last dated game release, made stable, and launcher-v0.17.0) are out: the go-day
steps are in docs/versioning.md.
Also, because they say 0.2 is out: the site's FAQ explains the numbers, Install says an old launcher
updates twice, hosting.md gives the one command for a server that says it is up to date, and the
launcher's Cargo.toml says 0.2.0 so builds from source (packages) are on the 0.2 line.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
release.yml builds every game release with HORIZON_VERSION, the tag without its v
(docs/versioning.md), as launcher.yml already does for the launcher: no version-bump commits.
- horizon_core::release::{RELEASE, VERSION}: the built-in version, or Cargo.toml's for a build from
source. (horizon_core::version stays Minecraft's, generated by gen_version_consts.py.)
- horizon-server and horizon-prepare --version, the server's /api/v1/status, horizon-ctl's MCP
serverInfo and the "Horizon/<version>" user agents use it; they all said 0.1.0.
- The client's crash reports and telemetry say v0.2.3 from the built-in version wherever the game
is installed; releases from before 0.2 (no version built in) still read the launcher's folder name.
- The launcher's benchmark telemetry used Cargo.toml's version instead of the release's.
- release.yml checks that horizon-prepare and horizon-server --version say the release's version.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
2026-10-22 01:20:13 +02:00
1123 changed files with 61116 additions and 17760 deletions
`tools/check_public_tree.py` rejects them; `git config core.hooksPath .githooks` runs it on commit.
- **Keep your runs away from real player data.** Any game or launcher you start gets fresh `mktemp -d`
folders in `HORIZON_HOME` (the game's data), `HORIZON_LAUNCHER_HOME`, `XDG_DATA_HOME` and
`XDG_CONFIG_HOME`, plus `HORIZON_LAUNCHER_NO_SELF_UPDATE=1` and `HORIZON_NO_AUDIO=1`. Never write to
`~/.local/share/{applications,horizon,horizon-launcher}` or `~/.local/bin`.
- **Mark what an agent posts.** A PR body starts with
`> 🤖 Opened by Ludovic's AI agent (Claude), not by Ludovic himself.` (issues: "Filed by", reviews:
"Review by"); every comment starts with `🤖 Agent:`. Never write as Ludovic.
- **Attribution.** Commits end with `Co-Authored-By: Claude <model> <[email protected]>`; PR bodies end
with `🤖 Generated with [Claude Code](https://claude.com/claude-code)`.
- Everything players run is Rust. Python/Java only for dev tools and harness drivers in `tools/`.
- Horizon-only features follow [docs/design-system.md](docs/design-system.md) and are off by default where
they would change the vanilla look.
- Small PRs (< ~500 lines). Labels: `feature` (players get something new), `skip-release` (they see no
change), none for a fix. Agents never add `release-now`: the maintainer decides it. Every merge to main ships
([docs/release-flow.md](docs/release-flow.md)).
- Contributors first: `gh pr list --state open --search <area>` before you start; never duplicate theirs.
## Crates
Game workspace (root `Cargo.toml`); `launcher/` and `crates/horizon-telemetry-server` are separate workspaces.
| Crate | Owns |
|---|---|
| `horizon-client` | The game: a lib (`src/lib.rs`, ~190 flat modules) and the bin (`src/main.rs` → `game.rs` + `game/`), plus `horizon-ctl` (`src/bin/horizon-ctl/`). Also the dedicated server (`--dedicated`) |
| `horizon-world` | Vanilla world data with no gameplay loop: NBT, regions, chunks, `datafix/`, `worldgen/`, `level/` (server-side block behaviour, ticks, redstone), entity physics, items, effects, food, damage |
| `horizon-friends`, `horizon-relay` | Play with Friends: invite codes, encrypted tunnel, the relay ([docs/play-with-friends.md](docs/play-with-friends.md)) |
@@ -46,7 +46,7 @@ The [documentation](https://playhorizon.dev/docs/) covers each of these.
- **The game:** singleplayer with vanilla's world generation, mobs, villagers, raids, redstone, every command, the HUD and the recipe book, on saves in vanilla's format. Worlds from 1.8 to 26.2 are upgraded as vanilla does. [Status table](#project-status).
- **Horizon's own extras:** new menus (vanilla's are one setting away), graphics presets, distant terrain up to 1024 chunks, controller support, optional zoom, minimap and more, all off by default where they would change the vanilla look. See [Settings](https://playhorizon.dev/docs/settings).
- **Resource packs, shader packs and data packs:** resource packs work, Java shader packs run their full-screen effects (experimental), and data packs load functions, predicates and item modifiers.
- **Playing with others:** invite friends from the pause menu with a code, Open to LAN, and joining offline-mode servers. Horizon cannot sign in to a Microsoft account yet, so it cannot join servers that check accounts. See [Multiplayer](https://playhorizon.dev/docs/multiplayer).
- **Playing with others:** invite friends from the pause menu's Friends button with a code, Open to LAN, and joining offline-mode servers. Horizon cannot sign in to a Microsoft account yet, so it cannot join servers that check accounts. See [Multiplayer](https://playhorizon.dev/docs/multiplayer).
- **Mods:** Horizon has its own mod system (Rust compiled to WebAssembly). Java mod jars do not run. See [Mods and data packs](https://playhorizon.dev/docs/mods).
Screenshots, captured in the Rust client itself with vanilla assets and no mods or shaders:
@@ -56,8 +56,8 @@ Screenshots, captured in the Rust client itself with vanilla assets and no mods
<td colspan="2"><img src="docs/media/sunset.webp" alt="Sunset over a savanna river valley and a plains village"></td>
</tr>
<tr>
<td width="50%"><img src="docs/media/menu-title.webp" alt="Horizon's title screen over a veiled snowy landscape"></td>
<td width="50%"><img src="docs/media/settings.webp" alt="The Video settings in their Simple view: a preset, render distance, field of view, framerate, VSync and GUI scale, each with what it costs"></td>
<td width="50%"><img src="docs/media/survival.webp" alt="Survival at dusk: a zombie walks up, with the hotbar, health, hunger and experience"></td>
<td width="50%"><img src="docs/media/settings.webp" alt="Vanilla's Video Settings screen with Horizon's own buttons at the top: Horizon Graphics, Advanced Graphics and Shader Packs"></td>
</tr>
<tr>
<td width="50%"><img src="docs/media/village.webp" alt="A plains village among savanna hills, with farms, villagers and an iron golem"></td>
Loaded 100 of 1123 files, more files were not shown because too many files have changed in this diff.
Show more
Reference in new issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Ludovic Marie