Priyanshu KumarandClaude Sonnet 5.5 6633b90497 doc: the select ADR's counts come from ternary_census.py, and the noreturn guard's cost and depth are stated with their measurements ...

Co-Authored-By: Claude Sonnet 5.5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01456qmqCGpCWJkrg6JUZNNM

2026-10-24 03:12:11 +05:30

..

drafts

Roadmap W1: stable op identity, written-lane result widths, provenance facts, and the shell/visual surface (#67)

2026-10-06 15:18:20 +05:30

reviews

ROADMAP: restructured by owner around the end-to-end audit, the vendored core first

2026-10-14 21:28:13 +05:30

adr-access-syntax.md

D: CI grades one pipeline; docs drop pdim, pdih and the legacy stages

2026-10-10 17:21:49 +05:30

adr-aggregates.md

B7: r2engine's integration tests are one binary, each file a module

2026-10-15 10:00:24 +05:30

adr-byte-relation.md

Roadmap W1: stable op identity, written-lane result widths, provenance facts, and the shell/visual surface (#67)

2026-10-06 15:18:20 +05:30

adr-decompiler-rewrite.md

D65: a stub is the stated import's only where its transfer reaches that import

2026-10-23 17:21:44 +05:30

adr-fixpoint.md

D: CI grades one pipeline; docs drop pdim, pdih and the legacy stages

2026-10-10 17:21:49 +05:30

adr-floating-point.md

D3: a float truncated to an integer is a cast inside C's range and a residual outside it

2026-10-09 14:33:45 +05:30

adr-frame-model.md

FS: the loads of one slot are one copy class, so a guard on any of them bounds the rest

2026-10-22 17:24:28 +05:30

adr-language-profile.md

P7: a value returned in two registers is one result

2026-10-14 04:57:09 +05:30

adr-listing-spelling.md

r2sleigh-lift: alias boundary tests for ubfx, sbfiz, lsl, cmn, tst and mvn; the ADR's duplicate row gone

2026-10-17 01:44:21 +05:30

adr-machine-profile.md

Roadmap W1: stable op identity, written-lane result widths, provenance facts, and the shell/visual surface (#67)

2026-10-06 15:18:20 +05:30

adr-measurement-discipline.md

ROADMAP: DecBench 2026-10-23 corrects the ranking, and D67 ranks the next fixes by measured ceiling

2026-10-23 17:23:30 +05:30

adr-memory-model.md

ROADMAP: restructured by owner around the end-to-end audit, the vendored core first

2026-10-14 21:28:13 +05:30

adr-next-pass.md

LX3: the reaching module's comment is two lines; the region solve's argument is in the ADR

2026-10-12 18:28:36 +05:30

adr-one-ir.md

Review: stack #165's open review threads

2026-10-14 04:57:30 +05:30

adr-partition-first.md

Roadmap W1: stable op identity, written-lane result widths, provenance facts, and the shell/visual surface (#67)

2026-10-06 15:18:20 +05:30

adr-provenance.md

C2: a function's return and argument types are Facts; return_unproven goes

2026-10-13 21:05:01 +05:30

adr-query-database.md

Q: a number that folds to zero is no formed address, whatever the image maps there

2026-10-22 17:24:28 +05:30

adr-radare2-function-walk.md

ROADMAP, ADR: the derived callee-recursion cap measured with no payoff; #26995 to #26997 open

2026-10-15 22:37:56 +05:30

adr-register-identity.md

Roadmap W1: stable op identity, written-lane result widths, provenance facts, and the shell/visual surface (#67)

2026-10-06 15:18:20 +05:30

adr-renderer-printer.md

Roadmap: hostile input, a performance gate and a decompiler rewrite enter the path

2026-10-08 12:10:07 +05:30

adr-resolved-bodies.md

P5: a call's target set is the table captured for that read, not what the slots hold

2026-10-22 17:24:28 +05:30

adr-scripting-surface.md

Docs: the command gap, the debugger and the scripting surface, reviewed

2026-10-09 19:35:30 +05:30

adr-select-respelling.md

doc: the select ADR's counts come from ternary_census.py, and the noreturn guard's cost and depth are stated with their measurements

2026-10-24 03:12:11 +05:30

adr-semantic-preservation-kernel.md

Roadmap W1: stable op identity, written-lane result widths, provenance facts, and the shell/visual surface (#67)

2026-10-06 15:18:20 +05:30

adr-stable-identity.md

D42: the fact collection returns the exhausted pass, so PreparedFunctionFacts and RecoveredInterface hold no flag

2026-10-17 16:45:19 +05:30

adr-structure-dominator-tree.md

doc: the select ADR's counts come from ternary_census.py, and the noreturn guard's cost and depth are stated with their measurements

2026-10-24 03:12:11 +05:30

adr-verified-layers.md

ADR D44: the oracle breadth ratchet tracks constructors as well as mnemonics

2026-10-22 12:44:00 +05:30

adr-written-lanes.md

Roadmap W1: stable op identity, written-lane result widths, provenance facts, and the shell/visual surface (#67)

2026-10-06 15:18:20 +05:30

command-gap.md

Merge pull request #146 from radareorg/docs-143-review

2026-10-16 12:26:51 +05:30

debugger.md

Docs: the command gap's rule says why ?e is skipped; parallel indexing counts its serial checkpoint pass

2026-10-11 22:24:51 +05:30

decompiler.md

D: CI grades one pipeline; docs drop pdim, pdih and the legacy stages

2026-10-10 17:21:49 +05:30

engine-vision.md

D: CI grades one pipeline; docs drop pdim, pdih and the legacy stages

2026-10-10 17:21:49 +05:30

r2il.md

Roadmap W1: stable op identity, written-lane result widths, provenance facts, and the shell/visual surface (#67)

2026-10-06 15:18:20 +05:30

rewrite_quality_gates.md

GT T3: every Dylint aims at a live path, a test keeps it so, CI ratchets the count

2026-10-15 10:00:24 +05:30

roadmap-history.md

ADR, ROADMAP: D58 lands; D62 to D64 carry what is left

2026-10-23 16:23:28 +05:30

ssa.md

ADR Count-driven tail: cursors as values, AArch64's offset, nine of twelve shapes.c builds measured; ROADMAP W14's rest

2026-10-17 17:52:53 +05:30

testing.md

r2s tests: a pipe a test closes stays open in a spawn in flight, so spawns and the two closing tests take one lock

2026-10-24 03:12:09 +05:30

types.md

Review #220: a pointer base wakes its readers only when the class stops being a pointer

2026-10-14 10:29:07 +05:30