fix(resolve): assigning a variable from a function call erased the type it was declared with by joyful-ii-V-I · Pull Request #278 · redhat-et/ripwire

joyful-ii-V-I · 2026-09-17T11:44:45Z

Defect

Rule 2 records a C++ assignment from a call as the variable's type, so that x = Foo() types x. A constructor call
and a function call are one grammar node, so the record held the callee's last name whatever it was:
PHI = PHINode::Create( … ) recorded Create, s = GetDBOptionsFromMap( … ) recorded GetDBOptionsFromMap, and
t = llvm::cast<T>( y ) recorded cast. Rule 2's flat per-function table tombstones a variable whose records disagree.
So the non-type erased the declared PHINode *PHI / Status s, and PHI->addIncoming( … ) / s.ok() fell to the
name ladder. On the fixture and on both corpora that ladder usually declines. The same record:

tombstoned the field use-site index's table (--uses=Target.count lost its owner pin);
made a member assigned from a call (cur = ns::cast<Target>( y )) read as a local, so Rule 2b refused the member's
declared type.

graph.h's varType note said such a type "simply never produces a type::method hit". That holds for the narrow but not
for the tombstone. (The comment above ctorNameNode also claimed a class-name filter in buildGraph that did not exist.)

Change

ingest (ingest_binds.h): the assignment arm marks its record with assignedTypeOf →
RawBind / Binding::isFromAssignment.
buildGraph (resolve.h assignmentNamesNoClass): drops such a record unless a class of that name exists. It uses the
class-name set Rule 2c already built, which is now built first (classNameSet). The record is dropped from:
- Rule 2's varType table;
- collectFieldUseSites' table;
- the local-name set that Rules 2b/2c and the external veto read.
Declarations are unchanged. A declaration's callee-read type (auto t = makeFoo()) is still recorded. That
declaration exists with an unrecorded type, and its conflict with a sibling declaration keeps one block's type off
the other block's calls (arm 48).
Format: the bind record gains one u8.
- kCacheVersion 22 → 23; kMinBindRecordBytes 26 → 27; cachefuzzcheck's blob walker reads the new byte.
- kParserVer 99 → 104 with both quality.h mirrors. 100–103 are declared by lanes queued ahead (train 2b: 102,
  fix(ingest): a receiver typed with template arguments recorded no type, or the wrong class #276: 103); the landing train assigns the number.
- test/qschemetrip.hash re-pinned, with a RE-PIN LOG entry.
graph.h stays line-neutral above rankGraphTeleport (showcase seed).

Measured

Method: --pin-census --no-cache; C rows joined on (caller id, callee, line); main fe28fd49 vs this branch.

corpus	sites changing target	bound calls	newly bound	retargeted	lost
rocksdb `0e2801ac3`	1,871	+1,562	1,544	327	0
llvm-project `4d5358b1d`	4,046	+2,984	2,859	1,187	0

On rocksdb, 1,494 of the 1,871 are Status::ok.

Blinded grading. A seeded sample of 60 (seed 20260917; 25 rocksdb, 35 llvm-project) was graded against source by six
independent readers. Each saw the two answers as A and B, with the order randomised per site; the key was kept outside
the grading directory.

51 better: 39 NONE→RIGHT, 5 WRONG→RIGHT, 4 PARTIAL→RIGHT, 2 NONE→PARTIAL, 1 WRONG→PARTIAL.
6 same: 4 WRONG→WRONG, 2 RIGHT→RIGHT.
3 worse. All three are resolver floors that the erased type had been hiding, not wrong recovered types:
- rocksdb Iterator* narrowing onto the memtable's same-named Iterator classes (the floor fix(resolve): a member call through a typed parameter was pinned to the caller's own class #248 records);
- two llvm-project calls where arity picked the wrong overload of the right class (getFirstInsertionPt, StringRef::find).
Wilson 95%: better 73.8–91.9%, worse 1.7–13.7%.

The local-name-set part on its own (this branch vs a build without that filter): 23 rocksdb and 132 llvm-project
sites. 15 graded: 10 better, 2 same, 3 worse. One worse site is the same-named-class floor (PHINode). Two show a floor
this unmasks but does not cause: Rule 2c reads a member named like a class (std::unique_ptr<ToolOutputFile> OutputFile; OutputFile->keep() → vfs OutputFile::keep). That is filed as a follow-up.

Built and rejected

Dropping a declaration's callee-read type too. It moves 89 more llvm-project sites (0 on rocksdb). 15 graded: 11
better, 2 same, 2 worse (both the same-named-class floor). Rejected on mechanism: it hands one block's written type to a
sibling block's auto t = ns::cast<Decoy>( y ), a precise edge to the wrong class (arm 48 is red on it).
Lifting fix(ingest): a receiver typed with template arguments recorded no type, or the wrong class #276's floor (narrowcheck arm 39d, the unqualified Vec<T>() constructor spelling). Tested on a scratch
composition of fix(ingest): a receiver typed with template arguments recorded no type, or the wrong class #276 50b97b7b + this change (never pushed). Reading that spelling moves 245 llvm-project sites: 170
lost, 0 gained (779 lost without this change); rocksdb 0. The remaining losses are declaration conflicts, e.g.
auto *LI = cast<LoadInst>( … ) beside another LI, which this change keeps on purpose. The floor stays.

Gates

test/narrowcheck.sh gains arms 44–51, with the fixture generated in-gate:

arm	checks
44	Rule 2 narrow through `t = ns::cast<Target>( y )`
45	the same through a plain function call
46	a member assigned from a call keeps Rule 2b
47a / 47b	controls: a class-named assignment still types and still conflicts
48	a declaration initialised by a call still tombstones a sibling declaration
49	`--uses=Target.count` pins
50	census mechanism is receiver-rule
51	cold == cold == warm

Red first:

main fe28fd49: 44, 45, 46 (no edge), 49 (owner_candidates="2"), 50 (NO-CENSUS-ROW); 47a/47b/48/51 green.
Variants of this change (env toggles in a scratch build, never committed):
- declarations dropped too → 48 red (pick@lib/target.h:1);
- local-name-set filter off → 46 red;
- new byte not persisted → 51 red;
- guard off → 44/45/46/49/50 red.

Results:

narrowcheck ALL PASS (46 rows).
qschemetripcheck, qextractionkeycheck, cacheidentitycheck, portablecachecheck, cachefuzzcheck: ALL PASS
(cachefuzz's ASan sweeps skip locally: no ASan tree, by instruction).
rocksdb map byte-identical across two cold runs and warm vs cold; xmllint --noout clean; the same for this repo's own map.
--quality-delta=$(git merge-base origin/main HEAD)..HEAD: gating="0" after 3 acks written through the binary
(--ack-only=buildFieldNarrowTables,classNameSet; the ledger diff is exactly three +ack rows). One minor verbosity
row is left.
Full suite, python3 test/pargates.py . ./build/ripwire -j 6 at 379d9fa: gates=639 pass=636 skip=3 fail=0. The skips are
editchecknotecheck, argvdiffcheck and g1freshcheck (no asan/ tree). The first run, on a binary built before the commit,
failed versioncheck alone (built_from=fe28fd49e+dirty): a true positive, fixed by rebuilding at HEAD, then the whole
suite was re-run.

Pins moved

kCacheVersion 22 → 23, kParserVer 99 → 104, and both quality.h mirrors; the train assigns the final number.
test/qschemetrip.hash 72f930b8 → f7ffb332.
.ripwire_quality_acks +3.

Train notes

test/narrowcheck.sh: union arms 26–38 (fix(resolve): a call through an interface pointer landed on unrelated nested classes of the same name #268), 39–43 (fix(ingest): a receiver typed with template arguments recorded no type, or the wrong class #276) and 44–51 (this).
src/ingest_binds.h: the ctorNameNode comment conflicts with fix(ingest): a receiver typed with template arguments recorded no type, or the wrong class #276's rewrite; keep fix(ingest): a receiver typed with template arguments recorded no type, or the wrong class #276's text and add this lane's
assignment sentence.
kParserVer / qschemetrip: re-derive on the merged tree.

🤖 Generated with Claude Code

Summary by CodeRabbit

Bug Fixes
- Fixed type resolution for variables assigned from functions, constructors, and factory calls.
- Prevented function names from being incorrectly treated as types when narrowing fields and local variables.
- Preserved valid type narrowing for known class-based calls and declaration initializers.
- Improved handling of conflicting local declarations and field-use analysis.
- Added coverage for assignment-based narrowing, deterministic results, and cache behavior.

…pe it was declared with Rule 2 records a C++ assignment from a call as the variable's type, so `x = Foo()` types `x`. A constructor call and a function call are one grammar node, so the record held the callee's last name whatever it was: `PHI = PHINode::Create( … )` recorded `Create`, `s = GetDBOptionsFromMap( … )` recorded `GetDBOptionsFromMap`, `t = llvm::cast<T>( y )` recorded `cast`. Rule 2's flat per-function table tombstones a variable whose records disagree, so the non-type erased the declared `PHINode *PHI` / `Status s`, and the call fell to the name ladder. The field use-site index (--uses=Owner.field) lost the same pin, and a MEMBER assigned from a call read as a local, so Rule 2b refused its declared type. graph.h's varType note said such a type "simply never produces a type::method hit"; that was true of the narrow and false of the tombstone. The fix: ingest marks an assignment's record (RawBind/Binding::isFromAssignment, ingest_binds.h assignedTypeOf), and buildGraph drops it unless a class of that name exists (resolve.h assignmentNamesNoClass, with the class-name set Rule 2c already built, now built first) — from Rule 2's varType table, from collectFieldUseSites' table and from the local-name set Rules 2b/2c and the external veto read. A declaration's callee-read type is kept: that declaration exists with an unrecorded type, and its conflict with a sibling declaration is what keeps one block's type off the other's calls. Measured, `--pin-census --no-cache`, C rows joined on (caller id, callee, line), main fe28fd4 vs this change: rocksdb 0e2801ac3 1,871 sites change target, bound +1,562 (1,544 newly bound, 327 retargeted, 0 lost; 1,494 are Status::ok); llvm-project 4d5358b1d 4,046 sites, bound +2,984 (2,859 newly bound, 1,187 retargeted, 0 lost). Seeded sample of 60 (seed 20260917; 25 rocksdb, 35 llvm), graded blind by independent readers with A/B randomised per site: 51 better, 6 same, 3 worse (39 NONE->RIGHT, 5 WRONG->RIGHT, 4 PARTIAL->RIGHT, 2 NONE->PARTIAL, 1 WRONG->PARTIAL; 4 WRONG->WRONG, 2 RIGHT->RIGHT). The three worse are resolver floors the tombstone had masked: rocksdb `Iterator*` narrowing onto memtable namesake `Iterator` classes, and two arity picks of the wrong overload of the right class. The local-name-set part alone moves 23 rocksdb and 132 llvm sites; 15 graded 10 better, 2 same, 3 worse — two of those expose Rule 2c reading a member named like a class (`std::unique_ptr<ToolOutputFile> OutputFile`) as that class, a floor this change unmasks and does not cause. Built and rejected: dropping a DECLARATION's callee-read type too. +89 llvm sites (0 rocksdb), 15 graded 11 better, 2 same, 2 worse, but it hands one block's written type to a sibling block's `auto t = ns::cast<Decoy>( y )` — arm 48. Considered for #276's floor (narrowcheck arm 39d, unqualified `Vec<T>()` constructors): a scratch composition of #276 50b97b7 + this change reading that spelling moves 245 llvm sites, 170 lost and 0 gained (779 lost before this change), all declaration conflicts like `auto *LI = cast<LoadInst>( … )` beside another `LI`. The floor stays. Format: the bind record gains one u8 (kCacheVersion 22 -> 23, kMinBindRecordBytes 26 -> 27, cachefuzzcheck's walker reads it); kParserVer 99 -> 104 with both quality.h mirrors (100-103 are declared by lanes queued ahead; the train assigns); test/qschemetrip.hash re-pinned with a RE-PIN LOG entry. graph.h stays line-neutral above rankGraphTeleport. Gate: test/narrowcheck.sh arms 44-51 (fixture generated in-gate). Red on main fe28fd4: 44, 45, 46, 49, 50 (no edge, owner_candidates="2", no census row). Red on variants built from this change: 48 when declarations are dropped too, 46 without the local-name-set filter, 51 when the new byte is not persisted. Green here, with arms 1-25. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

…sted class-name loop buildFieldNarrowTables gains the class-name set (api-surface 1 -> 2): its local-name set drops an assignment's callee-read type that no class is called, the rule Rule 2's varType table applies. classNameSet (duplication and new-clone-of-reused-helper against layout.h buildAggIndex, 93 tokens) is the Class/Struct/Interface loop buildGraph already carried inline for Rule 2c, hoisted so collectFieldUseSites shares it; layout.h includes graph.h, so resolve.h cannot reuse buildAggIndex, whose value type differs. Written through --quality-ack --ack-only=buildFieldNarrowTables, classNameSet; the ledger diff is exactly three +ack rows. The minor verbosity row (collectFieldUseSites 171 -> 173) is not gating and is left unacked. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

coderabbitai · 2026-09-17T11:45:00Z

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Advanced

Run ID: c1bbacb4-6c56-4d01-8075-437821d90c1f

📥 Commits

Reviewing files that changed from the base of the PR and between fe28fd4 and 379d9fa.

⛔ Files ignored due to path filters (1)

test/qschemetrip.hash is excluded by !test/*.hash

📒 Files selected for processing (12)

.ripwire_quality_acks
CHANGELOG.md
src/graph.h
src/ingest_binds.h
src/ingest_cache.h
src/ingest_model.h
src/model.h
src/quality.h
src/resolve.h
test/cachefuzzcheck.sh
test/narrowcheck.sh
test/qschemetripcheck.sh

Included review availability: Your plan provides up to 8 included reviews per hour; 6 remain after this review.

📝 Walkthrough

Walkthrough

Changes

The change preserves assignment-origin metadata for inferred type bindings. Graph analysis now ignores non-class assignment callees while retaining class-based narrowing. Cache formats and extraction identities are updated, with new narrowing, determinism, and cache tests.

Assignment type narrowing

Layer / File(s)	Summary
Binding provenance and cache format `src/ingest_binds.h`, `src/ingest_cache.h`, `src/ingest_model.h`, `src/model.h`, `src/quality.h`, `test/cachefuzzcheck.sh`	Assignment-derived type bindings carry `isFromAssignment` through ingestion, model emission, and cache serialization. Cache and parser versions increase, and cache walkers track the added byte.
Class-name filtering in graph analysis `src/resolve.h`, `src/graph.h`	Graph construction builds a shared class-name set and excludes non-class assignment callees from receiver, local, parameter, and field-use narrowing.
Validation and release support `test/narrowcheck.sh`, `test/qschemetripcheck.sh`, `CHANGELOG.md`, `.ripwire_quality_acks`	Tests cover assignment-derived narrowing, field-use indexing, deterministic output, and cache behavior. Release notes, re-pin logs, and quality acknowledgements document the change.

Priority: ⬇️ Low

Estimated code review effort: 3 (Moderate) | ~25 minutes

Change: Bug fix

Suggested reviewers: andriytyurnikov

Merge Risk: ⚪ Minimal · up to 379d9

The assignment-type narrowing change includes the necessary cache invalidation and validation coverage, with no current merge-blocking risk identified.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 45.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 20 functions across 10 files. (2 skipped:…	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and concisely describes the main resolver fix: preventing assignments from function calls from erasing a variable's declared type.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

Full details: Docstring Coverage

Explanation

Docstring coverage is 45.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 20 functions across 10 files. (2 skipped: 2 unsupported.)

Fix all pre-merge checks with AI

✨ Finishing Touches 💡 1

📝 Generate docstrings 💡

Create stacked PR
Commit on current branch

🧪 Generate unit tests (beta)

Create PR with unit tests
Commit unit tests in branch lane/ctor-inferred-type

_{Comment @coderabbitai help to get the list of available commands.}

joyful-ii-V-I · 2026-09-17T11:49:07Z

This PR lands through integration train 3 (with #268 and #276), built on train 1b (#277), after an independent review. I labelled it train-member and cancelled its full-matrix run: CI is the landing bottleneck right now, and the train PR runs the full matrix on the composed tree.

Contributor commits merge as-is. Resolutions: - src/model.h: RecvKind::JavaTypeCandidate is appended after train 2b's Lit* kinds, its note joins the enum's comment block, and kRecvKindCount names it as the last enumerator. - src/graph.h: javaNarrower is built with scopedRecvDecls like the train's Narrower (1b's constructor); the S6-C locality guard keeps !identityClaim and !isJsTsLitRecv and adds the JavaTypeCandidate exclusion. buildFieldNarrowTables merged without conflict and was read by hand against #278's change there: #278 skips only a C++ assignment's Type binding (isFromAssignment), and #235's localShadowSpans records only VarDecl bindings, so neither change reaches the other's rows. - kParserVer and kIngestParserVerMirror: 106 after #278's 105; the PR's two steps (declared 97 and 98 over main's 96) land as one, their notes kept as "PR step". The PR's 96 note moves into descending order. - test/qschemetripcheck.sh: the PR's two RE-PIN LOG entries, renumbered, under train 3's. test/qschemetrip.hash keeps the train's pin; re-derived once on the final tree. - test/callformcheck.sh header: both documented-absent edits (member templates, Java method references) kept. - test/regression.sh: loop union (+javamethodrefcheck = 633); README.md, docs/EVALS.md and present/deck5_ripwire_build.js regenerated by docs/gatecount_build.py. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

Contributor commits merge as-is. Resolutions: - src/graph.h / src/resolve.h: #278 hoisted buildGraph's class-name loop into resolve.h classNameSet(), shared by Rule 2c and its assignment guard; #267 widened that same loop to Ruby modules. The train keeps the classNameSet() call and the Ruby-module clause and its note move into classNameSet() (the guard was already a name-only, cross-language test). - kParserVer and kIngestParserVerMirror: 108 after #233's 107 (the PR declared 96 -> 97); the mirror gains a note; the PR's note says kCacheVersion stays #278's 23. - CHANGELOG: the PR's entry moves after the train's other entries. test/regression.sh: loop union (+rubyrecvnarrowcheck = 635); counts regenerated. test/qschemetrip.hash keeps the train's pin until the final re-derive. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

…he composed tree #278 removes the ASSIGNMENT half of the floor's stated reason (`Spec = cast<FunctionDecl>( F )` no longer records `cast`), so the numbers were re-measured on this tree: a variant binary that reads an unqualified template-id callee in ctorNameNode, `--pin-census --no-cache` against the train binary, joined on (caller id, callee, line). llvm-project 4d5358b1d: 463 sites move, 324 edges lost, 137 retargeted, 2 gained; rocksdb 0e2801ac3: none. Sampled against source, every loss is a DECLARATION conflict — a cast helper's name beside the declaration's written type (`const ConstantInt *CI = dyn_cast<ConstantInt>( V )`, `StructType *STyL = cast<StructType>( TyL )`) or beside a second declaration (`CallInst *I = cast<…>` and `auto *I = dyn_cast<Instruction>( V )`), and the four retargets sampled were all worse. The floor stays and arm 39d still passes; its reason is rewritten in ingest_binds.h ctorNameNode, narrowcheck's arm 39 header and #276's CHANGELOG entry, and #278's entry names the base of its own 245/170 measurement. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

…ons move up one Train 1b's fix push bumps kParserVer 102 -> 103 (a signed numeric literal receiver, `(-1).toFixed()`), so every train 3 assignment moves up one in merge order: #276 = 104, small-fixes = 105, #278 = 106, #235 = 107, #233 = 108, #267 = 109. - src/ingest_cache.h, src/quality.h: kParserVer and its mirror 109; every train 3 note renumbered; 1b's 103, which its push recorded only in the RE-PIN LOG, gains a note in both headers; kCacheVersion's note names #278's parser version 106. - test/qschemetripcheck.sh: the TRAIN 3 entry names 1b's fix push and the new numbers, the member entries are renumbered, and 1b's own entry follows them. - test/qschemetrip.hash: re-derived on this merged tree (neither side's pin hashed 109). - CHANGELOG.md: the train 3 entries' parser-version sentences and headings renumbered. Everything else in the fix push (MCP builders, --for terms_total, regex alerts, the impactpartition/nulbyte/timeout gate fixes) merged without conflict. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

joyful-ii-V-I · 2026-09-17T13:11:34Z

This PR is part of integration train 3, #281, and will show as merged when that train lands.

integration: train 3 — interface/template/assignment receiver types (#268 #276 #278), small fixes, Java method refs (#235), GDScript (#233), Ruby constant receivers (#267)

…a local, so Rule 2b refused its declared type Rule 2b's local-shadow veto refused the member whenever buildFieldNarrowTables' localNameSet held (method, name), and that set is built from every binding record. C++ assignments mint records: the assignment arm of bindsVisitNode records the callee's name as a Type binding for `x = Foo()` and `x = std::make_unique<T>( … )`, and the L3 capture records FnAssign for `x = other` and a clobber tombstone for `x = nullptr`. An assignment declares nothing, so a member assigned in a method and then called lost its narrow (llvm-project LVReader.cpp:175, `OutputFile->keep()` in LVSplitContext::open, no edge; now ToolOutputFile::keep). The set's value is now the evidence it holds (resolve.h localNameEvidence): any record sets kLocalNameBound, a VarDecl, ParamType or FnDecl record also sets kLocalNameDeclared, and Rule 2b vetoes on the second bit alone. Rule 2c and the Phase 5 external veto keep reading presence: they ask whether the name is a variable, which an assignment proves. The previous commit made every local-declaring shape emit a declaration record except the vexing-parse local, pinned as a floor. Measured, --pin-census --no-cache, joined on (caller id, callee, line) against the previous commit: rocksdb @ 0e2801ac3 118 retargeted (67 gained / 51 changed / 0 lost), llvm-project @ 4d5358b1d 683 (521 / 162 / 0), all decided by Rule 2b (instrumented scratch build). Vetoing evidence before: FnAssign or clobber 88 / 479, a Type naming a class 0 / 13, only Type records naming no class 30 / 191 (the population redhat-et#278's guard also releases). Seeded blinded sample of 40 graded against source: 35 better, 4 same, 1 worse (a namesake base class across namespaces in the base walk). Rejected variant, measured: Rule 2c reading declarations only moves 0 rocksdb and 7 llvm-project sites, graded in full: 6 worse, 1 same. Composition checked with lane/rule2-irbuilder-local (scratch merge, rerere off): 24 more llvm-project sites retarget, all better — members the old per-method veto refused because a vexing-parse local of the name lives in one nested block; no retargeted site on either corpus sits inside such a block. Gate: test/fieldnarrowcheck.sh arm v. v1 x5 and v3's Rule 2b pickup red on the previous commit; the 12 v2 controls red on a scratch build with the veto removed, 3 of them (reference-returning parameters, attributed declarator) red on this veto without the previous commit; the 3 v3 rows red on a scratch build whose Rule 2c ignores assignments. graph.h stays line-neutral above rankGraphTeleport. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

… 113, kCacheVersion 24 PR redhat-et#282 ad7a9d5 (signed off) on main a6868f7. The PR declared kParserVer 103 -> 104 and kCacheVersion 22 -> 23 over its base a5ce95e; main had spent 112 and redhat-et#278's 23, so integration/train-5 assigns 113 and 24 (the ref record grows viaArrow, kMinRefRecordBytes 39 -> 40), with both quality.h mirrors and a RE-PIN LOG entry. The qschemetrip hash is re-derived once on the final train tree. Resolutions (unions; nothing from either side dropped): - src/ingest_relations.h: main's type-alias capture (redhat-et#280) and redhat-et#282's stdSmartPointee sit side by side. - src/graph.h: the member-field loop skips an alias record (redhat-et#280) and a smart pointer's viaArrow pointee (redhat-et#282). - src/ingest_cache.h: redhat-et#282's readRef (viaArrow) with main's writeBind/readBind (isFromAssignment). Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

…nto train 5 — kParserVer 114 lane/rule2b-assignment-veto 626810b (signed off; redhat-et#282 + 8f16d3e + 626810b). The lane declared kParserVer 112 over redhat-et#282's 104; main had spent 112 and redhat-et#282 takes 113, so integration/train-5 assigns 114 (mirror moved; no record layout change, kCacheVersion stays 24). Resolutions: - src/resolve.h: main's type-alias bases, classNameSet and redhat-et#278's assignmentNamesNoClass beside the lane's localNameEvidence; rule2cClassNameRecv keeps main's ClassNameRecvNames signature (redhat-et#280's memberFields) with the lane's wording for condition (2). - src/graph.h: localNameSet takes the lane's evidence bits AND keeps main's VarDecl localShadowSpans (Java shadows). - A composition call, not a textual one: main's localNameSet loop skipped a record assignmentNamesNoClass drops (redhat-et#278: a member assigned from a call read as a local, so Rule 2b refused its declared type). The lane solves that same refusal with the declared bit, and its arm v3 needs Rule 2c to keep the assignment as evidence the token is a variable (`Widget = makePane();`), which the skip would drop. The loop now reads every record; assignmentNamesNoClass still guards the varType and field use-site tables, and its comment says the local-name set joined the two. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

joyful-ii-V-I and others added 2 commits September 17, 2026 07:13

joyful-ii-V-I added the train-member Lands through an integration train; runs the light CI set label Sep 17, 2026

joyful-ii-V-I mentioned this pull request Sep 17, 2026

fix(ingest): a member held by std::unique_ptr or std::shared_ptr recorded no type, so calls through it guessed #282

Merged

joyful-ii-V-I merged commit 4bf9ac6 into main Sep 17, 2026
15 of 45 checks passed

joyful-ii-V-I mentioned this pull request Sep 17, 2026

integration: train 5 — Rule 2b reads smart-pointer members (#282), assigned members and base-class members #284

Merged

Navigation Menu

fix(resolve): assigning a variable from a function call erased the type it was declared with - #278

fix(resolve): assigning a variable from a function call erased the type it was declared with#278
joyful-ii-V-I merged 2 commits into
mainfrom
lane/ctor-inferred-type

joyful-ii-V-I commented Sep 17, 2026 •

edited by coderabbitai Bot

Loading

Uh oh!

coderabbitai Bot commented Sep 17, 2026 •

edited

Loading

Walkthrough

Changes

❌ Failed checks (1 warning)

Uh oh!

joyful-ii-V-I commented Sep 17, 2026

Uh oh!

joyful-ii-V-I commented Sep 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

joyful-ii-V-I commented Sep 17, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Defect

Change

Measured

Built and rejected

Gates

Pins moved

Train notes

Summary by CodeRabbit

Uh oh!

coderabbitai Bot commented Sep 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

❌ Failed checks (1 warning)

Uh oh!

joyful-ii-V-I commented Sep 17, 2026

Uh oh!

joyful-ii-V-I commented Sep 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

joyful-ii-V-I commented Sep 17, 2026 •

edited by coderabbitai Bot

Loading

coderabbitai Bot commented Sep 17, 2026 •

edited

Loading