fix(ci): pin the root-dominance corpus to the shadow-stack lowering

[proggeramlug]

GC Root Dominance has been unable to return a verdict since #7370 made statepoints the default. Last green a3b35664c, first red 148f97bc2 = #7370.

It is gate breakage, not a dominance violation

error: 0 root store(s) in the corpus, need at least 1500. The subject of this
check never ran — a clean verdict here means the IR was not the IR you think it is
=== checked 2378 functions / 144 modules (144 .ll files, 0 root stores)

The checker's entire vocabulary is call void @js_shadow_slot_bind(...). Under the stack-map lowering, the final IR pass resolves those indices to native allocas and removes the calls — FunctionCodegen::stack_map_slot_count says so directly. So the corpus compiled fine and contained zero of the thing being counted.

The --min-binds floor did exactly its job: it caught that the gate's own subject never ran, rather than reporting a clean violations: 0. CLAUDE.md's fourth hazard, working as designed.

The fix, and why pinning is sound

Both corpus scripts now set PERRY_RS4GC=0. #7340 split the root-set analysis from its lowering; this gate is about the analysis, which both backends share. The shadow stack also remains the production lowering wherever the runtime cannot walk frames. PERRY_RS4GC is already in the object-cache key, so no stale-object risk.

Measured — same binary, same source, only the knob differing:

arm=default   js_shadow_slot_bind calls = 0     ← reproduces the CI signature exactly
arm=rs4gc0    js_shadow_slot_bind calls = 9

This exact breakage was already fixed once

#7370 handled it for the unit tests — helpers.rs records that eight tests broke when the default flipped and were given NativeRootsPin::shadow(). The corpus shell scripts were the same breakage in a different idiom, and were missed.

Summary by CodeRabbit

• Documentation
  • Documented the configuration used by the GC Root Dominance corpus scripts.
  • Recorded the rationale, measured bind counts, and corresponding unit-test adjustment.

[coderabbitai]

Caution

Review failed

The pull request is closed.

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 5c88a45b-22cd-4f20-879f-ce5c87dc0530

📥 Commits

Reviewing files that changed from the base of the PR and between 3ceffe6 and 39de4d1.

📒 Files selected for processing (1)

• changelog.d/7415-dominance-corpus.md

---

📝 Walkthrough

Walkthrough

The changelog records the GC Root Dominance corpus configuration. It documents pinning PERRY_RS4GC=0 because default statepoint lowering removes js_shadow_slot_bind calls and causes the --min-binds gate to reject the corpus.

Changes

GC Root Dominance corpus

Layer / File(s)	Summary
Document corpus validation configuration changelog.d/7415-dominance-corpus.md	Documents the default-statepoint gate failure, measured bind counts, and pinning PERRY_RS4GC=0 in the corpus scripts.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Possibly related PRs

• PerryTS/perry#7198: Relates to the GC Root Dominance checker workflow and its --min-binds validation.
• PerryTS/perry#7370: Relates to the statepoints-default change documented here.
• PerryTS/perry#7348: Relates to the PERRY_RS4GC=0 shadow-stack configuration.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/gc-root-dominance-corpus

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

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