fix(gen-zig): assert invariant predicates, return tail values, discard value calls (owner exception, #5980) - #6579
Merged
Merged
Conversation
gHashTag
added
the
owner-approved-foreign
This was referenced Oct 5, 2026
…ression, discard a value call (Closes #6315) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Refs #6315 Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…rced tests-fail seals (#6315, #6560) Refs #6315 Regenerated on the Railway t27c lab at 2cab1b5 (tri seals drift --fix, sync-twins). tri27_machine, clock_domain_tb and gf16_accel_tb now compile but their own tests fail or hang; sealed with --force (failures recorded in the seal) and ledgered as tests-fail. check_seal_currency and check_seal_coverage pass on the lab. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
gHashTag
force-pushed
the
claude/gen-zig-value-ignored
branch
from
October 5, 2026 17:57
7d4aaf0 to
26c013c
Compare
Contributor
This was referenced Oct 5, 2026
Contributor
PR DashboardGenerated at: 2026-10-05 18:09:01 UTC
Summary
Seal Status
|
This was referenced Oct 5, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #6315
What
t27c gen-zig produced Zig that would not compile ("value of type 'bool' ignored" and related errors). It now:
_ = f(...)).specs/compiler/zig_value_ignored.t27is the regression spec. It replaces the hand-written Rust testbootstrap/tests/value_ignored_zig.rs, which this PR deletes.Owner exception
This PR edits hand-written Rust (
bootstrap/src/compiler.rs), which the only-t27 rule forbids. The owner approved it on 2026-10-05, relayed by the coordinator ("fix it and do not stop, do what is best"): the parked gen-zig fixes land as exceptions, one class per PR, with the labelowner-approved-foreign. The debt is #5980 (t27core self-host); these fixes move to t27core there.The entries added to
tools/policy/foreign-exceptions.txt(listed here as the coordinator asked; they stay, because the next gen-zig PRs #6451/#6532, #6533 and #6295 need them):Once the last of those PRs lands, the owner should remove it.
File classification
bootstrap/src/compiler.rsbootstrap/stage0/FROZEN_HASHbootstrap/tests/value_ignored_zig.rsbootstrap/tests/bare_semicolon_zig.rsside(1);, which Zig rejects becausesidereturns u32; it now pins_ = side(1);specs/compiler/zig_value_ignored.t27.trinity/seals/*.json(about 190)tri seals drift --fix,t27c seal --save --force,tri seals sync-twins)tools/seal_baseline.txttests-failrowstools/policy/foreign-exceptions.txtdocs/now/2026-10-06-gen-zig-value-ignored.mdLab numbers (t27c reference: gen-zig + zig test over specs/, Railway t27c lab)
@intCastpanic) and the trios-cli config (the spec's test contradicts its own default). Both are filed in gen-zig: same-width sign cast panics (tri27_machine); trios config test contradicts its default #6560.CI notes
bare_semicolon_zigis fixed in this PR (see the table). The other two newly failing tests,the_dead_code_census_names_what_it_skippedandcorpus_classifier_matches_lean_completeness(ar_ternary_logic), fail the same way on another PR at the same master (claude/t27c-gen-typecheck-gate-rebased). They come from master, not from this change.ring-096-rustvsspecs/numeric/formats.t27DRIFTED, the same failure as on master (spec-guards: ring-096 drift and 8 published-figure pins were hidden behind the seal step #6553).cargo install wasm-objdumpcannot find the crate, an infrastructure problem.Seals
Seals were regenerated on the lab (2cab1b5; a drift pass at 26c013c on the newer master found nothing more). Three specs that now compile have failing or hanging tests: tri27_machine (1 of 10 fail), clock_domain_tb (hangs) and gf16_accel_tb. They were sealed with
--force, which records the failures in the seal'stestsobject. Their 5 seals (twins included) are ledgered astests-failintools/seal_baseline.txtwith #6560. On the lab:check_seal_currencypasses, andcheck_seal_coveragereports OK (1454 seals, 1325 hold, 129 known-broken).spec-guards has been red on master since 2026-09-16 for reasons this PR does not touch (#6553).
🤖 Generated with Claude Code