test(compiler): measure host stack headroom instead of assuming ten times - #212
Merged
julia-script merged 2 commits intoAug 16, 2026
Merged
Conversation
…imes The host-control case of the shadow-stack bound asserted that a frameless recursion carries ten times the reporting depth before V8 refuses. That multiplier is an engine constant in disguise: V8 sizes its execution stack differently across versions and platforms, and node 26 on darwin/arm64 no longer fits ten thousand frameless frames (this container's node 22 fits about 10,400 — the old assertion was passing by under five percent). The control now takes its depth as an exported parameter, so one instance proves the fact the test actually needs: the exact depth whose reservation reported completes framelessly, and doubling the depth until the host refuses shows at least twofold headroom. The headroom is measured off the running engine like every other depth in the file is measured off the emitted module, instead of being written down. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BiBYbB23Jj8GKS2AEVEHX4
…atalog Five codes — PAR0002, SEM0054, SEM0089, SEM0094, SEM0103 — had an empty "Reported as" cell in the generated diagnostic index because the extractor only understood `message:` followed by one string literal. Their wording was therefore ungated by documentation:check, undermining the rule that tests assert codes while the catalog gates wording. The extractor now walks the message expression by nesting rather than by pattern: a conditional contributes every branch's complete template (the condition's literals are discriminants, not wording), and an expression with no literal of its own resolves the message helper it calls and takes the templates its return statements produce. A code with several templates lists each, and a template containing a backtick uses the padded double-backtick span form. Two factories are restructured so each variant is a complete literal instead of hiding wording inside an interpolation hole — PAR0002's expectation/context fragments and SEM0089's `&mut` access marker. Every produced message stays byte-identical. All 151 codes now carry at least one pinned template, and perturbing any of the five previously unpinned wordings fails documentation:check. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01BiBYbB23Jj8GKS2AEVEHX4
|
Bugbot is not enabled for your account, so this pull request was not reviewed. Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs. |
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
2 participants
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.
The host-control case of the shadow-stack bound asserted that a frameless
recursion carries ten times the reporting depth before V8 refuses. That
multiplier is an engine constant in disguise: V8 sizes its execution stack
differently across versions and platforms, and node 26 on darwin/arm64 no
longer fits ten thousand frameless frames (this container's node 22 fits
about 10,400 — the old assertion was passing by under five percent).
The control now takes its depth as an exported parameter, so one instance
proves the fact the test actually needs: the exact depth whose reservation
reported completes framelessly, and doubling the depth until the host
refuses shows at least twofold headroom. The headroom is measured off the
running engine like every other depth in the file is measured off the
emitted module, instead of being written down.
Co-Authored-By: Claude Fable 5 noreply@anthropic.com
Claude-Session: https://claude.ai/code/session_01BiBYbB23Jj8GKS2AEVEHX4