Skip to content

fix: recognize Unicode identifiers in binding analysis - #207

Merged
rs545837 merged 1 commit into
Ataraxy-Labs:mainfrom
dajiaohuang:c1-3148-unicode-ident-boundaries
Oct 28, 2026
Merged

rs545837 merged 1 commit into
Ataraxy-Labs:mainfrom
dajiaohuang:c1-3148-unicode-ident-boundaries

Conversation

@dajiaohuang

Copy link
Copy Markdown
Contributor

Summary

  • Keep identifier boundaries character-aware throughout binding analysis, including parsed definition names and footprint tokenization.
  • Add synthetic regressions for a Unicode call that is a prefix of a longer definition, Unicode definition extraction, and Unicode producer/consumer footprints.

The existing non-ASCII call scanner already advances by Unicode character; this closes adjacent byte-oriented paths that could lose names or misclassify boundaries.

Validation

  • cargo fmt --all -- --check
  • cargo test --locked -p weave-core
  • cargo clippy --locked -p weave-core --all-targets -- -D warnings

All commands passed in task-local Cargo home/target directories. No CI or review status is asserted.

@inspect-review inspect-review Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

inspect review

Triage: 6 entities analyzed | 0 critical, 0 high, 6 medium, 0 low
Verdict: standard_review

Findings (0)


Reviewed by inspect | Entity-level triage found 0 high-risk changes

@rs545837
rs545837 merged commit d9bf676 into Ataraxy-Labs:main Oct 28, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants