Skip to content

analyze: prove document offsets past a non-ASCII character - #5

Merged
rogulia merged 1 commit into
mainfrom
test/analyze-span-offsets
Sep 18, 2026
Merged

rogulia merged 1 commit into
mainfrom
test/analyze-span-offsets

Conversation

@rogulia

@rogulia rogulia commented Sep 18, 2026

Copy link
Copy Markdown
Contributor

Test-only, closing a proof gap left by #4, in parallel with polytypo/polytypo-js#14.

Every html/markdown offset assertion that landed with analyze() was ASCII-only, in all five runtimes — so none of them separated a code-point document offset from a byte one through the span adapters, which is where a byte-oriented runtime would get it wrong. Python is not byte-oriented and the code is right; the five ports share one proof set, so the assertion should discriminate here too.

café in the first span puts the UTF-8 byte offset one ahead of the code-point offset, and the test asserts both numbers.

Checks: 18 analyze tests pass, ruff clean.

🤖 Generated with Claude Code

Every html/markdown offset assertion added with analyze() was ASCII-only, so
none of them discriminated a code-point document offset from a byte one through
the span adapters — the place a byte-oriented runtime actually gets it wrong.
This runtime is not byte-oriented, but the five ports share one proof set.

`café` in the first span puts the UTF-8 byte offset one ahead of the code-point
offset, and the test now asserts both numbers rather than only the one it wants.

Test-only. 18 analyze tests pass.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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.

1 participant