Repository navigation
fix(tinylinkify): leave www out of the URL search when it doesn't link www. - #36
Conversation
…www. With fuzzyLink off, every www match was turned down, and each one read the rest of its host again: 80 KB took 3.7 seconds. The search now has no www to find then. The fix for rejected candidates in #35 grew the root and tinylinkify/gfm past their size budgets, by about 120 B gzip, so the budgets go to 1985 B and 2460 B and the sizes in the docs to match.
Codex Review Summary: ❌ FailedThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
tinyaml
tinyini
tinyjson5
tinylinkify
tinyswap
commit: |
There was a problem hiding this comment.
✅ No new issues found. No P0 or P1 findings remain.
Reviewed changes — Reviewed all seven changed files, including the GFM finder, its public callers, the regression test, and bundle-size updates.
- Option-aware URL search: Uses an HTTP-only regex when
fuzzyLinkis disabled, avoiding repeated scans of rejectedwwwcandidates while preserving capture groups and URL processing. - Performance regression coverage: Exercises the long repeated-
wwwinput withfuzzyLink: falseand checks that the following HTTPS link is preserved. - Size budgets and documentation: Raises the root and
/gfmbudgets and updates published sizes to match the measured bundles.
Validation passed: all 36 tests against both source and built exports, formatting/lint/type checks, and bundle-size checks. An additional 800,048 comparisons against the pre-change implementation agreed across all eight option combinations; the 80 KB reproducer improved from about 4.1 seconds to 0.63 ms without changing its results.
gpt-6.1-sol(xhigh) | 𝕏

With
fuzzyLink: false,tinylinkify/gfmstill stalls on text like':// ' + 'www._'.repeat(16000) + '.a.b': 80 KB takes 3.7 s. The URL search turns down everywwwmatch whenwww.links are off, and each one reads the rest of its host again, so the fix for rejected candidates in #35 never kicks in.This gives the URL search no
wwwalternative whenwww.links are off, since every such match would be turned down anyway. The same input now takes under a millisecond. A new test fails without the fix, andfindon main and on this branch agree on 400,000 fuzzed strings under all 8 combinations of options.#35's fix also put the root and
tinylinkify/gfmabout 120 B gzip over their size budgets, which CI doesn't check. This raises the budgets to 1985 B and 2460 B and updates the sizes in the docs.Follow-up to #35
Reviewed with sol-loop: 1 round with gpt-6.1-sol on xhigh. Sol verdict: approved