[GH-ISSUE #6731] [Docs] Dead same-file anchors in the Malayalam README and unambiguous documentation errors from a full-repo sweep #4094

Closed
opened 2026-10-31 00:46:46 +00:00 by bsakaguchi · 3 comments
Owner

Originally created by @ashyyhere on GitHub (Oct 23, 2026).
Original GitHub issue: https://github.com/morluto/rea/issues/6731

Before submitting

  • I searched the existing issues for a related documentation report.
  • I removed credentials, tokens, private paths, and sensitive target data.

Documentation location

  1. Three same-file anchors in docs/readme/README_ml.md:
    • The Node.js requirements badge link near the top (line 14) targets #what-you-can-analyze.
    • The documentation bullet (line 250) targets #showcases.
    • The agent setup bullet (line 278) targets #update-rea.
  2. Prose errors found in the same full-repository documentation sweep:
    • docs/il2cpp-metadata.md: merged words (bounded to1GiB, limited to65,536, a64MiB, exact64-bit, characters,32MiB file,1,000,000).
    • docs/windows-ghidra-p0.md: dropped predicate ("Caller priority is preserved, and caller CPU affinity whenever ...").
    • docs/browser-observation.md: a paragraph and three sentences begin in lowercase.
    • docs/installation.md: "the former remains the diagnostic fallback" follows a four-item list with no pair to refer to.
    • docs/agent-environments.md: nonstandard "another origin than".
    • docs/releasing.md: an h3 sits directly under the h1 while every sibling section is h2.
    • docs/testing.md: a bullet ends without terminal punctuation.
    • docs/apple-application-analysis.md, docs/process-capture.md, website/README.md, website/figures.md: stray Unicode right single quotes where the files otherwise use ASCII apostrophes.

Problem

The Malayalam README defines none of the three anchors its own links target — those were the English section names the translation no longer uses — so clicking them on GitHub does nothing. The prose errors above mislead or distract readers; the merged words render as run-together tokens (for example to1GiB), and "the former" has no antecedent.

Expected content

  • Repoint the three links at the localized headings:
    • line 14 → #എന്തൊക്കെ-വിശകലനം-ചെയ്യാം (heading at line 144)
    • line 250 → #ഉദാഹരണങ്ങൾ (heading at line 191)
    • line 278 → #rea-അപ്‌ഡേറ്റ്-ചെയ്യുക (heading at line 112)
  • Correct the listed prose errors with the minimal edits described above.

Supporting evidence

  • Each replacement target was taken from GitHub's rendered user-content- heading ids for docs/readme/README_ml.md on main, not from a recomputed slug. GitHub retains U+200C (zero-width non-joiner) in ids such as #rea-അപ്‌ഡേറ്റ്-ചെയ്യുക; the npm github-slugger package strips it, and README_fa.md is correct as-is for the same reason.
  • tests/conformance/readmeConsistency.test.ts currently guards only the Persian anchors, so the Malayalam forms had no regression coverage.
  • A fix is already open: #6383 (adds the anchor repairs, the prose corrections, and a Malayalam case in readmeConsistency.test.ts).

Fixes the reporting gap for these findings; the fix PR is #6383.

Originally created by @ashyyhere on GitHub (Oct 23, 2026). Original GitHub issue: https://github.com/morluto/rea/issues/6731 ### Before submitting - [x] I searched the existing issues for a related documentation report. - [x] I removed credentials, tokens, private paths, and sensitive target data. ### Documentation location 1. Three same-file anchors in [`docs/readme/README_ml.md`](https://github.com/morluto/rea/blob/main/docs/readme/README_ml.md): - The Node.js requirements badge link near the top (line 14) targets `#what-you-can-analyze`. - The documentation bullet (line 250) targets `#showcases`. - The agent setup bullet (line 278) targets `#update-rea`. 2. Prose errors found in the same full-repository documentation sweep: - `docs/il2cpp-metadata.md`: merged words (`bounded to1GiB`, `limited to65,536`, `a64MiB`, `exact64-bit`, `characters,32MiB file,1,000,000`). - `docs/windows-ghidra-p0.md`: dropped predicate ("Caller priority is preserved, and caller CPU affinity whenever ..."). - `docs/browser-observation.md`: a paragraph and three sentences begin in lowercase. - `docs/installation.md`: "the former remains the diagnostic fallback" follows a four-item list with no pair to refer to. - `docs/agent-environments.md`: nonstandard "another origin than". - `docs/releasing.md`: an `h3` sits directly under the `h1` while every sibling section is `h2`. - `docs/testing.md`: a bullet ends without terminal punctuation. - `docs/apple-application-analysis.md`, `docs/process-capture.md`, `website/README.md`, `website/figures.md`: stray Unicode right single quotes where the files otherwise use ASCII apostrophes. ### Problem The Malayalam README defines none of the three anchors its own links target — those were the English section names the translation no longer uses — so clicking them on GitHub does nothing. The prose errors above mislead or distract readers; the merged words render as run-together tokens (for example `to1GiB`), and "the former" has no antecedent. ### Expected content - Repoint the three links at the localized headings: - line 14 → `#എന്തൊക്കെ-വിശകലനം-ചെയ്യാം` (heading at line 144) - line 250 → `#ഉദാഹരണങ്ങൾ` (heading at line 191) - line 278 → `#rea-അപ്‌ഡേറ്റ്-ചെയ്യുക` (heading at line 112) - Correct the listed prose errors with the minimal edits described above. ### Supporting evidence - Each replacement target was taken from GitHub's rendered `user-content-` heading ids for `docs/readme/README_ml.md` on `main`, not from a recomputed slug. GitHub retains U+200C (zero-width non-joiner) in ids such as `#rea-അപ്‌ഡേറ്റ്-ചെയ്യുക`; the npm `github-slugger` package strips it, and `README_fa.md` is correct as-is for the same reason. - `tests/conformance/readmeConsistency.test.ts` currently guards only the Persian anchors, so the Malayalam forms had no regression coverage. - A fix is already open: #6383 (adds the anchor repairs, the prose corrections, and a Malayalam case in `readmeConsistency.test.ts`). Fixes the reporting gap for these findings; the fix PR is #6383.
Author
Owner

@XinwangTao commented on GitHub (Oct 24, 2026):

Good catch, and the U+200C detail is the part worth preserving: GitHub keeps zero-width non-joiners in rendered heading ids while github-slugger strips them, so any automated anchor check that relies on slugger output will keep producing false "missing anchor" reports (or miss real ones) for scripts with ZWNJ. Glad #6383 adds a Malayalam case to readmeConsistency.test.ts — I'd suggest the test also explicitly pin the ZWNJ-retention expectation with a comment, so future languages (e.g. Arabic/Hindi adjacent scripts) inherit the same guard instead of rediscovering the discrepancy. The prose fixes look minimal and unambiguous; nothing to add there.

<!-- gh-comment-id:6378918797 --> @XinwangTao commented on GitHub (Oct 24, 2026): Good catch, and the U+200C detail is the part worth preserving: GitHub keeps zero-width non-joiners in rendered heading ids while github-slugger strips them, so any automated anchor check that relies on slugger output will keep producing false "missing anchor" reports (or miss real ones) for scripts with ZWNJ. Glad #6383 adds a Malayalam case to readmeConsistency.test.ts — I'd suggest the test also explicitly pin the ZWNJ-retention expectation with a comment, so future languages (e.g. Arabic/Hindi adjacent scripts) inherit the same guard instead of rediscovering the discrepancy. The prose fixes look minimal and unambiguous; nothing to add there.
Author
Owner

@morluto commented on GitHub (Oct 26, 2026):

Thank you for the careful documentation sweep and the rendered-heading evidence. The three Malayalam links now use the reported localized fragments on main, and the reported prose corrections have also landed. #6383 remains open for the separate regression test that guards the Malayalam heading and zero-width character. I am closing the documentation report; this triage checked the current source and did not rerun GitHub rendering or the test suite.

<!-- gh-comment-id:6428683087 --> @morluto commented on GitHub (Oct 26, 2026): Thank you for the careful documentation sweep and the rendered-heading evidence. The three Malayalam links now use the reported localized fragments on main, and the reported prose corrections have also landed. #6383 remains open for the separate regression test that guards the Malayalam heading and zero-width character. I am closing the documentation report; this triage checked the current source and did not rerun GitHub rendering or the test suite.
Author
Owner

@XinwangTao commented on GitHub (Oct 27, 2026):

同意关闭。三个锚点修复已落地,#6383 的回归测试确实是关键——ZWNJ 导致 slugger 与 GitHub 渲染不一致的问题不只影响 Malayalam,阿拉伯语/印地语等含零宽连接符的脚本都会踩到。建议 #6383 在固定 Malayalam 用例的同时,把锚点校验做成面向所有本地化 README 的通用检查(slugger 归一化 + GitHub 实际 heading id 双通道比对),避免后续新增语言时同类死链复发。

<!-- gh-comment-id:6431140621 --> @XinwangTao commented on GitHub (Oct 27, 2026): 同意关闭。三个锚点修复已落地,#6383 的回归测试确实是关键——ZWNJ 导致 slugger 与 GitHub 渲染不一致的问题不只影响 Malayalam,阿拉伯语/印地语等含零宽连接符的脚本都会踩到。建议 #6383 在固定 Malayalam 用例的同时,把锚点校验做成面向所有本地化 README 的通用检查(slugger 归一化 + GitHub 实际 heading id 双通道比对),避免后续新增语言时同类死链复发。
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
morluto/rea#4094
No description provided.