Fix misplaced anchors for remark and proof-like divs

[d-morrison]

Port div-anchors extension fix from Morrison-Lab/qwt#160 (Morrison-Lab/qwt#159).

Problem

Quarto theorem environments render titles with class .theorem-title, while proof-like environments (proof, remark, solution) render titles with class .proof-title. Previously, div-anchors.js only queried .theorem-title when relocating AnchorJS anchor links inline, causing remark and other proof-like divs to leave the anchor link at the bottom of the div rather than next to the title (for example, at linear-algebra.html#rem-zero-vector-identity).

Solution

• Support .proof-title alongside .theorem-title in div-anchors.js.
• In div-anchors.js, trim trailing whitespace from .proof-title (which ends with '. ') so the non-breaking separator and anchor stay attached to the title, and ensure a trailing space separates the anchor from subsequent body content if not already spaced.
• Add rem and sol aliases and crossref ID prefix matching (rem-, thm-, def-, etc.) to is_theorem_div in both Lua and JS.

[github-actions]

Claude finished review — View run

---

The diff is small and limited to the two files I already reviewed in full — nothing else is touched. This confirms the review is complete.

Summary of findings: none. The PR ports a fix verbatim (byte-for-byte identical diff, confirmed via gh pr diff 160 --repo Morrison-Lab/qwt) from the already-merged upstream Morrison-Lab/qwt#160, which addresses a real, reproducible problem — numerous rem-/sol-/def- divs exist across this repo's content (_subfiles/linear-algebra/_rem-zero-vector-identity.qmd and dozens of others) whose anchors were misplaced before this fix. Both the JS and Lua implementations are internally consistent (same class list, same ID-prefix list, same logic structure), the JS passes a syntax check, and the diff touches only the two vendored extension files — no unrelated scope creep, no build artifacts (_site/, _freeze/), no spell-check/link-check surface, and no R code involved (lab-manual R rules don't apply here). PR description explains the why clearly and matches CLAUDE.md's "Relationship to rme"-style provenance expectations (no internal-lab-provenance text leaked to readers). No prior review comments exist on this PR.

Verdict

Ready for merge

Structured Review Data (JSON)

{
  "schema_version": "1.1",
  "reviewer": "claude",
  "commit_sha": "da87d6bc01aaad4d4509f3e92a664d0293b14742",
  "verdict": "CLEAN",
  "findings": [],
  "detailed_assessment": "Checked _extensions/div-anchors/div-anchors.js and div-anchors.lua line by line; the diff is byte-identical to the already-merged upstream Morrison-Lab/qwt#160 fix, with no divergence or regression introduced during the port.",
  "holistic_assessment": "Scope stays confined to the vendored anchor-placement extension, matches the described bug (rem-/sol- prefixed divs like rem-zero-vector-identity exist throughout chapters), and carries no build-artifact or CI surface risk."
}

Reviewed commit: da87d6b

[github-actions]

💰 Cost: $0.3780 (review) — run

[github-actions]

PR Preview Action v1.8.1
Preview removed because the pull request was closed.
2026-10-07 00:02 PDT