spec: Enable heading numbering for section references in shiroa

[RobinJadoul]

No description provided.

[github-actions]

Codex Code Review

Findings

• Low: spec/book.typ introduces a global per-page heading numbering scheme, but it is still overridden by the existing local rule in spec/logup.typ. That leaves logup.typ with different numbering from the rest of the Shiroa pages, so the PR does not actually make heading numbering consistent across all chapters. If the intent is standardization, this needs to be reconciled.

No security vulnerabilities or significant performance issues stood out in this diff.

[claude]

Review of spec/book.typ

This is a spec-only documentation change — no security implications.

Whitespace fix (line 147): Unambiguously correct. Fixes cross-references rendering as Section1.2 → Section 1.2.

Heading numbering (lines 188–191): Correct in structure — position() cannot return none because line 161 already guards that, and counter(heading).update(0) is placed after the phantom xref-include calls so the reset is clean.

Issues

Low — logup.typ overrides the new numbering

logup.typ line 4 has #set heading(numbering: "1."). In Typst, a set rule inside #body re-overrides the one set before it by book-page, so logup.typ headings will still display plain 1., 1.1. instead of the chapter-prefixed numbers. Fix: remove the #set heading(numbering: "1.") from logup.typ since book-page now handles it.

Low — heading depth limited to 2 levels (see inline comment on line 189)

The "1.1" pattern silently drops counter values beyond depth 2. logup.typ uses === (depth-3) headings, so those render chapter.a.b instead of chapter.a.b.c. Consider "1.1.1" if deep sub-section numbering is desired.

Low — shadowed args parameter (style)

book-page takes ..args as its own variadic parameter, and the new heading numbering closure also names its parameter ..args, shadowing the outer one. Not a bug (the outer args is already consumed before this point), but renaming the inner one to e.g. ..nums would improve readability.