docs(const,vendor): six :rfc:`959#section-4.1` citations name an anchor RFC 959 does not have

[JarryShaw]

Describe the bug

Six :rfc: citations name 959#section-4.1. The fragment is well-formed but the anchor does not exist, so each renders a live link to nothing — the same user-visible defect as #942's secion-3, reached a different way. #943's new shape invariant cannot catch this class, by design: it checks that a fragment looks like an anchor, never that the anchor is real.

RFC 959 predates per-subsection anchors. Its rendered HTML carries only eight section- ids — section-1 through section-8 — on both mirrors:

$ curl -s -A 'Mozilla/5.0' -L https://www.rfc-editor.org/rfc/rfc959.html | grep -o 'id="section-[0-9.]*"' | sort -u
id="section-1" id="section-2" id="section-3" id="section-4" id="section-5" id="section-6" id="section-7" id="section-8"
$ ... | grep -c 'id="section-4\.1"'
0

Identical result from https://datatracker.ietf.org/doc/html/rfc959.

Reproduction

$ grep -rn '959#section-4\.1' pcapkit/ | grep -v __pycache__
pcapkit/const/ftp/command.py:187        pcapkit/vendor/ftp/command.py:199
pcapkit/const/ftp/command.py:261        pcapkit/vendor/ftp/command.py:273
pcapkit/const/http/method.py:268        pcapkit/vendor/http/method.py:173

Three template/generated pairs. The string is a literal in each vendor/ template, not synthesised, so both halves are hand-editable — no regeneration or IANA crawl is needed.

Expected behavior

A fragment that resolves. This needs a judgement rather than a spelling fix, which is why it is its own issue:

• #section-4 — resolves, and §4 "FILE TRANSFER FUNCTIONS" does contain §4.1 "FTP COMMANDS". Loses precision.
• #page-N — page- is a real anchor form (id="page-5" verified present) and Sphinx titles it, per sphinx.roles._format_rfc_target, whose known set is {'appendix', 'page', 'section'}. Precise but brittle against re-pagination.
• Drop the fragment and cite :rfc:959`` plainly.

My lean is #section-4, but the call is the maintainer's.

Notes

• Found by the Opus cross-review of docs(const,vendor): fix RFC anchor typos and missing section- prefixes #943, which spotted it while checking whether that PR's stricter regex was too strict.
• Related: docs(const,vendor): FEATCode's docstring cites RFC 5797 "#secion-3", a typo that renders a wrong anchor #942 (the secion typo), docs(const,vendor): fix RFC anchor typos and missing section- prefixes #943 (the shape invariant and the missing section- prefixes). This is a third, distinct class: right shape, dead target.
• Worth considering with it: whether the shape test should grow an existence check for a fixed allowlist of cited RFCs. That would need network access in CI, so probably not — but the limitation should at least be stated in the test's own docstring.

System information

• pcapkit commit 9ea0d6a5a, Python 3.14.7, CPython, Linux.

[JarryShaw]

Sure section-4 is good.

[JarryShaw]

Ruling from the maintainer — #section-4 it is. Verbatim:

Sure section-4 is good.

So all six sites become ``:rfc:959#section-4```. #page-N` and the bare-`:rfc:` option are both declined.

Re-verified on current main (d7489c61c, after #943 and #945 merged) — all six are still present and the count is unchanged:

pcapkit/const/ftp/command.py:187    pcapkit/vendor/ftp/command.py:199
pcapkit/const/ftp/command.py:261    pcapkit/vendor/ftp/command.py:273
pcapkit/const/http/method.py:268    pcapkit/vendor/http/method.py:173

Three template/generated pairs, each a literal in the vendor/ template, so both halves are hand-edited and no crawl or regeneration is needed. #943's shape invariant already accepts section-4, so the fix is inside the existing test's grammar.

Labels: needs: decision off, wip on. A worker is on it.