test(project): close the gaps the merge-wave reviews left in the guard tests (#1902)

[JarryShaw]

Please follow the guide below

• You will be asked some questions, please read them carefully and answer honestly

• Put an x into all the boxes [ ] relevant to your pull request (like that [x])

• Use Preview tab to see how your pull request will actually look like

• Searched for similar pull requests

• Followed the coding style (make pylint, make mypy, make isort)

• make test passes, and a test case covers the change. Narrow runs only, listed below.

• Added a changelog entry under docs/source/changelog/ and regenerated CHANGELOG.md, if the change is user-visible: N/A — centralised in docs(changelog): shared 1.5.0 changelog — long-lived, merges last (#610, #616, #617, #618, #620) #657

What is the purpose of your pull request?

Tick the commit type your subject line carries.

• fix — corrects a defect
• feat — adds a feature
• perf — changes performance, not behaviour
• refactor — changes neither behaviour nor performance
• test — tests only
• docs — documentation only
• ci — workflows or build tooling
• release — bumps the version or rolls up a distribution
• chore — anything else

These close the four gaps the merge-wave cross-reviews left.

• ast.unparse gate (test(project): keep the #1851 and #1878 tests green on the KTLO legs (#1898, #1899) #1900): a use counts as gated only in the branch of a version check that runs on 3.9 or later, or in a test whose skip decorator restricts it there. The else: of a < gate counts too. getattr(ast, 'unparse') now counts as a use.
• Directive regex (docs(const): drop version directives from generated tables (#1884) #1895): \.\.\s+version…, so .. versionchanged:: written with two spaces is caught.
• Conda install test (ci(conda): move platform pins up and give them a review-by date (#1851) #1885): a pip_installs() helper splits with comments=True and matches requirement names. A requests after a # no longer counts, and requests[socks] and requests>=2 now do.
• time_limit(1) tests (test(protocols): the #1889 swallowed-raise test can see the deadline arrive without the swallow #1897): they hold back the deadline's first arm until the body is inside its try, as test(protocols): keep the #1889 spin's first raise inside its try (#1897) #1901 did. A new test that stalls before the try fails without this.

Each new check fails with its fix reverted. tests/project, tests/test_support_helpers.py and the tier guard pass on 3.14, and under plain unittest on 3.8 the touched test classes pass.

Closes #1902

[JarryShaw]

Verdict: GOOD TO GO at c54724d17. A Sonnet cross-review checked all four fixes:

• ast.unparse scan: it gave the right answer on 28 probe snippets, including elif chains, nested gates, a (3, 8) threshold, conditional expressions, aliases and getattr with a default. The real tree still passes.
• Directive regex: it finds exactly what the old one found in every pcapkit/**/*.py file, and it now also catches two spaces or a tab.
• pip_installs(): it drops anything after a # and reads requests[socks] and requests>=2 as requests.
• held_deadline(): it holds only the first arm. With time_limit's re-arm removed, the repeat tests fail. TimeLimitTests passed in 20 of 20 loops.
• Suites: tests/project, tests/test_support_helpers.py and the tier guard pass. The touched classes pass on 3.8.

Not blocking: the scan accepts version_info[1:] >= (9,), and the regex also matches .. followed by a newline, which is a reST comment. Neither occurs in the tree.

[github-actions]

Coverage: 90.07% (unit tier, Python 3.15, c54724d17, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	2	20	2	96.77%
pcapkit/const	19595	881	2618	867	92.11%
pcapkit/corekit	2504	70	792	36	96.54%
pcapkit/dumpkit	276	3	100	3	98.40%
pcapkit/foundation	2916	164	1016	34	93.59%
pcapkit/interface	138	7	52	5	93.68%
pcapkit/protocols	17411	224	4580	189	98.10%
pcapkit/toolkit	872	177	298	8	77.18%
pcapkit/utilities	522	2	158	4	99.12%
pcapkit/vendor	4431	2341	1018	156	43.64%

Per-file detail: the coverage-html artifact of this run.