[JarryShaw]

• Searched for similar pull requests
• Followed the coding style
• make test passes, and a test case covers the change
• Changelog entry: 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?

• ci — CI configuration

---

Description

Closes #2218. Milestones stopped being set after #1068. The closed items have been backfilled by hand onto PyPCAPKit v1.5, and this workflow keeps them set from now on.

• milestone.yml: runs on issues and pull_request_target (opened, reopened, closed) and on workflow_dispatch, with its logic in util/milestone.py. Opening an item sets the milestone, per the owner on ci: set milestones automatically on close #2218.
  • An open item, a merged PR or a completed issue gets the open milestone titled PyPCAPKit v<major>.<minor>. The version is parsed from the base branch's pcapkit/__init__.py, not imported.
  • An unmerged PR, or a not-planned or duplicate issue, loses its milestone.
  • With no matching milestone, it warns and creates nothing.
  • A milestone already on an item, for example one set by hand to a later release, is kept. Only a removal overrides it.
• Backfill: workflow_dispatch covers open and closed items above the previous milestone's highest number, above Prevent maintenance workflow push races #336 today. It defaults to dry_run: true.
• Security: it checks out only github.sha (the base tip) with persist-credentials: false, fetching only the three files it needs. permissions: {} at the top; the job holds only contents: read, issues: write and pull-requests: write. The item is re-read through the API, and only its number comes from the payload.
• Docs: workflows.rst documents the workflow, and process.rst names the real milestone titles.
• Fail-without-fix: the new tests cannot collect on origin/main. With only the script present, they give 11 failed (all workflow and step tests). Against the close-only version, the open, reopen and precedence tests give 16 failed.
• Tests run: the new file gives 49 passed. The workflow-wide tests/project selection gives 733 passed. actionlint is clean.
• Overlap: docs(contributing): sixth whole-repo pass over the non-API docs (#1103) #2213 and ci(compat): compile the 3.8-floor util scripts on the 3.8 leg (#2206) #2214 also edit workflows.rst, and git merge-tree merges each of them with this cleanly.

[JarryShaw]

Review: GOOD TO GO at d55eafa3f (Sonnet). Three non-blocking follow-ups are being folded in, after which a delta review follows.

• Security: it checks out github.sha only, sparse, with persist-credentials: false. The item number goes through env:, with no ${{ }} in run:. __version__ is read with ast.parse. Permissions are minimal. The hardening and pin tests pass.
• Rule: action_for sets on a merged PR or a completed issue and removes on an unmerged PR or a not-planned or duplicate issue. It does nothing on no reason or a reopen. This matches the existing v1.5 milestone.
• Versions: 1.5.0b14 maps to v1.5, 1.10.0 to v1.10 and 2.0.0.post1 to v2.0. A malformed version raises. A missing or closed milestone gives a warning and no write.
• Dry-run, live with a GET-only wrapper: 7 GETs and no writes. The bound is Prevent maintenance workflow push races #336, and nothing pre-1.5 is touched.
• Tests: 445 passed across 19 files. Fail-without-fix: the tests do not collect on origin/main.
• Follow-ups:
  • Backfill PATCHes have no throttle or retry, so a large run could hit the secondary rate limit.
  • Concurrency lacks queue: max, unlike project-status.yml.
  • A header comment is change-relative.

[JarryShaw]

Review: GOOD TO GO at 86dbdad92 (Sonnet delta review of d55eafa3f→86dbdad92).

• Throttle: 1.0s between backfill writes only, with none on a dry run or a single item.
• Retry:
  • The --include parsing reads a synthetic 403 with Retry-After: 30 as rate-limited, and a plain 403 as not, so a plain 403 is not retried.
  • Waits are at most 3 (60/120/240s, or Retry-After), then it raises. The PATCH is idempotent.
  • If gh printed no headers on failure, it would fall back to the previous fail-fast behaviour.
• Timeout: timeout-minutes: 60 covers about 7 minutes of throttled writes plus retry stalls.
• Concurrency: queue: max matches project-status.yml, and the Lint step's existing actionlint ignore covers it.
• Comments: timeless and accurate.
• Tests: test_milestone_2218.py gives 45 passed. The workflow-wide selection gives 243 passed.

[JarryShaw]

Review: GOOD TO GO at 95feb5679 (Sonnet delta review of 86dbdad92→95feb5679, set-on-open per #2218).

• Triggers: [opened, reopened, closed] on both events. A fork PR's opened still checks out only github.sha (base) with persist-credentials: false, and only the number passes through env. Concurrency is per item with queue: max, so a burst cannot starve other items.
• Rule: set on an open item, a merged PR or a completed issue. remove on unmerged, not-planned or duplicate. No change on a close with no reason.
• Precedence: set writes only when the item holds no milestone. A different milestone is kept and logged, so no path overwrites a hand-set later release. remove takes whatever the item holds.
• Backfill: state=all&milestone=none, with the same bound. Dry-run writes nothing.
• Tests: 49 passed in test_milestone_2218.py, and 247 passed in the workflow-wide selection. Fail-without-fix: 16 failed against 86dbdad92.

[github-actions]

Coverage: 91.56% (unit tier, Python 3.15, 95feb5679, Unit Tests run success)

Package	Statements	Missed	Branches	Partial	Cover
pcapkit (top level)	104	4	20	2	95.16%
pcapkit/const	19599	881	2618	867	92.11%
pcapkit/corekit	2589	74	816	37	96.51%
pcapkit/dumpkit	282	3	102	3	98.44%
pcapkit/foundation	2994	23	1038	25	98.81%
pcapkit/interface	138	4	52	3	96.32%
pcapkit/protocols	17433	218	4584	185	98.15%
pcapkit/toolkit	869	36	298	10	95.20%
pcapkit/utilities	504	7	146	5	98.15%
pcapkit/vendor	4432	1988	1018	167	51.87%

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

[JarryShaw]

Ready to merge at 95feb5679. Closes #2218.

• Review: Sonnet GOOD TO GO, after two folded rounds: the throttle and retry, then set-on-open per the owner on ci: set milestones automatically on close #2218.
• CI: all checks complete, Required checks passed SUCCESS, 0 failed.
• KTLO: 12 legs, no warnings.
• Ordering: merged into main and run. project leg: 2347 tests, 0 failures.
• Conflicts: none against the current main. docs(contributing): sixth whole-repo pass over the non-API docs (#1103) #2213 and ci(compat): compile the 3.8-floor util scripts on the 3.8 leg (#2206) #2214 also edit workflows.rst, and git merge-tree merges each with this cleanly.
• First runs: after the merge, every open, reopen and close sets or settles the milestone. All open items already carry PyPCAPKit v1.5, and the closed backlog has been backfilled, so a workflow_dispatch dry run should find nothing to change.