Skip to content

Unblock the release: compare the roadmap against the published minor, not the proposed one - #44

Merged
sgjholt merged 1 commit into
mainfrom
claude/roadmap-priorities-pe63kc
Sep 16, 2026
Merged

sgjholt merged 1 commit into
mainfrom
claude/roadmap-priorities-pe63kc

Conversation

@sgjholt

@sgjholt sgjholt commented Sep 16, 2026 •

Copy link
Copy Markdown
Owner

Fixes the six failing jobs on #43.

What broke

test_the_roadmap_is_not_behind_the_released_minor, added in #42, read the version from .release-please-manifest.json. release-please bumps that file as part of preparing a release, so on #43 the manifest says 0.4.0 while the roadmap's newest section says v0.3.0, and the assertion fails:

AssertionError: v0.4.0 is released and the newest version docs/roadmap.md names is v0.3.0.
assert (0, 3) >= (0, 4)

That took out test on all five matrix entries, floors, and the ci gate with them. lint, typecheck, docs, build and notebook were unaffected.

The demand was circular rather than merely mistimed. The roadmap documents shipped work; nothing has shipped until the release pull request merges; the check required the page to name the version before it could honestly do so. Every minor release was unmergeable until someone hand-edited the roadmap into release-please's own branch.

The fix

The released versions now come from CHANGELOG.md, and the comparison is against the second-newest minor: the newest is whatever the open release pull request proposes, and the one before it is what a reader could already have installed.

Verified in both directions against the real #43 tree:

  • as it stands — manifest 0.4.0, changelog carrying 0.4.0, roadmap at v0.3.0 — 19 passed, where one failed before;
  • with a 0.5.0 section prepended, which is the original staleness one release on — fails again, naming the published minor.

The cost is that a stale page is caught one release after it goes stale rather than immediately. The docstring and the failure message both say so.

After this merges

#43 re-runs and should go green, and 0.4.0 can ship. The roadmap then owes a ## Shipped in v0.4.0 section for the smoothing work — a follow-up once the release is actually out, which is the order the page's own rule asks for.

Verification

pytest -m "not dataset and not notebook"    1175 passed, 15 skipped
ruff check / ruff format --check            clean
mypy                                        no issues in 65 source files
pytest tests/test_release_config.py         19 passed, on main and on the #43 tree

🤖 Generated with Claude Code

…roposed one

`test_the_roadmap_is_not_behind_the_released_minor` read the version from
`.release-please-manifest.json`, which the release pull request bumps as part
of preparing a release. On #43 that made the manifest say 0.4.0 while the
roadmap's newest section said 0.3.0, and the test failed — on the release pull
request itself, in six of the thirteen jobs.

The demand was circular. The roadmap documents shipped work, nothing has
shipped until the release merges, and the check required the page to name the
version before it could honestly do so. Every minor release was unmergeable
until someone hand-edited the roadmap into release-please's own branch.

It now reads the released versions from `CHANGELOG.md` and compares against
the second-newest minor: the newest is whatever the open release pull request
proposes, the one before it is what a reader could already have installed.

Verified in both directions against the real #43 tree: 19 passed where one
failed, and with a 0.5.0 section prepended — the original staleness, one
release on — it fails again. The cost is that a stale page is caught one
release after it goes stale rather than immediately, which the docstring and
the failure message both say.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
@read-the-docs-community

Copy link
Copy Markdown

Documentation build overview

📚 SpecMod | 🛠️ Build #34597387 | 📁 Comparing 1e34b4e against latest (2273219)

  🔍 Preview build  

2 files changed
± roadmap.html
± tutorial/specmod-tutorial.html

@sgjholt
sgjholt marked this pull request as ready for review September 16, 2026 21:53
@sgjholt
sgjholt merged commit 9b54544 into main Sep 16, 2026
14 checks passed
@sgjholt
sgjholt deleted the claude/roadmap-priorities-pe63kc branch September 16, 2026 21:53
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants