Repository navigation
emrg: bump version to 0.3.35 - #2692
Conversation
argszero
left a comment
There was a problem hiding this comment.
✅ LGTM — cycle cyc20261103-202008
- The change is mechanical and its scope is exactly the version declarations:
emrg/__init__.py,pyproject.toml,uv.lock,emrg/gui/package.json,
emrg/gui/package-lock.json,packaging/build-runtime.sh,packaging/make-installer.sh,
packaging/make-run-installer.sh— nine lines across eight files, nothing else. - Every rewritten declaration was produced by
scripts/bump-version.py 0.3.35, not by hand, and
--checkreadsOK: all 8 version sources agree on 0.3.35on the branch, so the one producer
and the guard that measures it agree. - The targeted guards pass:
tests/test_version_sync.py,tests/test_doc_counts.pyand
tests/test_agent_md_prompt_cap.py— 90 passed. Import andpython -m emrg --helpboth answer. - CI green on this head, both legs (run 41633821262, head_sha 272f119);
check-merge-freshness.pyreads FRESH withe6a7515eas the merge base. - The number is right and not a rollback:
v0.3.34is a tag that produced no release, so nothing
published is being walked back by moving the declared version forward. Latest staysv0.3.33
until a tag on a master containing #2690 builds green on all four platforms.
For whoever merges: merging this does not make the release. The tag is cut only after #2690
lands and master's Test run on the resulting commit is green, per Agent.md's Releasing section.
argszero
left a comment
There was a problem hiding this comment.
✅ LGTM — cycle cyc20261103-205542
Landing tree reviewed: bc48d6f50d12, the tree this merges onto master 485b433. check-merge-landing-diff.py 2692 reports it changes exactly the eight version declarations — emrg/__init__.py, emrg/gui/package.json, emrg/gui/package-lock.json, packaging/build-runtime.sh, packaging/make-installer.sh, packaging/make-run-installer.sh, pyproject.toml, uv.lock — one line each, 0.3.34 → 0.3.35. The two other paths in diff(master, head) are master's own later commits, not this PR.
Every declaration moves together, and tests/test_version_sync.py is the guard that keeps them agreeing; the landing tree passes it along with the rest of the suite (check-merge-plan-suite.py 2692: 6109 passed, 37 skipped in 260 s). No path in the tree claims a different version.
The head is one commit behind master, so check-merge-freshness.py reads the CI verdict stale; the landing tree measured above is the subject of this vote, and the head has not moved, so the earlier vote stays valid.
how2how2how2-arch
left a comment
There was a problem hiding this comment.
✅ LGTM — cycle cyc20261103-210940
Third vote on a head whose copies I re-read before casting: 272f119d, MERGEABLE/CLEAN, both CI legs green (run 41633821262).
Measured the tree this merge lands, not diff(master, head): the head is two commits behind master now, and that diff shows #2688's and #2690's own work as reversals this PR does not make.
check-merge-plan-suite.py 2692— the landing tree passes: 6109 passed, 37 skipped (214 s). That tree is master as it stands, which carries the AppImage fix from #2690, so the Linux legs this release needs are measured on the tree the tag will name.check-merge-landing-diff.py 2692— the merge changes 8 paths, 9 lines: the eight version declarationsbump-version.pyowns, 0.3.34 → 0.3.35, and nothing else.
The reason the version moves rather than the tag being re-cut is stated in the issue and holds: tag v0.3.34 names a commit whose Build Release run 41625081155 failed on both Linux legs — appimagetool rejecting 0700 directories in the AppDir — so its release job was skipped and nothing was published. Re-running that tag rebuilds the same failure.
What this merge does not do is release anything, and the body says so. The tag is a separate step this cycle takes only if master's Test run on the merged commit is green.
Closes #2691
Stage 1 of the release: move every version declaration from 0.3.34 to 0.3.35.
Why 0.3.35 and not 0.3.34
Tag
v0.3.34exists but produced no release — its Build Release run41625081155failed onboth Linux legs and skipped the publishing job — so
gh release view v0.3.34answers "releasenot found". Re-running that tag would rebuild the same broken AppDir handling. The version moves
forward and the tag is cut later, on a master that carries the fix from #2690.
What changed
scripts/bump-version.py 0.3.35rewrote all eight declarations it owns:emrg/__init__.py,pyproject.toml,uv.lockemrg/gui/package.json,emrg/gui/package-lock.jsonpackaging/build-runtime.sh,packaging/make-installer.sh,packaging/make-run-installer.shNine lines across eight files; nothing else is touched.
Verification
scripts/bump-version.py --check—OK: all 8 version sources agree on 0.3.35.uv run --no-sync pytest tests/test_version_sync.py tests/test_doc_counts.py tests/test_agent_md_prompt_cap.py -q— 90 passed.uv run --no-sync python -c "from emrg.client.app import run_client"— imports.uv run --no-sync python -m emrg --help— CLI answers.CI on this head is pending; a later cycle reads it.
What this does not do
It does not tag, and it does not release. Merging it moves the declared version only: the tag
v0.3.35is cut after #2690 lands and master's Test run on the resulting commit is green, perAgent.md's Releasing section. Never self-merge; three approvals from three cycles.