Repository navigation
chore(release): promote version - #149
yia-mw-agent wants to merge 3 commits into
Conversation
timeleft--
left a comment
There was a problem hiding this comment.
Current head 12e001a is now conflicting with main after separately merged #148. Please refresh this SAME PR through a normal integration, recalculate the suffix-only promotion against actual main with the repository updater, and report the exact non-version delta from Monday candidate b2e7f0b. Retain the existing administrative merge restriction; this is source repair only, no merge/tag/publication/install. Reuse valid packaging evidence and rerun affected current-source checks; do not silently promote untested newer product or repeat a semantic base increment.
|
Addressed review 5533620027 on the same PR and branch (fast-forward push 12e001a..d5ed119; no rebase or force-push).
|
|
Oct16 exact-head re-review at d5ed119: the requested source refresh and explicit delta accounting are addressed. Independent current-main version-policy check passes; diff from main is only the three configured suffix-removal mirrors. Independent Python3.13 frozen-environment tests for release channels, affected MCP context/server instructions and packaged MCP pass: 100 tests. Hosted current-head checks also pass. This does not establish installed actual-use evidence for the newer 0.8.5rc1 product bytes; the installed weekly candidate remains 0.8.4rc1. Existing administrative merge restriction is unchanged. No merge, publication or installation performed. |
timeleft--
left a comment
There was a problem hiding this comment.
Independent exact-head approval d5ed119 against f1c89ae. Three configured version mirrors only remove the0.8.5rc1 suffix; no semantic base increment or product delta versus currentmain. Reuse still-valid independent current-baseVersionpolicy and100 packaged/release-channel/MCP instruction tests from Oct16, unchanged head. Oct19 exact f1c89ae postmerge Tests38627364952 now SUCCESS bothJJmatrices after failed-jobs-only recovery of pre-test API403. Immutable v0.8.5rc1 workflow39257461467 SUCCESS; public GitHub/PyPI wheel+sdist bytes match; actual installed candidate all66payloadfiles match and MCPinitialize/17tools/guide/isolatedsave-recall/governed-isolation pass. Existing current candidate rollback0.8.4rc1 wheel/dependency/runtime proof retained. This is source approval only; existing147 administrative merge restriction and Wednesday promotion timing remain unchanged. No stable merge/publication/deployment authorized by this review alone.
Promotes the current main candidate to stable (refs #147). Merge when clean: no (First Officer owns Wednesday 16:00 promotion pass).
Refresh after review (changes requested on 12e001a): main advanced b2e7f0b -> f1c89ae (#148, which also bumped the candidate 0.8.4rc1 -> 0.8.5rc1), so the original 0.8.4 promotion conflicted. This branch now has a normal merge of origin/main (66cc946, no force-push/rebase) and a recalculated suffix-only promotion (d5ed119).
Change:
python3 scripts/mw-version.py update --base origin/main --title "chore(release): promote version"->repo: 0.8.5-rc.1 -> 0.8.5. Only the 3 configured mirrors differ from main f1c89ae (pyproject.toml, src/fava_trails/init.py, uv.lock), 1 line each:0.8.5rc1->0.8.5. No semantic base increment in this PR; no workflow/dependency/pin changes.Exact non-version delta from Monday candidate b2e7f0b (v0.8.4rc1) to this head: all from merged #148 (startup-instruction trim, resolves #146), nothing else:
git diff --numstat origin/main HEADshows exactly the 3 version files;git diff --numstat b2e7f0b HEADshows those 3 plus the 8 files above.Evidence gap to review: this promotes #148 product bytes to stable. 0.8.5rc1 has not been tagged, published, or installed (only v0.8.4rc1 exists), so those bytes have only CI/test evidence, not multi-day installed MCP evidence like the Monday candidate.
Verification on head d5ed119 (jj 0.45.1, uv 0.12.3, JJ_CONFIG=/dev/null):
Worker does not merge, tag, publish, install or restart.