v0.15.3 post-publication: record digests, flip state, sweep pointers - #367
Merged
Conversation
v0.15.3 is published. PyPI reports it as the latest version, the GitHub Release is stable rather than a draft, and the project page now shows the AliceMemory repository URLs, which is the reason this release existed. Digests recorded in docs/release/v0.15.3-checksums.txt, verified from three sources that agree rather than copied from one: the PyPI JSON API advertises them, downloading the GitHub Release assets and hashing them locally produces the same values, and the published SHA256SUMS asset is a third reading. ae46ee24...c33d538f alice_memory-0.15.3-py3-none-any.whl acea1ccd...aef1932f alice_memory-0.15.3.tar.gz The post-rename publish path is now proven end to end. The trusted publisher scoped to the pypi environment authenticated on first attempt, which was the main thing this release was cutting early to find out. One process correction, and it is a correction rather than an addition. The rule recorded after v0.15.2 was "the release gate runs twice, from the branch and from the tag". That is necessary and it is not sufficient. The publish job requires every one of twelve exact-SHA checks to be complete AND successful on the release commit, and the semantic gate is only one of them. This release dispatched publish while Integration tests was still in progress on the release commit, and the gate correctly refused: identity verification failed with "latest required exact-SHA check did not succeed", every publish step skipped, nothing uploaded. The cost was one wasted dispatch. The rule is now to enumerate all twelve and confirm each is completed/success before dispatching.
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
There was a problem hiding this comment.
Your trial has ended. Reactivate Greptile to resume code reviews.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
v0.15.3is published. PyPI reports it as latest, the GitHub Release is stable, and the project page now shows the AliceMemory URLs, which is the reason this release existed.Digests, three agreeing sources
PyPI JSON API, GitHub Release assets downloaded and hashed locally, and the published
SHA256SUMSasset all agree.The post-rename publish path is proven
The trusted publisher scoped to the
pypienvironment authenticated on first attempt. That was the main thing this release was cutting early to find out, and it is why the publish action bump (#349) was held.The old
samrusani/AliceBotpublisher can now be deleted on PyPI. It carriesEnvironment name: (Any)and the owner id is unchanged by a rename, so until it is gone, anyone creating a repo at the freed old name with apublish-pypi.ymlcould mint a publishing token foralice-memory.Process correction
The rule recorded after v0.15.2 was "the gate runs twice, branch then tag". Necessary, not sufficient.
The publish job requires all twelve exact-SHA checks to be complete and successful on the release commit; the semantic gate is one of them. This release dispatched publish while
Integration testswas still in progress, and the gate correctly refused withlatest required exact-SHA check did not succeed. Every publish step skipped, nothing uploaded, one wasted dispatch.The rule is now: enumerate all twelve and confirm each is
completed/successbefore dispatching.