release(1.6.1): accept setuptools sdists and build the pure-python wheel once - #1592
Merged
Merged
Conversation
…eel once Release run 35958433737 published every crate and then failed in the release job's Python asset verification: a setuptools sdist carries a second PKG-INFO under *.egg-info, and a pure-python build yields one universal wheel that three OS legs collapse into. atm-core's 1.6.0 copy of release_artifacts.py carried both fixes (sc-publish#74, #75); the phase-bc kit re-render dropped them. Match only the root-level sdist PKG-INFO again, and build hermes-atm on a single wheel leg so the expected and collected wheel counts agree. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
rand-lee
approved these changes
Sep 24, 2026
This was referenced Sep 24, 2026
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
2 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Release run 35958433737 published all 15 crates to crates.io and then failed in the
releasejob at Verify collected Python distributions:A standard setuptools sdist contains the root
PKG-INFOand a second one undersrc/hermes_atm.egg-info/, so the kit's check rejects every well-formed setuptools sdist. Behind it, the same step would then have failed on the wheel count: the manifest lists three OS legs for hermes-atm, but a pure-python build yields onepy3-none-anywheel whose identical filename deduplicates when the legs are collected.Cause. atm-core's 1.6.0 copy of
release_artifacts.pycarried both release-branch fixes (sc-publish#74 and #75, from v1.4.4). The phase-bc kit re-render (b381e09) replaced that copy with upstream, which never took them. 1.6.1 is the first release on the re-rendered copy. Same story as #1591.Fix, consumer-side only, kit untouched.
.github/scripts/release_artifacts.py: match only the root-level<rootdir>/PKG-INFOin_python_distribution_name_from_sdist, the 1.6.0 behaviour (sc-publish#74), kept to two lines so the script stays at the 1000-line ceiling its own test enforces.release/publish-artifacts.toml: hermes-atm builds on one wheel leg (ubuntu-latest). A pure-python wheel is platform-independent, so three legs built the same file three times; with one leg the expected and collected counts agree without the atm read hides valid inbox messages when message_id is absent #75 code.Verified locally against the artifacts downloaded from run 35958433737:
validate-manifestpasses, the wheel matrix keeps three legs for the maturin distributions and one for hermes-atm, andverify-python-release-assetspasses on the patched script where the current one fails, and the kit script tests pass (293 passed, 13 skipped). The publisher retries the root release against this commit;gate-and-tagreuses tag v1.6.1 and the crates step skips the 15 already-published crates.🤖 Generated with Claude Code