Repository navigation
release 0.10.0: the version the A0 work goes out under - #27
Merged
Merged
Conversation
0.9.0 was released on 2026-10-19 (tag v0.9.0, db7057c, published to NuGet) and the A0 work landed on top of the same number: yesterday's release prep edited the 0.9.0 section in place, moving its date to 2026-10-29, rewriting its intro to name macula-go v0.29.0, and appending two subsections to a release that had already shipped. This restores that section to exactly what the tag's own CHANGELOG says — verified by comparing the block at v0.9.0 against the block here, byte for byte — and puts the new material in a `## [0.10.0] - 2026-10-29` section above it, with the A0 re-seal as `### Changed`, and the honesty text ("what this release does not change") carried over unchanged because it is still true. Macula.csproj's Version is 0.10.0, which is what release.yml's guard reads (version-root src/Macula) and compares the tag against. The README gains the same paragraph, and `dotnet add package Macula` resolves 0.10.0. Raf's ruling, 2026-10-29: dotnet goes out as 0.10.0. Co-Authored-By: Claude Code <noreply@anthropic.com>
Mercurius r=1, two findings, both his. 1. release.yml pushed to NuGet with --skip-duplicate, so a version that is already published is skipped and the job reports success: a release that published nothing looks exactly like one that published everything. That is how 0.9.0 could be prepared a second time on 29 Oct under a number that went live on 19 Oct. Removed, with the reason above the step. A re-run after a partial publish now fails loudly, which is what we want. The sweep across the other repositories stays in macula-codebase-checks#37. 2. The README's 19 Oct banner said `dotnet add package Macula` resolves 0.9.0 directly under the new paragraph saying it resolves 0.10.0. Past tense now: "resolved 0.9.0 until 0.10.0". Co-Authored-By: Claude Code <noreply@anthropic.com>
Supervisor's instruction: the --skip-duplicate removal is the same defect class as this week's miss — a duplicate version "succeeding" without publishing — so it must not hide inside a version bump. It now has its own `### Changed` entry in the 0.10.0 section, saying what the step does now, what it did before, and that a release which published nothing used to look like one that published everything (macula-codebase-checks#37). Co-Authored-By: Claude Code <noreply@anthropic.com>
…ot happen Mercurius r=2b, one exact clause, in two places. Both said the silent skip "is what let 0.9.0 be prepared a second time on 29 Oct". That claims a cause that never operated: no v0.9.0 publish was attempted that day, and the second preparation happened at review, before any publish job ran. The flag was never reached — it would only have hidden the mistake at publish time. Both now say what would have happened rather than what did: had 0.9.0 been tagged again on 29 Oct, under a number live since 19 Oct, the job would have gone green and published nothing. Co-Authored-By: Claude Code <noreply@anthropic.com>
This branch was successfully deployed
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
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.
Refs the 29 Oct release batch.
Purpose (reliability): the A0 work reaches users under a version number that has not already been published, and a release that publishes nothing stops reporting success.
0.9.0 was released on 2026-10-19 (tag
v0.9.0, db7057c, published to NuGet at 2026-10-19T18:29Z) and the A0 work landed on top of the same number. Yesterday's release prep then edited that shipped section in place: date moved to 2026-10-29, intro rewritten to name macula-go v0.29.0, two subsections appended to a release that had already gone out.This PR
The version.
src/Macula/Macula.csprojon<Version>0.10.0</Version>— the filerelease.yml's guard reads through version-rootsrc/Macula— and a README paragraph withdotnet add package Macularesolving 0.10.0.The record. The
## [0.9.0] - 2026-10-19section is restored byte for byte to what the tag's own CHANGELOG holds (compared programmatically, and Mercurius reproduced it independently), and the new material goes into## [0.10.0] - 2026-10-29above it — the A0 re-seal as### Changed, and the "what this release does not change" text carried over unchanged because it is still true.The release workflow (named as its own change, not folded into a version bump).
release.yml's NuGet push no longer passes--skip-duplicate. Before this, a version already on NuGet was skipped and the job reported success: a release that published nothing looked exactly like one that published everything. That is the same defect class as this week's miss — three SDK versions prepared under numbers published nine days earlier — and it is what would let it happen again silently. It is a### Changedentry in the CHANGELOG as well as a line in this PR body, on the Supervisor's instruction, so a reader of the release notes sees a release-path change rather than a version bump. The reason is written above the step in the workflow so it is not put back. The sweep across the other repositories stays in macula-codebase-checks#37.No library code changes.
Registry check, before review
v0.10.0does not exist locally or on origin;Rounds
r=1 findings from Mercurius, both fixed: the README's 19 Oct banner contradicted the new "resolves 0.10.0" line (now past tense), and this workflow change, which he asked to fold in.
Raf's ruling, 2026-10-29: dotnet goes out as 0.10.0, tag after green main CI.
🤖 Generated with Claude Code