Repository navigation
release: prepare v0.1.0-rc.2 - #93
Merged
Merged
Conversation
v0.1.0-rc.1 was tagged on 2026-11-09 and never released: three faults in the release workflow kept its run from the signing jobs, and a tag's run uses its own commit's workflow files, so no re-run can release it. rc.2 is the same tree with those fixes (#91, #92). - The workspace version, the lockfile's workspace crates, the Helm chart, its image tag, the enclosure manifests and the air-gap bundle example name 0.1.0-rc.2. - docs/releases/v0.1.0-rc.2.md: rc.1's notes, with why there is a second candidate and how this one is cut. - CHANGELOG: the unreleased entries are 0.1.0-rc.2's, under a preface; 0.1.0-rc.1 is recorded as tagged and never released. - README points at the rc.2 notes and says no release is published yet. - claims.toml: changelog anchors moved by the preface, and the rc.2 notes cited beside rc.1's; CLAIMS.md re-rendered. NUMBERS.md regenerated. - clients/shells/audit-waivers.txt: the owner's waiver for 0.1.0-rc.2, as for rc.1; GA_READINESS row B names both pre-releases.
phillips1707
approved these changes
Nov 10, 2026
phillips1707
left a comment
There was a problem hiding this comment.
Approving. This is the rc.2 preparation, and it does what the description says.
Checked
- Commit
117896ahas a verified signature and sits ond60346d(#92), the tip ofmain. Sincev0.1.0-rc.1(3568b9b), the only changes are #91, #92 and this commit, so "rc.2 is rc.1 with the workflow fixes" holds. - Version.
Cargo.toml, all 70 workspace entries inCargo.lock, the chart'sversion,appVersionand image tag, the five images inenclosure.yamland the air-gap README all move from0.1.0-rc.1to0.1.0-rc.2. Every0.1.0-rc.1left in the tree is an example, a test fixture, a historical record or the waiver for rc.1 itself. - Waiver.
audit-waivers.txthas a0.1.0-rc.2line matchingCargo.toml's version, so the shell-audit gate can pass and write theverdict#92 reads.shell-audit-waiver.test.shandshell-audit.test.shpass locally. - Claims anchors. I checked all 61 changed anchors mechanically. Each
CHANGELOG.mdline moved by +10 holds the same text it did onmain, and each newv0.1.0-rc.2.mdanchor holds the same text as its rc.1 counterpart. - Release notes. The rc.2 notes are rc.1's, with "Why a second candidate" added and the pre-tag owner checklist replaced by "How this candidate is cut".
- Local checks.
numbers.sh --checkandgate-record.sh --checkpass locally. I couldn't runclaims.sh(it needs bash 4) orcargo metadataon this machine; "scripts, provenance, secrets, manifests" passes in CI.
Not blocking
build and test,reproducible buildandrelease gate, dry runare still running as I approve.- As the description says,
docs/launch/and the two decks still say rc.1. - Next: the gate at the merge commit, the records-only PR, then
tag-release.shforv0.1.0-rc.2.
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.
Prepares
v0.1.0-rc.2.v0.1.0-rc.1was tagged on 2026-11-09 at3568b9band never released. Three faults in the release workflow kept its run from the signing jobs. They are fixed by #91 and #92. A release tag cannot be moved, and a tag's run uses the workflow files of its own commit, so no re-run of rc.1 can release it. rc.2 is the same tree with those fixes.What changes
Cargo.toml, the workspace crates inCargo.lock, the Helm chart'sversion,appVersionand image tag, the five images indeploy/k8s/enclosure.yaml, and the air-gap bundle README's example.cargo metadata --locked --offlinepasses, so the lockfile needs nothing else.docs/releases/v0.1.0-rc.2.md. rc.1's notes, with a section on why there is a second candidate. The closing section now says how this candidate is cut.CHANGELOG.md. The unreleased entries become 0.1.0-rc.2's, under a short preface. 0.1.0-rc.1 is recorded as tagged and never released. A new, empty Unreleased section sits on top.README.md. Says no release is published yet, and points at the rc.2 notes. Line counts are unchanged.clients/shells/audit-waivers.txtrecords the owner's decision for0.1.0-rc.2, for the same reason as rc.1: no shell has a buildable project yet. GA_READINESS row B names both pre-releases. The waiver file is covered by the gate, so it lands here, before the gate runs.claims.tomlmove with it. Each one was checked to point at the same text. The rc.2 notes are cited beside rc.1's.CLAIMS.mdis re-rendered andNUMBERS.mdregenerated.Left as it is
docs/ops/UPGRADING.md, and the test fixtures.docs/launch/,docs/design/site/) still says rc.1, including the two.pptxdecks. Those directories are outside the gate and can follow separately.Still needed before the tag
tag-release.sh, push the tag, and a second maintainer approves each signing job.Checks
claims.sh --check --strict,numbers.sh --checkandgate-record.sh --check: passtag-release.sh --check: ready except the gate record, which does not cover this commit yet, and the uncommitted treecargo metadata --locked --offline: passes with the bumped lockfiletools/ci/tests/*.test.sh: 35 of 36 pass on a Windows checkout.mg-install.test.shfails five symlink and file-mode cases there, the same as on main. CI runs it on Linux.