Repository navigation
release(mobile): push candidate tags as the operator, not the App - #9084
Conversation
Mobile candidate tags were created by a workflow holding the shared buzz-release-bot App key. Move mobile off that key so the only way to create a mobile-v* tag is a person on the mobile release team. scripts/mobile-release.sh candidate now verifies a dedicated mobile tag ruleset, creates the annotated tag locally, and pushes exactly that tag. A rejected push removes the local tag. The ruleset ID is a placeholder that fails loudly until an administrator records the real one. Remove the mobile-release-candidate workflow, its publisher script and test, and update the contract test, CI filters, and RELEASING.md. The desktop, relay, and chart lanes keep using the App unchanged. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Tom Brow <tomb@squareup.com>
…e branch ruleset Record the Restrict Mobile Tags ruleset ID, accept either refs/tags/mobile-* or refs/tags/mobile-v* as its include, and describe the operator requirement as being a bypass actor (the mobile release managers) instead of a placeholder team. The Mobile Release Branches ruleset now lists the same bypass actors, so drop its no-bypass requirements and keep checking the review and status check policy. Update the contract test stubs and RELEASING.md to match. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Tom Brow <tomb@squareup.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 6cfe9cc305
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
🔐 Codex Security Review
Review SummaryOverall Risk: NONE
FindingsNo concrete security, correctness, or reliability findings were identified. Notes
Generated by Codex Security Review | |
… fetches git push origin honors remote.origin.pushurl and url.*.pushInsteadOf, so a checkout fetching canonical block/buzz could publish the candidate tag into a fork. Require the effective push URL to equal the effective fetch URL. Signed-off-by: Tom Brow <tomb@squareup.com>
…low clone A --commit SHA verified as a source-branch ancestor can lie outside a shallow operator clone's history, so creating the local tag failed. Fetch that exact commit first. Signed-off-by: Tom Brow <tomb@squareup.com>
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 3c2193e472
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
…a GitHub An insteadOf rewrite can send a canonical-looking origin's fetch and push to a fork, so the git-side checks all pass against the fork. After pushing, require GitHub's block/buzz ref to hold the exact local annotated tag object, and remove the local tag otherwise. Signed-off-by: Tom Brow <tomb@squareup.com>
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: ddb7db3b82
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
git remote get-url --push lists only the first pushurl, while git push sends to every one, so a canonical first entry hid a fork after it. Compare all effective push URLs with the fetch URL. Signed-off-by: Tom Brow <tomb@squareup.com>
|
@codex review |
A concurrent fetch of main in the same checkout could replace FETCH_HEAD between the source fetch and its reads, so a stable candidate could be tagged on a main-only commit that then passed the ancestry check against itself. Read the tip with ls-remote and fetch that exact commit. Signed-off-by: Tom Brow <tomb@squareup.com>
|
@codex review |
wesbillman
left a comment
There was a problem hiding this comment.
🤖 Carl, an automated reviewer, commenting via Wes's GitHub account.
No blockers at cacdef0. All earlier findings are resolved. I did not submit an approval.
- Push sends extra tags: fixed.
--no-follow-tagsplus an explicit refspec, and the local tag is deleted only whenls-remoteconfirms origin doesn't have the candidate. Mongo reran the reproduction against the exact script. A stray annotated tag stays unpublished. A competing rc at a different commit is left untouched and the attempted local tag is kept. A lost push acknowledgement and a failed remote lookup both keep the local tag, and a confirmed-absent remote allows cleanup. - Ruleset
14378754: fixed. It now excludesrefs/tags/mobile-*(verified with a read-onlygh apicall), and the new preflight and its contract case cover that. - Docs: fixed. RELEASING.md now says release managers can tag stable-branch commits that were never reviewed.
cacdef0(snapshotting the source tip without relying on the shared FETCH_HEAD) looks correct. The new race test pins it.- I'm fine leaving the
insteadOfnit, given the test-harness reason and the GitHub-side OID check.
Merge criteria
- Remove Integration 4349119 (
buzz-release-bot) from26588677and19321162before merging, as the PR body commits to. Until then,release-rulesets.sh:4, RELEASING.md:108/:342 and the auto-tag header are inaccurate. - CI green on
cacdef0. It was still pending when I checked.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: cacdef001e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
…using it When a push reported failure but origin had the tag, the script told the operator to use it without the canonical block/buzz check, though an insteadOf rewrite could make origin a fork. Continue to the canonical checks instead. Signed-off-by: Tom Brow <tomb@squareup.com>
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: afd7e99ea4
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Co-authored-by: studio <dc2fc481343672df31a689840c09299972e704dac592e0da6f18486646a7ee3e@localhost> Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Signed-off-by: Tom Brow <tomb@squareup.com>
|
@codex review |
|
Codex Review: Didn't find any major issues. Chef's kiss. Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
* origin/main: (54 commits) release(mobile): push candidate tags as the operator, not the App (#9084) test(mobile): check posterless video previews renew the relay assertion (#9144) feat(relay): log WebSocket EVENT writes on the access target (#9154) perf(db): plan multi-value #d reads with their values (#9131) fix(mobile): confirm before starting a Huddle from the header (#9155) feat(cli): send NIP-FI evidence via credential chain and Buzz desktop (#8999) fix(mobile): find verified people who have no profile (#9086) feat(protocol): define compact membership discovery (BUZZ-615) (#9117) perf(db): hydrate search hits with exact event timestamps (#9126) perf(db): prune post-commit push enqueue by signed event timestamp (#9128) perf(db): use exact event timestamps for transactional deletion (#9130) perf(redis): index mesh ready records and share one Redis connection (#9113) fix(mobile): delay iOS message actions until a deliberate hold (#9049) feat(cli): user status expiry and status read (#8777) feat(cli): send a thread message to the channel (#8778) feat(relay): add kind 30180 custom emoji entry (#8783) feat(cli): read-only unread state (NIP-RS) (#8789) feat(cli): report a message to relay moderators (#8829) buzz-db: read explicit #h sets one channel at a time (#9110) feat(cli): list huddles and their participants (#8830) ... Signed-off-by: Taylor Ho <taylorkmho@gmail.com>
Mobile release candidates (
mobile-vX.Y.Z-rc.N) are currently created by themobile-release-candidate.ymlworkflow, using thebuzz-release-botApp key that also serves the desktop, relay and chart lanes. This PR takes mobile off that key. After it lands, the only way to create a mobile candidate tag is for a person on the mobile release team to push it.What changes
scripts/mobile-release.sh candidatekeeps all its existing checks: version, source branch, clean tree, canonical origin, ancestry and next rc number. It no longer dispatches a workflow. It now:refs/tags/mobile-*orrefs/tags/mobile-v*, and bypassable by the current user), and that the App-only general tag ruleset 14378754 excludesrefs/tags/mobile-*;--commitmissing from a shallow clone;--no-follow-tags, explicit refspec);block/buzzholds that exact tag object, then verifies the remote tag as before.A rejected push removes the local tag only when origin lacks it, so a retry starts clean. When the push reports failure but origin has the tag, the script keeps the local tag and still runs the canonical GitHub check, reporting success only if it passes. When GitHub can't confirm the tag, the local tag is kept and the message says how to reconcile it.
scripts/release-rulesets.sh:require_mobile_release_tag_rulesetreplaces the App-oriented check.MOBILE_RELEASE_TAG_RULESET_IDis26588677.Removed
.github/workflows/mobile-release-candidate.yml,scripts/publish-mobile-release-candidate.shand its test, and their CI references. UpdatedRELEASING.mdand the auto-tag workflow header comment.The desktop, relay and chart auto-tagging is unchanged.
Rollout
Tag ruleset 26588677 ("Restrict Mobile Tags") is active with the two mobile release managers as bypass actors. The App (Integration 4349119) is temporarily a bypass actor of 26588677 and 19321162 so the current
mainworkflow keeps working; it is removed from both right before this merges. General tag ruleset 14378754 now excludesrefs/tags/mobile-*; without that, rulesets stack and it would reject release managers' mobile tag pushes. The branch ruleset 19321162 now has the same bypass actors, sorequire_mobile_release_branch_rulesetno longer requires "no bypass"; it still checks the review and status check policy. Release managers must be bypass actors of both rulesets.Before merge
buzz-release-bot) from the bypass actors of rulesets 26588677 and 19321162.Validation
shellcheck -x scripts/release-rulesets.sh scripts/mobile-release.sh scripts/mobile-release-source.sh scripts/test-mobile-release-contract.sh: clean.scripts/test-mobile-release-contract.sh: passes. It runs against a local bare remote with a stubbedghand covers:refs/tags/mobile-v*also accepted as the include, each failing before any tag is created;push.followTags=truewith a stray local annotated mobile tag (exactly one ref published), a push that lands but exits non-zero (tag kept), and a GitHub API outage after the push (tag kept, rerun publishes the next rc);--commitpinning.scripts/test-release-ref-contract.shandscripts/test-desktop-release-candidate.sh: pass.scripts/release-rulesets.shandscripts/mobile-release-source.sh, then ranrequire_mobile_release_tag_ruleset(including the 14378754 exclusion check) andrequire_mobile_release_branch_ruleset mobile-v0.27. Both returned success. No tag was created.🤖 Generated with Claude Code