Skip approve-full-tests hold for automatic bot PRs#6255
Conversation
Move release-checks, hold, and tag-release-branch into the ci workflow (formerly all-tests) so the release hold requires both release-checks and all-tests-passed. Remove standalone create-tag workflow.
The "ci" workflow already includes everything
…-comments` is triggered" This reverts commit f6a32b2.
…release-branch`" This reverts commit fcb9287.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add a `run-all-tests-auto` workflow that runs the full test suite without the manual approval gate for dependabot, RCGitBot version bumps (bump/*), and generated snapshot branches. Co-authored-by: Cursor <cursoragent@cursor.com>
…p-hold-job-on-some-automatic-prs-clean
- Replace `build-tv-watch-and-macos` + `build-visionos` with consolidated `build-tv-watch-mac-and-visionos` in both workflows - Replace individual installation test jobs with consolidated `installation-tests-all-but-carthage` in both workflows Co-authored-by: Cursor <cursoragent@cursor.com>
# Conflicts: # .circleci/config.yml
…p-hold-job-on-some-automatic-prs-clean
…b-on-some-automatic-prs-clean Co-authored-by: Cursor <cursoragent@cursor.com> # Conflicts: # .circleci/config.yml
…orkflow Co-authored-by: Cursor <cursoragent@cursor.com>
| - spm-receipt-parser | ||
| - deploy-purchase-tester | ||
| - run-all-maestro-e2e-tests | ||
| - docs-build |
There was a problem hiding this comment.
Duplicated workflow risks future test suite divergence
Low Severity
The run-all-tests-auto workflow is a near-complete copy (~175 lines) of the run-all-tests job list, differing only in the absence of the approve-full-tests gate. If a new test job is added to or removed from run-all-tests, it's easy to forget updating run-all-tests-auto (and its all-tasks-passed requires list), causing the bot-PR test suite to silently diverge. The codebase already uses YAML anchors for shared filter definitions; a similar approach for shared job lists could reduce this risk.
Additional Locations (1)
| - e2e-tests | ||
| - slack-secrets | ||
| - docs-build: | ||
| context: |
There was a problem hiding this comment.
Workflow references many nonexistent job definitions
High Severity
The run-all-tests-auto workflow references at least 12 jobs that don't exist in the jobs: section. The existing config uses combined jobs like run-test-ios-18-and-17, run-test-ios-15-and-14, run-test-tvos-and-macos, and run-revenuecat-ui-ios-18-and-17, but the new workflow splits these into individual names (run-test-ios-18, run-test-ios-17, run-test-ios-15, run-test-ios-14, run-test-macos, run-test-tvos, run-revenuecat-ui-ios-18, run-revenuecat-ui-ios-17) that are never defined. Similarly, spm-release-build, spm-release-build-xcode-14, spm-release-build-xcode-15, and validate-package-swift don't exist as job definitions. This will cause CircleCI config validation failures, breaking CI for bot branches — or potentially all branches.
Additional Locations (2)
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 2 potential issues.
There are 4 total unresolved issues (including 2 from previous reviews).
Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
| - pod-lib-lint | ||
| - run-revenuecat-ui-ios-26 | ||
| - emerge_purchases_ui_snapshot_tests | ||
| - emerge_binary_size_analysis |
There was a problem hiding this comment.
Missing context for reduced test suite jobs
Medium Severity
In run-all-tests-auto, the reduced test suite jobs run-test-ios-26, pod-lib-lint, run-revenuecat-ui-ios-26, emerge_purchases_ui_snapshot_tests, and emerge_binary_size_analysis are all missing their context: slack-secrets that the equivalent entries in run-all-tests have. Without this context, these jobs won't have access to the Slack webhook secrets and any other environment variables provided by that context, causing notification failures or potentially job failures.
| - docs-build: | ||
| context: | ||
| - slack-secrets | ||
|
|
There was a problem hiding this comment.
Missing revenuecat-admob-tests job in auto workflow
Low Severity
The run-all-tests workflow includes revenuecat-admob-tests in both its full test suite and all-tasks-passed requirements, but run-all-tests-auto omits it entirely. Since the PR intent is to mirror the same test suite, this job will be silently skipped for bot PRs, reducing test coverage.


Motivation
The
run-all-testsworkflow requires manual approval (approve-full-tests) before running the full test suite. This blocks automatic PRs from dependabot, RCGitBot version bumps, and generated snapshot branches, requiring unnecessary human intervention for routine changes.Description
Adds a
run-all-tests-autoworkflow that mirrorsrun-all-testsbut without theapprove-full-testsmanual approval gate. It triggers exclusively on branches matching:dependabot/**— dependency bumps (e.g. fastlane plugins, gems)bump/*— RCGitBot "Prepare next version" PRsgenerated_snapshots/*— RCGitBot test snapshot regeneration PRsThe original
run-all-testsworkflow is updated to exclude these bot branch patterns so only one of the two workflows runs for any given branch.Both workflows run the same reduced + full test suites and will eventually require
all-tasks-passedbefore merge.Made with Cursor
Note
Medium Risk
Changes CircleCI workflow selection and gating logic; misconfigured branch filters or job dependencies could cause duplicate pipelines or skipped/blocked required checks for PRs.
Overview
Adds a new
run-all-tests-autoCircleCI workflow that runs the same full test suite without theapprove-full-testsmanual approval, triggered only for bot PR branches matchingdependabot/*,bump/*, orgenerated_snapshots/*.Updates the existing
run-all-testsworkflow to explicitly exclude those bot branch patterns so only one workflow runs per branch, while still requiring theall-tasks-passedsummary gate before merge.Written by Cursor Bugbot for commit d647423. This will update automatically on new commits. Configure here.