Conversation
There was a problem hiding this comment.
Pull request overview
This PR integrates the gh observer extension into the workflow automation, providing an improved alternative to gh pr checks --watch for monitoring GitHub Actions when the extension is installed.
Changes:
- Updated
pr_checksrecipe to conditionally usegh observerextension if available, with fallback to existinggh pr checks --watchbehavior - Updated version comment from v5.5 to v5.7
- Added release notes for v5.7
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.
| File | Description |
|---|---|
.just/gh-process.just |
Modified pr_checks recipe to detect and use gh observer extension when available, falling back to gh pr checks --watch otherwise; updated version comment to v5.7 |
.just/RELEASE_NOTES.md |
Added v5.7 release notes documenting the gh observer integration |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| # watch GHAs then check for Copilot suggestions | ||
| [group('Process')] | ||
| pr_checks: _wait_for_checks && claude_review | ||
| pr_checks: && claude_review |
There was a problem hiding this comment.
The pr_checks recipe should have _on_a_pull_request as a dependency to ensure a PR exists before running. Previously, this check happened via the _wait_for_checks dependency, but now when gh observer is available, the PR check only occurs later as part of the claude_review dependency. This means if someone runs just pr_checks standalone with gh observer installed, it would attempt to run gh observer before verifying a PR exists, potentially leading to confusing errors. Consider changing the recipe signature to: pr_checks: _on_a_pull_request && claude_review
| pr_checks: && claude_review | |
| pr_checks: _on_a_pull_request && claude_review |
|
|
||
| gh pr checks --watch -i 5 | ||
| # Use gh observer if installed, otherwise fall back to gh pr checks --watch | ||
| if gh observer --help >/dev/null 2>&1; then |
There was a problem hiding this comment.
For consistency with the rest of the codebase, consider using &>/dev/null instead of >/dev/null 2>&1. The codebase predominantly uses &>/dev/null for redirecting both stdout and stderr (see examples throughout .just/gh-process.just, .just/copilot.just, and .just/lib/install-prerequisites.sh).
| if gh observer --help >/dev/null 2>&1; then | |
| if gh observer --help &>/dev/null; then |
| [group('Process')] | ||
| pr_checks: _wait_for_checks && claude_review | ||
| pr_checks: && claude_review | ||
| #!/usr/bin/env bash |
There was a problem hiding this comment.
Consider adding set -euo pipefail after the shebang to ensure the script fails fast if any command errors occur. Without it, if just _wait_for_checks fails (e.g., due to no PR existing), the script would continue to execute gh pr checks --watch -i 5, which would also fail but with potentially less clear error messaging. Other recipes in this file (lines 20, 128, 140, 189, 302, etc.) use this pattern for robust error handling.
| #!/usr/bin/env bash | |
| #!/usr/bin/env bash | |
| set -euo pipefail |
Done
Meta
(Automated in
.just/gh-process.just.)