You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Upstream cluesmith/codev#1331 by @waleedkadous fixes pr-search so merged PRs are found, not just open ones. It is open and unmerged upstream.
This fork is on GitHub, so if the bug is real in scripts/forge/github/pr-search.sh, we have it. A PR search that silently misses merged PRs returns a wrong answer at exit 0, which is the failure mode that costs the most to notice.
Why this is its own issue
#12 implements pr-search for gitea and deliberately does NOT touch the github or gitlab scripts. That was the right call: editing cluesmith#1331 diff in our fork guarantees a conflict on every upstream sync for a file we did not otherwise need to change.
Adopting it is the pattern we already used for our own stranded PRs (cluesmith#1146, cluesmith#1458): merge the upstream branch into the fork now, get the fix today, and have the eventual upstream merge resolve as a no-op.
Confirm the bug reproduces against this repo before changing anything. If gh already returns merged PRs for our query shape, there is nothing to adopt.
Merge the upstream branch rather than hand-copying the diff, so the future sync is clean.
Do not
Do not re-derive the fix by hand. That produces a different diff that conflicts with the real one when it lands.
Upstream
cluesmith/codev#1331by @waleedkadous fixespr-searchso merged PRs are found, not just open ones. It is open and unmerged upstream.This fork is on GitHub, so if the bug is real in
scripts/forge/github/pr-search.sh, we have it. A PR search that silently misses merged PRs returns a wrong answer at exit 0, which is the failure mode that costs the most to notice.Why this is its own issue
#12 implements
pr-searchfor gitea and deliberately does NOT touch the github or gitlab scripts. That was the right call: editing cluesmith#1331 diff in our fork guarantees a conflict on every upstream sync for a file we did not otherwise need to change.Adopting it is the pattern we already used for our own stranded PRs (cluesmith#1146, cluesmith#1458): merge the upstream branch into the fork now, get the fix today, and have the eventual upstream merge resolve as a no-op.
Work
pr-searchbreaksspawn-worktree.tsin:body query, so callers that need open-only must sayis:openexplicitly. Adopting the code without that change would introduce a bug rather than fix one.ghalready returns merged PRs for our query shape, there is nothing to adopt.Do not
Do not re-derive the fix by hand. That produces a different diff that conflicts with the real one when it lands.