From 5ff18b644493c5e79373a785e873988ded4ac006 Mon Sep 17 00:00:00 2001 From: Claude Date: Fri, 19 Jun 2026 10:31:40 +0000 Subject: [PATCH 1/2] fix(ci): always dispatch configurator deploy on release MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The downstream deploy was gated on commit_push.pushed == 'true', so when release-it already synced version artifacts to main before the tag fired, sync-main found nothing to commit and silently skipped the deploy — leaving the configurator stuck on the previous version. Remove the condition so deploy-configurator.yml and publish-dist.yml are dispatched on every release, whether or not main needed a version artifact update. Co-Authored-By: Claude Sonnet 4.6 Claude-Session: https://claude.ai/code/session_01EbKb1xhzQ6ctoq5T6dC6BM --- .github/workflows/release.yml | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index e282715b..cbb7688d 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -201,15 +201,15 @@ jobs: echo "pushed=true" >> "$GITHUB_OUTPUT" fi - name: Trigger downstream deploys - if: steps.commit_push.outputs.pushed == 'true' env: GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} run: | - # GITHUB_TOKEN pushes do not emit push events, so deploy-configurator - # and publish-dist would never see the new version. Dispatch them - # explicitly — GITHUB_TOKEN CAN trigger workflow_dispatch (actions:write - # permission granted above). Each call is independent so one failure - # does not prevent the other from running. + # Always redeploy on every release, regardless of whether main needed + # a version-artifact commit. When release-it already pushed the bump + # before the tag fired, commit_push.pushed is false but the deploy + # must still run so the configurator reflects the new version. + # GITHUB_TOKEN CAN trigger workflow_dispatch (actions:write granted above). + # Each call is independent so one failure does not prevent the other. gh workflow run deploy-configurator.yml --ref main echo "Dispatched deploy-configurator.yml" gh workflow run publish-dist.yml --ref main From 834c80737747d6e4ccfd128cf422012cb11b8591 Mon Sep 17 00:00:00 2001 From: Claude Date: Fri, 19 Jun 2026 10:39:50 +0000 Subject: [PATCH 2/2] fix(ci): ensure both downstream dispatches run even if one fails Both gh workflow run calls shared a single sequential run block, so a failure in deploy-configurator.yml would prevent publish-dist.yml from ever being dispatched. Use explicit error tracking so both dispatches always execute, then exit with a non-zero code if either failed. Co-Authored-By: Claude Sonnet 4.6 Claude-Session: https://claude.ai/code/session_01EbKb1xhzQ6ctoq5T6dC6BM --- .github/workflows/release.yml | 20 ++++++++++++++++---- 1 file changed, 16 insertions(+), 4 deletions(-) diff --git a/.github/workflows/release.yml b/.github/workflows/release.yml index cbb7688d..c7e43a2d 100644 --- a/.github/workflows/release.yml +++ b/.github/workflows/release.yml @@ -210,7 +210,19 @@ jobs: # must still run so the configurator reflects the new version. # GITHUB_TOKEN CAN trigger workflow_dispatch (actions:write granted above). # Each call is independent so one failure does not prevent the other. - gh workflow run deploy-configurator.yml --ref main - echo "Dispatched deploy-configurator.yml" - gh workflow run publish-dist.yml --ref main - echo "Dispatched publish-dist.yml" + fail=0 + if gh workflow run deploy-configurator.yml --ref main; then + echo "Dispatched deploy-configurator.yml" + else + echo "::error::Failed to dispatch deploy-configurator.yml" + fail=1 + fi + + if gh workflow run publish-dist.yml --ref main; then + echo "Dispatched publish-dist.yml" + else + echo "::error::Failed to dispatch publish-dist.yml" + fail=1 + fi + + exit "$fail"