Overview - Nx-cloud-workflows

7 days30 days

Latest CI Pipeline Executions

Fuzzy

Succeeded
main
75a16482 fix(checkout): scrub credentials from git output (#129) * fix(checkout): scrub credentials from git output * fix(checkout): restrict credential scrubbing to repository url * fix(checkout): simplify repository url scrubbing
by Caleb Ukle
Succeeded
main
1743863c feat(templates): add universal-large launch template (#115) Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com> Co-authored-by: Victor Savkin <claude@anthropic.com> Co-authored-by: James Henry <james@henry.sc>
by Victor Savkin
Succeeded
main
efb2b21e fix(install-browsers): add idle + absolute timeouts and surface playw… (#126) * fix(install-browsers): add idle + absolute timeouts and surface playwright install failures - 3min idle timeout, 5min absolute timeout, SIGTERM then SIGKILL after a 5s grace - Inject DEBUG=pw:install into the child env so the extraction step actually logs progress - On failure, print the active node version, resolved playwright version, and the last 10 lines of stdout/stderr so the cause of a hang is visible Supersedes #121. * fix(install-browsers): kill the whole process group on timeout Previous attempt used child_process.exec + proc.kill('SIGTERM'), which only signals /bin/sh — the pnpm/node/playwright descendants keep running and the 'close' event never fires, so timers are never cleared and printFailureContext never prints. Switch to spawn('sh', ['-c', cmd], { detached: true }) so the shell is the leader of its own process group, then signal the whole group with process.kill(-pid, signal). * fix(install-browsers): drop redundant last-stdout/stderr dump on failure The captured chunks are already in the live log right above the failure context, so re-printing them just looks like a second run and makes the output harder to scan. Keep only the bits that aren't visible elsewhere: the kill reason, the active node version, and the resolved playwright version.
by Rares Matei
Succeeded
main
6ad7eb0e handle promise errors in install-browsers step (#120) Mirror the defensive handling already used in the checkout step so an unhandled rejection can no longer hang the install-browsers step: - add a top-level main().catch() that logs and exits non-zero - reject runCmdAsync on the child process 'error' event so a failed spawn settles the promise instead of hanging the await forever Co-authored-by: rarmatei <matei.rar@gmail.com>
by polygraph-...
Succeeded
main
48d56ef3 harden release workflow and validations workflow (#119) * harden release workflow and validations workflow - release workflow: add environment: release gate - release workflow: validate version and commit SHA input formats - release workflow: require cherry-pick commits to be reachable from origin/main - release workflow: route inputs through env vars and quote shell expansions - validations workflow: add explicit permissions: contents: read * use github app token for release pushes - mint installation token via actions/create-github-app-token - pass token to actions/checkout so subsequent git pushes use it - set git user to the app's bot identity - drop default workflow permissions from contents: write to contents: read
by Rares Matei