Thanks a lot for the great work! This will make managing our PRs more efficient. It's impressive to see the number of PRs merged over the weekend.
Chen Li On Sun, May 3, 2026 at 8:15 PM Yicong Huang <[email protected]> wrote: > Hi Texera team, > > > I wanted to provide some updates on the recent CI changes that have been > implemented this weekend. These updates should boost our overall workflow! > > 1. *Per-component label-based CI*: CI can now be initiated based on PR > labels, allowing checks to focus solely on the components affected. For > instance, changes related to the frontend won't trigger backend checks. > Labels are automatically detected from the PR's diff, and committers can > manually add component labels to run extra checks if necessary. > 2. *Auto Queue*: PRs targeting main can be managed by Auto Queue once > approved and with auto-merge enabled. This helps maintain a good merge > flow. Please refrain from manually clicking “Update branch” in most > cases, > as it can cause unnecessary CI reruns and slow down the queue. Use the > *emergency* label to prioritize specific PRs when you need that to be > moved to the front of the queue and get merged soon. > 3. *Coverage report*: CI now uploads coverage reports for your > reference. You should see coverage reports as comments on PRs after CIs > are > done. Use them to gauge the test coverage impact of each PR and assist > in > improving our coverage gradually. You can also find the coverage badge > in > the README.md at the project root. > 4. *Nightly license check*: We conduct a nightly strict license-binary > check on main to monitor dependency and license drift. PR builds operate > in a relaxed mode to avoid blocking merges for minor dependency bumps, > while nightly builds perform a stricter check and log any drift issues. > > With these tools, we were able to merge *110 PRs in the past 48 hours*. > Here are some more details on the CI performance/time: > > Success runs: > *Metric* > *Before* > *After* > *Delta* > *Change* > mean > 16.9m > 11.5m > -5.4m > -32% > p50 > 16.6m > 12.6m > -4.0m > -24% > max > 17.8m > 16.9m > -0.9m > -5% > > Failure runs: > *Metric* > *Before* > *After* > *Delta* > *Change* > mean > 16.5m > 11.1m > -5.4m > -33% > p50 > 17.2m > 12.4m > -4.8m > -28% > max > 17.2m > 14.9m > -2.3m > -13% > > > Thanks to Xinyuan and Jiadong for their great contributions to make this > happen! > > Let us know if you notice any incorrect labels, unexpected CI behavior, or > missing coverage reports. Feel free to create issues for any CI-related > problems. > > > Best regards, > Yicong Huang >
