GitHub seems to have issue with workflow state caching that causes the
DNM step to not work properly in few cases and not detecting changes in
the DNM tag, forcing people to mess with tags or close/opening PRs,
which in turns restarts all workflows.
Convert the script to Python so that the tag data is guaranteed to be
fresh.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
This reverts commit 370e0882cb, the
condition is evaluated at workout creation time so this does not work at
all, plus some conditions don't have a label removal logic so pairing
this with the manifest run does not work anyway.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
The current CI setup sets the manifest DNM label in a workflow and
checks it in a different workflow. The one performing the check is
configured to rerun on label changes but it's been reported few times
that the rerun logic does not seem to always run effectively and there's
been cases where the manifest has been fixed but the label has not been
removed by the automation, resulting in a stale PR.
Since the manifest has its own dedicated label, fix this race condition
by checking for that label specifically in the manifest workflow rather
than in a separate one, this means that the check is always performed
after the script that sets the label.
It also means that on manifest changes the manifest step itself will
fail rather than the PR metadata check one.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
