Skip to content

ci: fix run_n_monitor single execution

Sergi Blanch Torné requested to merge sergi/mesa:ci-run_n_monitor-bugfix_single_job into main

What does this MR do and why?

There is an issue reported #11517 (closed) explaining that calling the tool ci-run_n_monitor with a single job as a target, it finishes the monitoring too early.

The source of the problem is when the job takes more than one pipeline monitoring cycle to change to the running state. When the job is enabled and stays in pending status (or in created) when the loop concludes (to run dependencies), the return of the monitor_pipeline() gets confused and returns the wrong structure.

This was here silently before !29917 (merged) and the introduced set of which are the considered running status helps to solve this problem.

cc: @gfxstrand @eric

Merge request reports