Hi, On Sat, 27 Jun 2026 at 14:43, Santiago Vila <[email protected]> wrote: > > On Sat, Jun 27, 2026 at 01:41:38PM +0300, Otto Kekäläinen wrote: > > > If you do want to use Salsa CI but struggle to get it fully green, you > > can also just disable the failing jobs without turning off the entire > > pipeline [2]. > > That's what I usually do: > > Disable job "foo" until someone figures out how to fix it. > > and I highly recommend it as well. > > In fact, I believe this should be the first recommendation, before > considering to disable the pipeline completely as you suggest in the > subject. So, while I agree that a disabled CI is better than a > persistently failing one, I think a CI which becomes green by dropping > a single job is a lot better than a disabled CI. > > I know there is also another school of thought which proposes to make > the failures non-fatal, but IMO that should only be done if there is > really a fix in the horizon. If there is not, there is little point in > wasting resources.
Thanks, that's exactly how I am thinking. If developers are not quick to fix or disable failing jobs, they will develop "alert fatigue" and stop caring when the CI is red. The longer it remains red, the less likely it is to ever turn green again. I also realize that the people who read debian-devel@ are on average the more diligent ones who take high pride in their Debian work. Those who spend less time on packaging quality or are simply sloppy by nature probably don't read or react to any suggestions on debian-devel@ anyway. What we should have is some kind of feedback loop from Salsa CI to the overall package lifecycle that incentivizes people to avoid permanently failing CI pipelines, but that does not exists currently.
