Closed b-sharpe closed 6 years ago
+1 for this idea. It opens up another vector for us to discover patch failures, which is cool for us from a maintenance standpoint and allows us to get ahead of rerolls.
The con here is that people may send us support requests when they, not us, are including invalid patches. But in those cases, I guess it should be quick to diagnose and close those requests.
So, I'm on board.
@balsama, what do you think?
I'm fine with this. There are situations where the existing behavior could help someone out. E.g. a new release of core comes out with one of our patches committed. In that case, the user would would get the same functionality even though the patch failed to apply.
In that case, they really should update Lightning too though.
Tests are failing though. I just restarted - if they pass, we should merge.
Currently patches just get skipped if they fail to apply. Since Lightning doesn't pin core, this is potentially hazardous to a core update in which a patch provided by lightning may no longer apply.
PR enables exiting composer when a patch fails to apply.