Conversation
mhvk
left a comment
There was a problem hiding this comment.
Looks good. Strange that the travis interpretation of .travis.yml changed like this. Might it be worth pinging them?
p.s. How coverage can change with this PR completely eludes me.
| os: | ||
| - linux | ||
|
|
||
| stage: Comprehensive tests |
There was a problem hiding this comment.
Yes, I had noticed this in my own package too. Does mean we should update package-template as well!
pllim
left a comment
There was a problem hiding this comment.
Ugh, I am not a fan of them silently changing behaviors like this. Thanks!
I don't think the coverage failure is relevant. I have learned to ignore it completely.
|
Remote data failures are hopefully transient. Let's merge this and then see. |
|
One can argue how silent it was, for a while there were config warnings (actually they have introduced the linter and its output which wasn't there before), we just didn't bother about them, not to follow their changelog, etc. |
|
Also, as I recall stages were beta, or at least very new when we started to use them, so having some reshuffle and getting rid of the default Anyway, all is good now, we just need to do some plumbing for CI from time to time. |
|
This is a mess for backporting, so I'll directly fix travis for the 4.0.x branch |
In the past week, two jobs has disappeared from our travis matrix, so it's rather timely to do this config cleanup to get rid of all warnings and issues.
This affects the bugfix branch, but quite possibly it will rather need a separate PR (I'll see which way is better during the attempted backport)