Inconsistent release 1.6.17

[froth]

Scala steward informed me about a new release 1.6.17.
Looking at https://search.maven.org/search?q=mockito-scala it seems that something went wrong in the process, some packages have the new version some the old. Even mixing for different scala versions.

Not a real problem for me, but I thought you might to know (if you did not already noticed it)

[ultrasecreth]

I noticed the release job failed yesterday, but on the bintray stage, didn't know that the maven release was corrupted...
As it was a dependency update I didn't pay that much attention, but what you say actually is a problem, gonna have to summon @mockitoguy here, as I'm not really familiar with the release process

Thanks for reporting it!

[mockitoguy]

Hey guys! This is a little strange.

• "mockito-scala_2.13:1.16.17" has been published to central (proof) but it does not show up in search.maven.org.
• "mockito-scala_2.12:1.16.17" (and 2.11) has not been published to Bintray, and therefore is missing from Central.

@bbonanno, what are the expected covariants of Scala we should be publishing?

[omarkilani]

For whatever reason, the .pom is missing from the published artifact:

https://dl.bintray.com/mockito/maven/org/mockito/mockito-scala_2.13/1.16.17/

Note that the .pom.md5 is there, so that's a little strange.

[mockitoguy]

The problem is that the CI build failed to publish all artifacts (details). My theory is that it's a bug in Gradle Bintray plugin: when one of the bintrayUpload tasks fails, the bintrayPublish task should not run (e.g. this looks like a bug in plugins handling of Gradle's finalizedBy/dependsOn). I haven't though about a remedy for this yet.

[mockitoguy]

@bbonanno, [ci skip] does not work, we need to use [skip ci], see the CI configuration. Same goes for skipping releases, it needs to be [skip release] ("skip" goes first ;-)).

[ultrasecreth]

@mockitoguy yeah, I noticed that after I merged, but in any case, it should have done a full release right? (I didn't care cause a full release out of that PR was not the end of the world).
I also noticed a PR failed the build on just the bintray upload (which is weird, why is a PR uploading anything)

The full release is a cartesian product of all the modules that produce a Jar x the 3 Scala versions we support

[ultrasecreth]

I just merged another PR, no comments, so let's see how the release goes

[froth]

went good :) the new release is consistent

[ultrasecreth]

Nice, let's hope it doesn't happen again!

[ultrasecreth]

Had to release again (another fix was made to the lib) so if that one's also successful I'll close this issue