Hi,
in my team, there is a discussion to parallel our Tycho build using the
Maven -T option. With Tycho 1.6, we get too often
ConcurrentModificationExceptions. One of my colleague saw that this
feature is planned for 1.7 release
he propose to fork current Tycho master branch, doing some fixes
That's a bad idea based on a lack of understanding of how open-source works best.
In the end, your forked branch and Tycho will diverge and you'll have to spend a lot of time re-synchronizing them whenever you want to use a new Tycho release. You'll quickly waste more energy in it than it would take to directly contribute the fixes to Tycho. The ROI of forking as OSS project vs contributing is always bad in the end.
This is even more true for Tycho which publish good snapshot builds whenever there is a change, so the "time to market" between a merged patch and you can try it for your build is less than 24 hours.
what are the remaining tasks for this feature and do
exist a possible release date for 1.7?
I don't know. Some discussion on the ticket may answer that. But I think at the moment, best would be that you test it, and report specific issues for each problem you face. As bug 380169 is already full of info, you can probably just open new tickets and mark them as "Blocks: 380169" in the dedicated Bugzilla field
I have a look at the related ticket [1]. I understand that still three
plugins are not thread-safe. Would be it enough to fix them with the
solution which the xtext-maven-plugin chose? [2] If applicable, I could
prepare a patch for them.
I don't know whether it would be enough, but in any case, any good contribution is welcome as long as it improves stuff (even if it doesn't fully fix things). So yes, please prepare a patch ;)
