Pali Rohár ha scritto: > On Monday 07 August 2017 10:43:17 Pali Rohár wrote: >> On Monday 07 August 2017 10:26:46 Luigi Toscano wrote: >>> Il 07 agosto 2017 09:34:37 CEST, "Pali Rohár" <pali.ro...@gmail.com> ha >>> scritto: >>> >>>> >>>> Anyway, I rebased kf5 branch on top of master with (semi-manual) >>>> conflicts resolution and found out that changes between kf5 branch and >>>> "kf5 rebased on top of master" differs. And I suspect that some merge >>>> commits which was done from master to kf5 were incorrectly. >>> >>> Do you mean merged, not rebased, right? >> >> I mean real git rebase but result in new temporary branch. > > Hi! Now I rebased current kf5 branch on top of current master and pushed > it into my cloned git repository. Available at: > > https://cgit.kde.org/clones/kopete/pali/kopete.git/?h=master-kf5 > > Reason for it: in kf5 branch are some older version of patches (with > errors) which were properly fixed in version of patches, which then were > integrated into master branch. Having periodical merge from master into > kf5 made 1) git history hard to interpret and 2) inability to use either > git bisect or git blame. It is really a bad idea to have two different > versions of one patch in two different branches in git which are then > merged... > > Now diff between above my branch (master-kf5) and kf5 seems to be clean > and I would propose to use this branch which would be merged into master > (instead of original kf5).
Not really a fan of that, as it's really faking the history. Yes, wrong history on that way, but it's what it is. On the other side, there is no more time to discuss, as the dependency freeze (minus exceptions) for Applications 17.12 is today. So if you think that this Frameworks-based branch is working at least not worse than the kdelibs 4.x master, please go with the merge, or ask for an exception on release-t...@kde.org -- Luigi
