> -----Original Message-----
> From: Daniel Shahaf [mailto:d.s_at_daniel.shahaf.name]
> Sent: zondag 28 augustus 2016 20:23
> To: Stefan <luke1410_at_posteo.de>
> Cc: dev_at_subversion.apache.org
> Subject: Re: [PATCH] Fix a conflict resolution issue related to binary files (patch
> Stefan wrote on Sun, Aug 28, 2016 at 13:31:39 +0200:
> > The regression test was tested against 1.9.4, 1.9.x and trunk r1743999.
> > I also tried to run the test against 1.8.16 but there it fails (didn't
> > investigate in detail).
> > Trunk r1758069 caused some build issues on my machine. Therefore I
> > couldn't validate/check the patch against the latest trunk (maybe it's
> > just some local issue with my build machine rather than some actual
> > problem on trunk - didn't look into that yet).
> For future reference, you might have tried building trunk_at_HEAD after
> locally reverting r1758069; i.e.:
> svn up
> svn merge -c -r1758069
> <apply patch>
> make check
> Stefan wrote on Sun, Aug 28, 2016 at 18:33:55 +0200:
> > Got approved by Bert.
> (Thanks for stating so on the thread.)
> > Separated the repro test from the actual fix in order to have the
> > possibility to selectively only backport the regression test to the 1.8
> > branch.
> Good call, but the fix and the "remove XFail markers" (r1758129 and
> r1758130) should have been done in a single revision: they _are_
> a single logical change. That would also avoid breaking 'make check'
> (at r1758129 'make check' exits non-zero because of the XPASS).
I do this the same way sometimes, when I want to use the separate revision for backporting... But usually I commit things close enough that nobody notices the bot results ;-)
(While the initial XFail addition is still running, you can commit the two follow ups, and the buildbots collapses all the changes to a single build)
I just committed the followup patch posted in another thread to unbreak the bots for the night...
> Thanks for seeing this through!
Received on 2016-08-28 23:32:38 CEST