> Please tell us how to reproduce the problem.
As I said I'm not really sure how to reproduce but the problem seems to be in
the argument order, have a look e.g. at
http://mail.openjdk.java.net/pipermail/build-dev/2008-January/000699.html.
> hgmerge.py is obsoleted by hg 1.0, which has all its functionality built-in.
I didn't know that, thanks.
I have a suspicion that there can be done incorrect merges if one uses kdiff3
for them. Some time ago it happened to me that I had some conflicts and
kdiff3 appeared on my screen but it had different "layout" than the last time
(at least I had to click "Merge" action myself, previously I hadn't, don't
ask what had changed, I have no idea [maybe hg 0.9.5 -> 1.0?]). So I just
normally merged all the conflicts and it was OK. But... the files on the disk
were different than the ones I saw in kdiff3 as a merge result. So I created
some sandbox repos and played a bit with them and yes, the problem was there.
I searched mercurial wiki and found this page:
http://www.selenic.com/mercurial/wiki/index.cgi/MergeProgram
and saw some notes about kdiff3 - so now I use hgmerge.py (hg repo
http://hg.borho.org/hgmerge/) and it works. I noticed that I'm not the only
person having these troubles (some time ago I helped my colleague with the
same issue).
*** IMPORTANT NOTE
The hg repo http://hg.borho.org/hgmerge/ doesn't contain hgmnerge.py file
anymore so the wiki page
http://www.selenic.com/mercurial/wiki/index.cgi/MergeProgram is not actual.
Thanks.