Below is the transcript of a bug that comes from pulling the resolution
of a three-way conflict onto a four-way conflict (the three-way conflict
plus another).  Specifically, there are four conflicting patches a, b,
c, and d.  The first three are combined in repo c, and then resolved by
patch rc.  All four are combined in repo d, then we try to pull rc from
repo c.

Andrew

~% mkdir 0
~% cd 0
~/0% darcs init
~/0% echo 0 > f           
~/0% darcs add f
~/0% darcs record -a -m 0 
Finished recording patch '0'
~/0% cd ..
~% for r in a b c d; do darcs get 0 $r; (cd $r; echo $r > f; darcs record --ignore-times -a -m $r); done
Copying patch 1 of 1... done!
Finished getting.
Finished recording patch 'a'
Copying patch 1 of 1... done!
Finished getting.
Finished recording patch 'b'
Copying patch 1 of 1... done!
Finished getting.
Finished recording patch 'c'
Copying patch 1 of 1... done!
Finished getting.
Finished recording patch 'd'
~% cd d
~/d% darcs pull -a ../a
We have conflicts in the following files:
./f
Finished pulling and applying.
~/d% darcs pull -a ../b
We have conflicts in the following files:
./f
Finished pulling and applying.
~/d% darcs pull -a ../c
We have conflicts in the following files:
./f
Finished pulling and applying.
~/d% cd ../c
~/c% darcs pull -a ../a
We have conflicts in the following files:
./f
Finished pulling and applying.
~/c% darcs pull -a ../b
We have conflicts in the following files:
./f
Finished pulling and applying.
~/c% echo rc > f
~/c% darcs record -a -m rc
Finished recording patch 'rc'
~/c% cd ../d
~/d% darcs pull -a ../c
Pulling from "../c"...
darcs: bug in get_extra commuting patch:
Fri Sep 29 15:37:15 PDT 2006  andrew.pimlott.ctr@metnet.navy.mil
  * d
zsh: exit 1     /home/andrew/tmp/darcs-unstable/darcs pull -a ../c

Hi Andrew!

On Fri, Sep 29, 2006 at 23:14:31 +0000, Andrew Pimlott wrote:
> Below is the transcript of a bug that comes from pulling the resolution
> of a three-way conflict onto a four-way conflict (the three-way conflict
> plus another).

Thanks for the bug report and especially the transcript.

I'm going to merge and defer this one until we solve issue175, which has
the same error message.  Will definitely note the transcript.

Note Tommy's message to darcs-devel on 2007-08-09

  http://article.gmane.org/gmane.comp.version-control.darcs.devel/6300

It has sort of a workaround (while waiting for the new patch theory stuff to
come out)

I just created an automated test for this and submitted it to the bugs/ directory. 

With the old fashioned repo format, it still fails as described above. I suspect
it would pass with the --darcs-2 format, but instead fails with a "Malformed
patch bundle" message, which I believe is separate bug that we already have a
test for. I'm marking the "Malformed patch" bug as a superseder for now.

On second thought, I'm going to go ahead and consider this
"resolved-in-unstable", since it appears to be a dupe of other bugs that are. 

If bugs/issue279.pl continues to fail once the Malformed Patch bug is fixed and
the --darcs-2 format is used, then we'll know we need to re-open this.

Marking as resolved. issue279.pl has been moved to the passing tests/ directory,
and darcs2 has long been released.