On Mon, Aug 27, 2007 at 12:37:02AM +0200, Tommy Pettersson wrote: > On Sun, Aug 26, 2007 at 06:56:02PM +0200, Eric Y. Kow wrote: > > Otherwise, one option might be to introduce some temporary version of > > the code, say, prefixed with the name issue494 and appended to the end > > of the relevant files where they are not likely to cause conflicts. I > > can then apply these patches, resolve the less hairy conflicts and then > > delete the functions. Do you think that might work, or would it just > > make things really bad for stable? > > This is a good idea. I will try it out when I get the time. I > think it might be possible to just rename the original function > to hide it, and then insert a changed copy of it somewhere where > it won't conflict.
When I took a closer look at this I found that the bad conflicts are all about the import lines at the top of the files, so it can't be worked around since it's only one line in each file that stirs up the conflict. This also makes it alot more likely we'll get a new conflict. :-( I constructed and pulled a deliberately conflicting test patch, which took twenty minutes of 100 % CPU on a modern desktop. I also made a little script to calculate the counts and nesting levels of previous conflicts in the darcs repo, and we've had worse conflicts than this one before. So without better solutions in sight (so far) I think the best option is to wait and pull the patch and its conflict resolution in as late as possible. This way I can amend-record it if something should conflict with it. -- Tommy Pettersson <[EMAIL PROTECTED]> _______________________________________________ darcs-devel mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-devel
