[darcs-users] Re: identical change conflict?

Deliverable Mail deliverable at gmail.com
Tue Mar 22 17:27:04 UTC 2005


On Tue, 22 Mar 2005 16:23:57 +0000 (UTC), Aaron Denney <wnoise at ofb.net> wrote:
> > If darcs were to check that the result of aplying the patch to the last
> > commit would indeed yield the "modified" file verbatim, perhaps it might
> see
> > there's no conflict and not bother with it?  At least I'd love to see
> > an option to tell it that's the case, --auto-resolve-identical or
> > something similar.
> 
> This has been brought up before.  It is in fact necessary that they
> conflict, given how darcs works.  See
> http://www.darcs.net/pipermail/darcs-users/2004-December/004649.html
> for example, and a few other places, where people complain about the
> horrible performance problems of merging.
> 
> This should probably be added to the manual, though the same issue
> cropping up in a different guise is discussed in the "unrecord" section.
> 
> --auto-resolve-identical is an interesting idea though, to automatically
> merge "identical conflicts".

Aaron -- thanks for the pointer!  It's interesting to see what each
SCM does in this situation.  By default, tla and bk also detect the
conflict.  In tla, it can be resolved via tla star-merge --three-way. 
In bk, I found that if I keep the repo in checkout:edit mode, and do
bk unedit prior to bk pull, when identical changes are already
injected by other SCMs, it does the trick.

So, only darcs at this moment has no easy way to avoid identical
changes conflict and has to go first in a pull sequence!  (It used to
be bk.:)

Cheers,
Alexy




More information about the darcs-users mailing list