migrating to git

Sittampalam, Ganesh ganesh.sittampalam at credit-suisse.com
Tue Jan 11 09:02:12 CET 2011

Simon Marlow wrote:

> The darcs team have been making great strides with performance, but
> conflict handling remains a serious problem.  The darcs roadmap
> doesn't show this being fixed in the near future  
>    http://wiki.darcs.net/Roadmap

I've just updated the roadmap for darcs 2.8 (the next major release, due
this Summer) to mention a few items about conflict handling:

 - "Include old text in conflict marking"

If you have a conflict where you start with xy, and in one branch change
that to x and in another change it to xyz, the current marking is pretty
hard to decipher:

v v v v v
^ ^ ^ ^ ^

The new marking will give you

v v v v v
^ ^ ^ ^ ^

which makes things a lot clearer

This is committed already so will definitely be in (in fact I'm now
tempted to include it in the upcoming patch release, 2.5.1)

 - "reduce the number of alternatives presented when there are
complicated conflicts"

Complicated conflicts often result in a mess of conflicting options
being presented to you, where some of the alternatives are a seemingly
randomly chosen subset of the one side of the conflict. I've got some
patches that improve this problem, though they don't eliminate it
completely - this will probably be in 2.8 (so far the patches only work
for v2 format, but it should be possible to make them work for v1 too).

 - "include patch names in conflict markers"

I have done some implementation work towards this and am hoping to have
something complete for 2.8, though there are still some issues to
resolve (might need a compatible patch format change, akin to the
old-fashioned -> hashed conversion process). At the very least the work
I've done so far will mean that it'll be available for conflicts arising
during a rebase.

There are other problems with conflict handling for which fixes are
further off and much more uncertain, which I'm happy to discuss further
if people are interested.



Please access the attached hyperlink for an important electronic communications disclaimer: 

More information about the Glasgow-haskell-users mailing list