[darcs-users] Darcs 2.5 roadmap

Simon Marlow marlowsd at gmail.com
Tue Feb 2 12:04:55 UTC 2010


I've just been testing darcs-beta on my NFS home dir at work, and it 
seems quite usable so far, and I very much like how fast it is to create 
new GHC branches now.  The only thing I noticed taking an inordinately 
long time was converting the GHC testsuite repo from darcs1 to hashed 
format, but that only has to be done once. I'll be switching to hashed 
repos for local work for the time being - this is a nice improvement, 
thanks guys!

On 01/02/2010 20:11, Eric Kow wrote:
> On Mon, Feb 01, 2010 at 19:21:33 +0000, Florent Becker wrote:
>> On a quite not unrelated note, do we have a roadmap for darcs 2.5?
>
> Yes: http://wiki.darcs.net/Roadmap
>
> I've taken the liberty of inserting new information into the map
> pre-empting some of the stuff we decided together in Vienna. IMHO,
> it would be a good idea for us to
>
>   - finish up Luca's caching work (perhaps we'll see him in Zurich?)
>
>   - address the needless writing of inventories problem that affects
>     repos with lots of history - http://bugs.darcs.net/issue1106 ,
>     notably GHC.  This problem makes some operations O(n) time with
>     respect to the number of patches in the repo, when we could
>     "easily" boil that n down to the number of patches since the last
>     tag.

I like the above two.  Also, if I'm allowed to vote for things that 
would be important for us:

1. Better information about conflicts. Currently, you don't even get to 
find out which two patches are in conflict, which seems a bit silly (you 
know which files have conflicts, and you can see one, but not the other 
patch(es)).

Conflict markers are almost useless, as the hunks are in a random order 
and not marked with the patch that they come from.

I'm aware that these may be difficult to fix - I think that was the 
conclusion last time I brought it up - but nevertheless we shouldn't 
forget that this is a serious problem with the darcs UI at the moment, 
and so despite its difficulty should IMHO be given a high priority 
(personally I care a lot more about this than hunk splitting, for example).

2. Rebase.

Cheers,
	Simon


More information about the darcs-users mailing list