Hi Zooko,

Zooko O'Whielacronx wrote:
> Folks:
> 
> The great thing about darcs is that cherry-picking is the native mode
> of operation. There seems to be one place where it doesn't do that
> though -- when you do a "darcs send" the patch that gets produced
> appears to depend on all patches that were in the repo and not in the
> target repo, instead of depending only on the subset that the patch
> really depends on.

The output of darcs send should always work fine with the target repo
(or any repo that's a superset of that).

I guess the problem is that people are trying to use it with other
repositories that are missing patches that are in the target repo.

You're right that in theory darcs could produce a patch with minimal
dependencies - but that would involve commuting it all the way back in
the history of the repo, which is computationally expensive (the
so-called "minimal contexts" problem).

The difference with cherry-picking during a pull or push is that in
those cases darcs knows what's in the source and target repos, so it
knows how far to commute.

A relatively simple workaround would be to maintain a repo with a
baseline that everyone is expected to have, and to always do 'send'
against that. In extremis that baseline could be an empty repo and
you'll get patches with minimal dependencies, but then you'll pay the
computational cost of that.

For a nicer fix, we could perhaps add a flag to darcs to send against a
tag, or something like that. Perhaps if you explain the workflow people
are trying to use here we could figure out what makes sense.

Cheers,

Ganesh

=============================================================================== 
Please access the attached hyperlink for an important electronic communications 
disclaimer: 
http://www.credit-suisse.com/legal/en/disclaimer_email_ib.html 
=============================================================================== 

_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users

Reply via email to