[darcs-users] darcs hang: infinite recursion?

[Felix Breuer]

On Mon, 13 Jun 2005 08:30:30 -0400
David Roundy <droundy at darcs.net> wrote:

> Alas, I *can* help explain the scenario.  What you've run into is the
> infamous O(2^N) behavior of darcs when it encounters certain sorts of
> conflicts. The code should eventually complete, but it's possible that our
> sun will become a red giant before that happens (which would most likely
> cause darcs to fail).  :(

Will that problem (hopefully) solve itself when you have developed this
new formalism?

Any way we can help solving this issue? Or can it only be explained to
someone who is up to his neck in patch theory? (I only read the chapter
in the manual :)


> The only solution is to lose some of the history, and rerecord some of the
> patches.  This can be a royal pain if there are many interleaved and
> interdependent patches involved.  If possible, unrecording the changes in b
> that add the new files to www/ and then move them to doc/ would be the
> simplest workaround.  If you have trouble working around this, speak up and
> I (or even better, someone else--I'm a bit busy) will be happy help you
> out.

Thanks for the offer, but I managed. Albeit losing some of the history.


> I am in the process of working out a new formalism for handling conflicts,
> but it'll be a while before it's finished.  On the plus side, I *am* making
> progress, and have a nice framework for generating test scenarios upon
> which to test my merge and commute code.

When you have worked it out and if you have time to spare, please write
a bit on this formalism in the "Theory of Patches" chapter. I am
curious!


Regards,
Felix