This is in 1.0.9rc2.  I do a recursive add to import a tree into a newly
initialized repo.  CPU time for `darcs add': 16 sec.  I `rm -rf _darcs',
reinitialize, do `cp _darcs/prefs/boring .boring', then `darcs setpref
boringfile .boring', then redo the `darcs add'.  CPU time this time: a whopping
996 sec.!

Any idea why?

On Sun, May 06, 2007 at 08:10:44PM +0000, Scott L. Burson wrote:
> This is in 1.0.9rc2.  I do a recursive add to import a tree into a newly
> initialized repo.  CPU time for `darcs add': 16 sec.  I `rm -rf _darcs',
> reinitialize, do `cp _darcs/prefs/boring .boring', then `darcs setpref
> boringfile .boring', then redo the `darcs add'.  CPU time this time: a whopping
> 996 sec.!
> 
> Any idea why?

No, this surprises me.  Any chance you could write up a script that
reproduces this? Ideally, I'd imagine something that generates the tree to
be added in the script itself.  I imagine just creating empty files will
work just fine to reproduce this.

And thanks for the bug report! I imagine once we can easily reproduce this,
it won't be too hard to track down the problem--which really has to be a
bug (albeit a performance bug), from what you describe.
-- 
David Roundy
http://www.darcs.net

Since no example to reproduce this has appeared in the last six months, I'm
considering it "presumed-dead" by marking it "wont-fix".

Please re-open if the bug wakes from the dead. 

   Mark