[darcs-devel] announcing darcs 2.0.0pre1,
the first prerelease for darcs 2
droundy at darcs.net
Thu Dec 13 22:25:08 EST 2007
On Thu, Dec 13, 2007 at 07:12:45PM -0800, David Roundy wrote:
> On Thu, Dec 13, 2007 at 09:50:50AM +0000, Simon Marlow wrote:
> > Simon Marlow wrote:
> > >David Roundy wrote:
> > >>Yikes! That's actually a very surprising bug. I'd be interested in
> > >>hearing if it shows up if you run a darcs2 optimize first? Either way,
> > >>of course, it's a serious bug, but that'd give a hint where the
> > >>trouble is.
> > >
> > >darcs2 check has nearly finished... yup, the repository is consistent.
> > >
> > >Trying optimise now...
> > darcs2 optimize was killed by the OS after 5 minutes for using up all my
> > memory :-(
> Okay, I'm pushing the fix for this now (and it should hit the hashed darcs
> repository in about an hour, and the old-fashioned darcs repository maybe
> an hour after that--tests need running). We were being way to strict in
> our reading of the repository. I needed to redefine the (poorly named)
> Hopefully data type in order to allow us to delay reading of patch files
> longer. Optimize now takes 3-6 seconds (depending on cache state) with
> profiling enabled, versus previously 30 minutes without profiling enabled
> (presumably I've got more RAM than you do).
I take this back. One test fails with the new changes, and I need to
sleep. So it won't be until some time tomorrow that you'll have a darcs 2
that's really worth more efficiency testing. Semantics testing should be
fine (but is more work).
Department of Physics
Oregon State University
More information about the Cvs-ghc