On Tue, Jul 03, 2007 at 02:10:05PM +0100, Simon Marlow wrote:
> I built the stable branch today on my Windows machine, and it fails to push 
> to an existing repository, with this error:
> 
> > c:/builds/darcs/darcs push [EMAIL PROTECTED]:/home/darcs/ghc-wind
> ows-builds
> darcs failed:  Not a repository: 
> [EMAIL PROTECTED]:/home/darcs/ghc-windows-builds ((scp) failed to fetch: 
> [EMAIL PROTECTED]:/home/darcs/ghc-windows-builds/_darcs/hashed_inventory)
> 
> Is this the correct behaviour, or should it be backwards compatible?  Do I 
> need to upgrade the remote repository?

No, this is not expected behavior, we should be backwards compatible.  :(
I'm not sure what causes this, but would definitely like to fix it.  It is
clearly a bug in the repository-identification code, but I'm not sure what
the bug is.

Does anyone else have an idea what the trouble here is? Also, does this
disappear if you add a _darcs/format having "darcs-1.0" in it? I would hope
that it would, but we should definitely find and fix the bug anyhow.  It
looks similar to the bug we've seen with the darcs repositories when a 404
message is sent over http, but obviously you aren't getting a 404...

> Nice to see this hashed inventory stuff going in, BTW.  I hope this will 
> make life a lot easier with partial repos.

It should, but there also are some (possibly severe) performance
regressions, which may be very serious, which relate to my goal of having
simple code that is clearly for the first implementation.  I'm not sure
when anyone will get around to tuning the new code, but don't expect to
work on that myself in any serious way until at least after the new
conflicts stuff is complete.  Obviously I didn't quite manage to make the
new code bullet-proof... although as always "it worked for me"!

Hopefully being all married now and having a few weeks to help Jason on
finish off his SoC project and then integrate it into darcs, this will move
along, but things are moving more slowly so far than we'd hoped.

Once we've got debugged and cleaned up the new conflicts code, and possibly
also integrated the GADT type witnesses all the way through the code
(they're only in the core right now, and are #ifdeffed out for compiling
the whole of darcs) in order to make refactors and optimizations safer.
-- 
David Roundy
Department of Physics
Oregon State University
_______________________________________________
darcs-devel mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-devel

Reply via email to