Re: Significant performance waste in git-svn and friends

["Shawn O. Pearce" <spearce@spearce.org>]

David Kastrup <dak@gnu.org> wrote:
> Junio C Hamano <gitster@pobox.com> writes:
> 
> > Mike Hommey <mh@glandium.org> writes:
> >
> >> The same things obviously apply to git-cvsimport and other scripts
> >> calling git-hash-object a lot.
> >
> > I wonder if letting fast-import handle the object creation is an
> > option, though.
> 
> I think it would be saner to give git-hash-object an operation mode
> that makes it usable as a pipe-controlled daemon, so that one needs
> not fork and exec for interning another object.  That way, porcelain
> commands could keep one bidirectional pipe (feed object type and
> source and whether to use -w into git-hash-project, receive object id)
> to git-hash-object around until they finish.

Aside from getting the hashes back from fast-import, that's what
fast-import is for.  I could also make it disable writing.  Hmm.
Junio and I were just talking about making fast-import send the
marks table back out on stdout.  This would make it easier for a
frontend process to stream a whole bunch of objects into the process,
then get back all of their SHA-1s.  Less context switches and more
parallel operation.

-- 
Shawn.