[PATCH] baz2bzr: direct_merges performance bugfix

[Robert Collins]

On Sun, 2005-10-16 at 22:27 +0200, David Allouche wrote:
> On Sun, 2005-10-16 at 20:00 +0200, David Allouche wrote:
> > Attached is patch that rewrites the direct_merges function used by
> > baz2bzr to
> >       * avoid unneeded network round-trips (by a avoiding using "baz
> >         ancestry-graph --immediate" unless absolutely needed)
> >       * not rely on the New-Patches header of continuation revisions,
> >         which have been historically unreliable
> >       * incidentally fix the explosive memory usage problem
> >       * use set operations instead of list operations
> >       * use pybaz Patchlog.new_merges feature instead of reimplementing
> >         it
> > 
> > I _think_ it is as correct as possible, but I might have missed some
> > obscure aspect of the problem.
> 
> Indeed, I missed the fact that imports are not continuations and have no
> ancestor. Fixed in the updated patch in attachment. 
> 
> Also, it occurred to me that the changes in direct_merge logic may cause
> imports to be different, so maybe the namespace should be bumped as
> well.

What changes are you expecting in the output?

Rob

-- 
GPG key available at: <http://www.robertcollins.net/keys.txt>.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20051017/8880aa53/attachment.pgp