pybaz/baz-import/baz bug with sealed branches

[John Arbash Meinel]

I'm not sure who exactly is to blame. But I found an interesting problem
while working with bzr baz-import.
If you have a branch which has been sealed (with commit --seal), so that
it has a 'version-0' patch, then 'baz ancestry-graph' returns an
incomplete graph.

Specifically, I have a long-lived branch, which I ended up sealing

$ baz ancestry-graph john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--version-0
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--version-0

But if I go to one of the patches, I get a full ancestry
$ baz ancestry-graph
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-10
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-10
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-10
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-9
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-9
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-8
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-8
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-7
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-7
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-6
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-6
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-5
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-5
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-4
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-4
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-3
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-3
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-2
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-2
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-1
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--patch-1
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--base-0
john at johnmeinel.com--2004/mifar-all--CVS-dev--0.7--base-0
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-7
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-7
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-6
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-6
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-5
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-5
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-4
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-4
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-3
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-3
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-2
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-2
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-1
john at johnmeinel.com--2004/mifar-all--CVS--0.6--patch-1
john at johnmeinel.com--2004/mifar-all--CVS--0.6--base-0
john at johnmeinel.com--2004/mifar-all--CVS--0.6--base-0

(there is actually something like 200 patches in this branch).
The problem is that pybaz.Revision.iter_ancestors uses graph_ancestry to
figure out what the ancestors are.

It looks like this might just be a baz 1.4.2 bug. As 1.5 seems to do the
right thing. Any chance that we will ever see a baz-1.5 release? (So
that I can just use apt-get rather than having to build from source?)

John
=:->
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 256 bytes
Desc: OpenPGP digital signature
Url : https://lists.ubuntu.com/archives/bazaar/attachments/20060112/57cd3535/attachment.pgp