[MERGE] Fix RemoteBranch to be used correctly in tests using bzr+ssh, to fire off Branch hooks correctly, and improve the branch_implementations tests to check that making a branch gets the right format under test.
Vincent Ladeuil
v.ladeuil+lp at free.fr
Wed Feb 18 08:01:41 GMT 2009
>>>>> "robert" == Robert Collins <robertc at robertcollins.net> writes:
robert> We have a per_repository test under discussion. You
robert> say some other tests need a comment copied to it, but
robert> you're not saying which ones :(.
Test *names* :-(
Let's use a more precise media then (I should have think about
that from the start, damn it, we *have* a tool for that :-( ):
test_clone_on_transport_preserves_repo_format
test_clone_unstackable_branch_preserves_stackable_repo_format
=== modified file 'bzrlib/tests/bzrdir_implementations/test_bzrdir.py'
--- bzrlib/tests/bzrdir_implementations/test_bzrdir.py 2009-02-13 00:52:18 +0000
+++ bzrlib/tests/bzrdir_implementations/test_bzrdir.py 2009-02-18 07:06:05 +0000
@@ -559,6 +559,8 @@
source_branch = bzrlib.branch.Branch.open(
self.get_vfs_only_url('source'))
if isinstance(target_repo, RemoteRepository):
+ # we have to look at the real repo to see whether RemoteRepository
+ # did the right thing.
target_repo._ensure_real()
target_repo = target_repo._real_repository
self.assertEqual(target_repo._format, source_branch.repository._format)
=== modified file 'bzrlib/tests/per_repository/test_repository.py'
--- bzrlib/tests/per_repository/test_repository.py 2009-02-13 00:52:18 +0000
+++ bzrlib/tests/per_repository/test_repository.py 2009-02-18 07:03:01 +0000
@@ -815,6 +815,8 @@
repo = repo._real_repository
target_repo = target.open_repository()
if isinstance(target_repo, remote.RemoteRepository):
+ # we have to look at the real repo to see whether RemoteRepository
+ # did the right thing.
target_repo._ensure_real()
target_repo = target_repo._real_repository
# The repository format is preserved.
And, avoiding bundle for BB sanity, available at
lp:~vila/bzr/format-on-push-with-comments.
>> Err, I'm talking about the 'call is None' part. It sounds a bit
>> surprising to record calls and their parameters and sometimes
>> just don't care about the call itself but still recording
>> it. Hence the proposal to update the comment.
robert> Well I don't care about the call at all
I realized that, but I found it suprising without reading the
rest of the patch, which is clear in the patch context. The
suggested clarification was for future readers.
robert> I'll add an explanation.
I can't find it in bzr.dev, oh well... never mind :-/
Vincent
More information about the bazaar
mailing list