[MERGE][Bug #50582] Pushing into a local no-trees repo no longer creates a working tree
John Arbash Meinel
john at arbash-meinel.com
Wed Nov 14 14:46:25 GMT 2007
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Daniel Watkins wrote:
> BzrDir.clone seemed to be under the impression that being able to
> create a working tree meant it should go ahead and do so. The attached
> should disabuse it of this notion by checking that the target repo
> wants it to make a working tree before it does so.
>
> I'm not entirely convinced that my test is ideal, but I wasn't sure how
> to improve it so I thought I'd get it looked at in review.
>
>
I think this part:
+ try:
+ repo = self.make_repository('repo', shared=True)
+ except errors.IncompatibleFormat:
+ return
Should be:
+ try:
+ repo = self.make_repository('repo', shared=True)
+ except errors.IncompatibleFormat:
+ raise tests.TestNotApplicable('must support shared repositories')
I would also say that the end of the test should be:
+ # We should have a branch, but no working tree
+ a_dir = tree.bzrdir.clone(self.get_url('repo/a'))
+ a_dir.open_branch()
+ self.assertRaises(errors.NoWorkingTree, a_dir.open_workingtree())
Also, I guess there was a way for result.find_repository() to not find a repo.
(hence your result_repo = None path). It would be good to have a test for that.
Otherwise it looks good.
BB:tweak
John
=:->
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.7 (Darwin)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFHOwpBJdeBCYSNAAMRAkIgAJ9xtQ1x3nyeJeBVnqXvqbDWHj1wTwCgr7lo
vpNjGc4bB7hqlzO4+M+Rofs=
=wyrg
-----END PGP SIGNATURE-----
More information about the bazaar
mailing list