~bzr-pqm/bzr/bzr.dev : changes

~bzr-pqm/bzr/bzr.dev » Changes from revision 6437.70.17

Rev	Summary	Authors	Date
6437.70.17	Change the attribute lookup, I missed it in the first pass. Change the attribute lookup, I missed it in the first pass.	John Arbash Meinel	12 years ago
6437.70.16	per_tree re-uses the per_workingtree scenarios. per_tree re-uses the per_workingtree scenarios. Which means we have to update the per_tree tests to handle the new model, as well as update the test_selftest code that explicitly tested the permutation code.	John Arbash Meinel	12 years ago
6437.70.15	rename some variables for clarity. rename some variables for clarity.	John Arbash Meinel	12 years ago
6437.70.14	Finish bug #1046697 and run all per_workingtree tests agains... Finish bug #1046697 and run all per_workingtree tests against a checkout. Specifically, a lightweight checkout of a RemoteRepository. We don't try to do all permutations, but we do try to at least do the default WT format against the default repository format. Update NEWS/release-notes to the associated bug.	John Arbash Meinel	12 years ago
6437.70.13	Use development-subtree so the tree reference tests pass. Use development-subtree so the tree reference tests pass. The support is actually determined by the repository. But the bug is that while Tree.supports_tree_reference() returns False, tree.add_reference() passes without complaint.	John Arbash Meinel	12 years ago
6437.70.12	Merge the get_file_text cleanup. Merge the get_file_text cleanup. There were a few tests that fail because the filtered content ends up coming back as the empty string, rather than actually coming back with content. Obviously, we want content :).	John Arbash Meinel	12 years ago
6437.70.11	record_stream wasn't being iterated to completion. record_stream wasn't being iterated to completion. This meant that we didn't always finalize a nested_progress_bar. Which meant that our test suite thought we were at a depth >1. Which meant that ProgressRecordingUIFactory was discarding the update messages. So wrap the progress bar in a try/finally. That way when the generator exits, we get a proper cleanup. Note that record_stream is generally driven by iterating the stream and stepping it, so it isn't too surprising that we weren't stepping that one final time to trigger generator teardown.	John Arbash Meinel	12 years ago
6437.70.10	WorkingTreeFormat6._matchingbzrdir wasn't actually creating ... WorkingTreeFormat6._matchingbzrdir wasn't actually creating a matching bzrdir. The issue is it was just inheriting the value from WTFormat4, so stuff like views don't work. This may not fix everything, but it helps. Also, update one skip-the-test that really meant 'this permutation is not applicable'.	John Arbash Meinel	12 years ago
6437.70.9	branch_builder builds in the branch/repository location, not... branch_builder builds in the branch/repository location, not in the wt location.	John Arbash Meinel	12 years ago
6437.70.8	trivially implement peek_lock_mode. trivially implement peek_lock_mode.	John Arbash Meinel	12 years ago
6437.70.7	one more test that wanted to have a branch reference one more test that wanted to have a branch reference and another test that assumed wt.branch was in the same location as wt.	John Arbash Meinel	12 years ago
6437.70.6	Fix a couple tests that wanted to directly create a wt where... Fix a couple tests that wanted to directly create a wt where the branch was. It wasn't actually needed for the tests, they just wanted a cheap way to not run the test against trees that didn't support checkouts. A couple other tests that assumed the location of the WT was the same location as self.get_transport() or the same location as wt.branch.	John Arbash Meinel	12 years ago
6437.70.5	user_url seems a cleaner way to do the same thing. user_url seems a cleaner way to do the same thing.	John Arbash Meinel	12 years ago
6437.70.4	Now that we have 2 possible locations where things exist, we... Now that we have 2 possible locations where things exist, we must cleanup both. test_kind_change was re-using the 'tree' location for each permutation of the test, and using osutils.rmtree() to cleanup. However that won't clean up the RemoteRepository location at all. So switch to using tree.bzrdir.root_transport.delete_tree(), and also check if the branch location needs to be cleaned up separately.	John Arbash Meinel	12 years ago
6437.70.3	Switch from 'clone' to 'sprout' for the executable test. Switch from 'clone' to 'sprout' for the executable test. 'cmd_branch' seems to use sprout anyway, and 'clone' from a RemoteBranch will create a target branch that doesn't have a WT, because the source doesn't have a WT.	John Arbash Meinel	12 years ago
6437.70.2	Merge the fix for having RemoteBranchFormat.initialize(local... Merge the fix for having RemoteBranchFormat.initialize(local_bzrdir) properly share the repository object that is passed in.	John Arbash Meinel	12 years ago
6437.70.1	Set up the infrastructure to start testing a lightweight che... Set up the infrastructure to start testing a lightweight checkout of a remote repository. Now some tests are failing which are genuine bugs. So start going through the test failures and fix them.	John Arbash Meinel	12 years ago
6437.63.6	(gz) Merge 2.4 into 2.5 (Martin Packman) (gz) Merge 2.4 into 2.5 (Martin Packman)	Patch Queue Manager	12 years ago
6437.63.5	(gz) Ignore bzrlib/locale directory created by build process... (gz) Ignore bzrlib/locale directory created by build process (Ross Lagerwall)	Patch Queue Manager	12 years ago
6437.63.4	(vila) Fixes '-d' handling for 'bzr config' (Wouter van Heys... (vila) Fixes '-d' handling for 'bzr config' (Wouter van Heyst)	Patch Queue Manager	12 years ago

Older »