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

~bzr-pqm/bzr/bzr.dev » Changes from revision 5847.4.35

Rev	Summary	Authors	Date
5847.4.35	Rename 'expect_fail' to 'invalid_delta' as suggested by Jelm... Rename 'expect_fail' to 'invalid_delta' as suggested by Jelmer Makes the actual value of the parameter clearer to understand.	John Arbash Meinel	13 years ago
5847.4.34	Get rid of rename_targets. Get rid of rename_targets. We weren't checking it yet anyway. Instead we use _id_index. We need it anyway to check for file_id uniqueness. And we need it for the pure-add case, vs a rename, and renames are rare, so it would only rarely help. Since we now have direct modifies going through changes, I don't think we need to worry about the performance.	John Arbash Meinel	13 years ago
5847.4.33	Clean up the _update_basis_apply_changes code Clean up the _update_basis_apply_changes code ensure that the code path is getting hit.	John Arbash Meinel	13 years ago
5847.4.32	Update the test suite to handle root changes. Update the test suite to handle root changes.	John Arbash Meinel	13 years ago
5847.4.31	Clean up test_inv. Clean up test_inv. The tests were failing for WT2/3 anyway. But instead of shotgunning slow tests in test_inv, do fast directed tests in test_dirstate. We might actually be in the clear, though I'd like to improve performance by using update_changed more often, and we may want a direct test of root-id changing?	John Arbash Meinel	13 years ago
5847.4.30	simplify the delete case. simplify the delete case. if real_delete was confusing the issue, we do the same thing either way. The only thing I can think of is if we get a directory entry, which shows up as a remove+add at the same location. But there shouldn't be a way to actually get that.	John Arbash Meinel	13 years ago
5847.4.29	Simplify a bit. Handle that when a directory is added, we ne... Simplify a bit. Handle that when a directory is added, we need to add its block. Handle that when a directory is removed, if it is the last reference to that path, the block should also be removed. I'm going to try simplying the remove code now.	John Arbash Meinel	13 years ago
5847.4.28	Some more direct tests. Fix some bugs in the create_entry co... Some more direct tests. Fix some bugs in the create_entry code.	John Arbash Meinel	13 years ago
5847.4.27	Handle missing parent, though bugs may still remain Handle missing parent, though bugs may still remain	John Arbash Meinel	13 years ago
5847.4.26	_check_delta_ids_absent thought it didn't have to set _chang... _check_delta_ids_absent thought it didn't have to set _changes_aborted=True but it really does. Because we may have called _*apply_deletes first, as part of splitting up renames into delete+add pairs.	John Arbash Meinel	13 years ago
5847.4.25	Trap an InconsistentDelta case. add can try to add another Trap an InconsistentDelta case. add can try to add another entry to the same location. Not too bad to detect, since we just need to look at nearby locations.	John Arbash Meinel	13 years ago
5847.4.24	Start handling an InconsistentDelta case exposed by test_inv... Start handling an InconsistentDelta case exposed by test_inv cases.	John Arbash Meinel	13 years ago
5847.4.23	Add the first invalid delta test. Add the first invalid delta test.	John Arbash Meinel	13 years ago
5847.4.22	Cleanup dirstate a bit. Cleanup dirstate a bit. Lots of places were fumbling around trying to raise InconsistentDelta, not all of them were properly setting _changes_aborted=True. So force them all through a helper. Also gives a little bit more room per line, though from what I can see, the reasons are always long enough to wrap to the next line anyway.	John Arbash Meinel	13 years ago
5847.4.21	We were raising exceptions improperly. A couple more tests c... We were raising exceptions improperly. A couple more tests cases, so far so good.	John Arbash Meinel	13 years ago
5847.4.20	Handle a bunch of rename cases. Handle a bunch of rename cases.	John Arbash Meinel	13 years ago
5847.4.19	Get the ugly 'exists, but at a different location' working. Get the ugly 'exists, but at a different location' working.	John Arbash Meinel	13 years ago
5847.4.18	And now we add a bunch more tests, and finally have a failin... And now we add a bunch more tests, and finally have a failing case.	John Arbash Meinel	13 years ago
5847.4.17	The 25ms was the time spent checking that we didn't alter th... The 25ms was the time spent checking that we didn't alter the containing WT. Switch to using temp files everywhere, since we don't actually need to write anything to disk. (Except DirState requires a disk file.) Now down to ~2ms per test, which is a great feeling.	John Arbash Meinel	13 years ago
5847.4.16	Switching from BranchBuilder to Inventory directly dropped u... Switching from BranchBuilder to Inventory directly dropped us from 73ms to 26ms. Probably as good as we'll get.	John Arbash Meinel	13 years ago

Older »