On Mon, Jul 02, 2012 at 05:35:30AM -0000, Neil Martinsen-Burrell wrote:
> > It might be worthwile to use self.revision_tree() rather than
> > self.branch.repository.revision_tree().
>
> self.revision_tree() doesn't seem to work the same. That is, it breaks lots of other test cases with e.g.
>
> ...
> NoSuchRevisionInTree: The revision id {added} is not present in the tree <WorkingTree6 of /private/var/folders/d8/wjzb13qx7ns9yjh46bq9ljs00000gn/T/testbzr-jPjTz6.tmp/bzrlib.tests.blackbox.test_push.TestPushStrictWithChanges.test_push_with_revision(out-of-sync-trees)/work/local>
Perhaps trying WorkingTree.revision_tree() first and if that raises
NoSuchRevision, falling back to Repository.revision_tree() ?
Without using WorkingTree.revision_tree() we're going to be fetching
the entire revision tree from the server in some cases, which is a
major performance regression.
On Mon, Jul 02, 2012 at 05:35:30AM -0000, Neil Martinsen-Burrell wrote: tree() rather than repository. revision_ tree(). tree() doesn't seem to work the same. That is, it breaks lots of other test cases with e.g. nTree: The revision id {added} is not present in the tree <WorkingTree6 of /private/ var/folders/ d8/wjzb13qx7ns9 yjh46bq9ljs0000 0gn/T/testbzr- jPjTz6. tmp/bzrlib. tests.blackbox. test_push. TestPushStrictW ithChanges. test_push_ with_revision( out-of- sync-trees) /work/local> revision_ tree() first and if that raises revision_ tree() ?
> > It might be worthwile to use self.revision_
> > self.branch.
>
> self.revision_
>
> ...
> NoSuchRevisionI
Perhaps trying WorkingTree.
NoSuchRevision, falling back to Repository.
Without using WorkingTree. revision_ tree() we're going to be fetching
the entire revision tree from the server in some cases, which is a
major performance regression.
Cheers,
Jelmer