'bzr blame' fails with ghost revisions and 'bzr check' output...

[John Szakmeister]

On Tue, Dec 29, 2009 at 3:39 PM, John Arbash Meinel
<john at arbash-meinel.com> wrote:
[snip]
> If you just want to do it from the test suite, you can do:
>
> tree.set_parent_ids([tree.last_revision(), 'ghost-revision-id'])
> tree.commit("commit with ghost")
>
> Note, however, that bzr-svn is currently the only way to introduce a
> ghost revision *and* to mark a file as last-modified in that revision. I
> think there would be a way to inject it if you manually set the values
> in an inventory and call Repo.add_inventory().

Ah.  That's probably more along the lines of what I want to do then.
I'd like to get some tests that replicate what you'd see in a real
situation with bzr-svn so we can get these issues resolved and prevent
regressing.

> The issue is that bzr-svn is preserving the last-modified information
> from the ghost that it isn't including in the ancestry. bzr itself, and
> other converters don't ever do that.

I'm not sure how to interpret that: is bzr-svn doing something wrong
then?  Or are you saying it's just different?

-John