remove LocalAbspathServer, LocalRelpathServer

[Martin Pool]

On 24 Aug 2006, John Arbash Meinel <john at arbash-meinel.com> wrote:
> Martin Pool wrote:
> > Getting that wrong makes me think that the current interpretation of -r
> > for bundle could be better.  Perhaps 'bzr bundle -r -2' should give the
> > bundle from -2, similar to bzr diff?
> 
> Possibly. But I think it is better to use the:
> bzr bundle ../bzr.dev
> 
> Form, because that one will give you a bundle that can be definitely
> applied.
> 
> Also, if you did 'bzr branch bzr.dev my-code'
> The parent is set correctly, so that a plain 'bzr bundle' gives the
> right answer.
> 
> I use that form them most (though I use bzr cbranch)

I see what you mean.  I think it's still a little unobvious but not an
outright bug.

> >  from bzrlib.transport.memory import MemoryTransport
> >  from bzrlib.transport.local import LocalTransport
> > +from bzrlib import urlutils
> > +
> > +
> > +# TODO: Should possibly split transport-specific tests into their own files.
> 
> Shouldn't this import be up higher? Possibly grouped with other 'from
> bzrlib import foo' imports?

There were no others, but probably better to put them at the top.  Done.

> ^- we already have a 'self.assertStartsWith()' which is better because
> it gives more understandable error messages.

thanks.

> ^- 'urlutils.local_path_to_url()' can also take a relative path, if it
> makes it more symmetric.

done.

OK, i'll let it wait for 0.11 to reopen on monday.

-- 
Martin