Repository referencing in command lines

[Erik Bågfors]

2006/2/11, Matthew D. Fuller <fullermd at over-yonder.net>:
> On Sat, Feb 11, 2006 at 08:15:45AM -0600 I heard the voice of
> John A Meinel, and lo! it spake thus:
> >
> > Except how do you mix working with a repository, with working without a
> > repository.
> >
> > For example:
> >
> > bzr get foo.bar
> >
> > Are you referencing a local directory (which may not exist), or are you
> > referencing the repository.
>
> Which is why I suggest the "repo:<repo_url>:<branch_name>" syntax in
> the first place    8-}
>
> # Standalone branch at path "./foo.bar/"
> % bzr get foo.bar
>
> # Same as above; 'standalone:' marker exists for disambiguation if
> # necessary
> % bzr get standalone:foo.bar
>
> # branch called 'foo.bar' in whatever repo is currently 'default',
> # based on current dir or env variable or whatever
> % bzr get repo::foo.bar
>
>
> Now, consider these:
>
> # branch called 'foo.bar' in repo at given URL
> % bzr get repo:"sftp://me@server/some/where/my/repos":foo.bar
>
> # Standalone branch at URL
> % bzr get sftp://me@server/some/where/my/repos/foo.bar
>
> I'm happy to have the latter syntax work just as well if that's a
> branch within a repo, as if it's a standalone (that is, the latter
> syntax would work in the former case, too).  In fact, I'd go along
> with you in saying that it almost _MUST_, for consistency and sanity.

I think we should start with supporting only the last syntax for both
standalone and none standalone and then see if that works well.  I
think it does and that we don't have to do anything else.  If not, we
implement the first thing as well..

/Erik