Repository referencing in command lines (was Re: 08:16 < abentley> Better phrasing: [...])

[Matthew D. Fuller]

On Fri, Feb 10, 2006 at 10:01:06AM -0500 I heard the voice of
Aaron Bentley, and lo! it spake thus:
> Matthew D. Fuller wrote:
> > 
> > Please, no.  That's high on the list of reasons I skipped SVN
> > altogether.
> 
> Could you expand on this?

I'll take this up a bit more in response to John's message.


> bzr init traditionally creates a branch in the cwd.  You're saying
> you want 'bzr init --repo=/foo/bar' to do nothing the the cwd at
> all?

Well, I followed up to myself on that murkiness.  "init" currently
does two things; it creates an empty branch, and sets the cwd (or
specified location) to be the working dir of the branch.  ISTM that
the process is the same whether you're storing the branch standalone
or in a repo, so it should be the same command.

So "bzr init --repo=/foo/bar" (or whatever other syntax we come up
with to mean the same thing) should just create and store the branch
in the given repository, instead of standalone in the current dir as
it would without the --repo arg.  $CWD would then become a checkout of
that branch in the repo, instead of the working dir built-in to the
standalone branch.  The only other glitch is that, when storing in the
repo, we really need to give the branch a name right off; it can't be
anonymous like a standalone.


This is distinct from my other thoughts in an earlier mail on "init"
being the command to initialize a repository, and suggests that a
"repoinit" or similar other command should be used for that.  I
retract my linking of 'bzr init' to 'cvs init' in that context, and
suggest a sound beating for my failure to think it through.


-- 
Matthew Fuller     (MF4839)   |  fullermd at over-yonder.net
Systems/Network Administrator |  http://www.over-yonder.net/~fullermd/
           On the Internet, nobody can hear you scream.