RFC: iter_changes with specific_files, status and commit

[John Szakmeister]

On Fri, Jun 26, 2009 at 4:38 AM, Martin Pool <mbp at sourcefrog.net> wrote:
>
> 2009/6/26 John Szakmeister <john at szakmeister.net>:
> > On Fri, Jun 26, 2009 at 4:13 AM, Martin Pool <mbp at sourcefrog.net> wrote:
> > [snip]
> >> I can't think of any problems at the moment.
> >
> > It's a bit weird for the user to provide a path asking about the
> > status of X, and then having the command also tell you the status of
> > Y.  That seems like a behavior we're going to have to explain to them.
> >  Is it a big deal?  Perhaps not.  Is it unintuitive?  Definitely.
>
> You asked for the status of b, and you're told some files that were
> moved into b (or out of b.)  It might not be exactly what you expected
> but it's not very surprising to me.

Yeah, I guess that's a better way of putting it.  I wouldn't expect
that output, but I can certainly understand why it's telling me about
it.

> And of course it would be nice to have this come about by reducing
> code divergence between commit and status, as is already happening
> with iter_changes focussed patches.

I can certainly appreciate the value of reducing code divergence. :-)

-John