20 minutes ago, Moritz Bunkus wrote:
> Hey,
>
> On Fri, Feb 3, 2012 at 16:13, Eli Barzilay <[email protected]> wrote:
>
> > I think that magit should leave the buffers as is
>
> I disagree. I like the current behavior and rely on the buffers
> being up to date with the working tree after having checkout out a
> revision. So if this is changed then it should be configurable
> (e.g. "revert without asking", "revert with asking", "don't revert
> at all").
Obviously, the current behavior should be available as an option. I
think that it's also obvious that there *should* be an option.
The subjective point is what should the default be. IMO, having the
revert-without-asking thing is dangerous enough that it shouldn't be
the default (but for myself I'd be happy regardless to have an
option).
Just now, PJ Weisberg wrote:
>
> This I can agree with. Make "asking" the default, and have a "never
> ask again" option, to keep all the convenience without the potential
> for surprising data loss.
That's a good way to have neither as a default, but there are two
problems with it:
1. How do you implement the "never ask again"? Editing .emacs files
to add options seems to be very out of fashion. (Maybe just invoke
the custom functionality?)
2. Also, this means that the "asking" that you're talking about is
going to be magit code -- not the asking that gets done by Emacs
when a buffer's file is modified. This is probably not too much of
a technical problem, but it might be confusing.
Either way, I'd be happy to write a patch that adds the option, and I
can try to do the asking part too if you have an idea about #1 above.
--
((lambda (x) (x x)) (lambda (x) (x x))) Eli Barzilay:
http://barzilay.org/ Maze is Life!
