Re: Adding refactoring capabilities to Emacs

[Dmitry Gutov]

On 26/09/2023 14:24, João Távora wrote:
> On Tue, Sep 26, 2023 at 11:57 AM Dmitry Gutov<dmitry@gutov.dev>  wrote:
> > On 26/09/2023 11:06, João Távora wrote:
> > > On Tue, Sep 26, 2023 at 6:36 AM Alfred M. Szmidt<ams@gnu.org>   wrote:
> > >
> > > > If you have a diff on file, you are most probobly going to apply it,
> > > > and also probobly going to remove a hunk or two or edit the diff in
> > > > some manner.  (That this is "relatively rare" I disagree from my own
> > > > usage and experience).  Not to mention that visiting a file on disk,
> > > > that is read-write, and Emacs making it read-only would be very
> > > > strange.
> > > I completely agree with these two points.  Even non-file diff-mode
> > > buffers, such as the ones provided by piping git diff into Emacs
> > > (yes, I can do that 😄 ) are generally better left read-write,
> > > since I frequently edit them to kill hunks I'm not interested in.
> > 'k' (or M-k), 'C-c C-s' and 'C-_' all work fine in a read-only diff-mode
> > buffer. 'C-x C-s' also works, of course.
> I think it's very inconsistent to have specialized commands to modify
> a buffers contents and not allow all the other regular commands that
> modify a buffer do their work.  I don't have unlimited brain address
> space for keybindings and I think C-SPC C-n a few times C-w does
> the job just fine.

Is that better than typing 'k'?

Or being able to use hunk navigation with n/p?

> Opening regular files of a special type read-only mode would be a
> spectacular failure in the basic ergonomics of an editor.

Like you said: you edit both kinds of diff buffers (vc and files on 
disk) in the same way. But we assign them different read-only status.