Re: C-x C-q more useful as toggle-read-only

[Andre Spiegel]

Being the VC maintainer, I probably should have commented on this
earlier on, but I couldn't follow this thread continuously over the
past few days.  The binding of C-x C-q was introduced before I took
over VC maintenance, so I am not to blame for anything :-).

There have already been lengthy discussions about this issue several
times in the past, and for the moment, I have simply chosen not to
have an opinion on the matter and implement whatever people consider
the best.

A few remarks though --

For one thing, I agree that vc-toggle-read-only is only meaningful for
files with locking, i.e. if unmodified files are kept read-only by the
version control system.  Under both RCS and CVS there are modes where
this is the case, and both backends can also work in a mode where this
is not the case.

So maybe it would be natural to have that C-x C-q binding for files
with locking only, and even in that case it could be made into a user
option, whereas C-x v v would be the general VC command for doing the
right thing.

Note though, that this requires a rather drastic change in the VC
documentation, where C-x C-q is consistently documented as the main VC
command.  For that reason, I am also skeptical of introducing a user
option "on the fly" as Eli suggested.  This is not a light issue,
because it would turn the meaning of a very important key stroke into
a mere option.

On the positive side, the new VC does handle the special case that
someone complained about: if a file is "locked" (or "modified") but
the buffer is read-only, then C-x C-q will simply make the buffer (and
if needed, also the file) writable, because that is the most sensible
thing to do in that situation.