Re: How do I tell when the current buffer is the minibuffer?

[Alan Mackenzie]

Hello, Eli.

On Sun, Mar 06, 2016 at 07:44:25PM +0200, Eli Zaretskii wrote:
> > Date: Sat, 5 Mar 2016 23:20:43 +0000
> > From: Alan Mackenzie <address@hidden>
> > Cc: address@hidden

> > On Sun, Mar 06, 2016 at 12:03:50AM +0100, Andreas Schwab wrote:
> > > Alan Mackenzie <address@hidden> writes:

> > > > On Sat, Mar 05, 2016 at 11:40:15PM +0100, Andreas Schwab wrote:
> > > >> Alan Mackenzie <address@hidden> writes:

> > > >> > How does one correctly create a buffer local variable in C?  What I 
> > > >> > have
> > > >> > at the moment is:

> > > >> >     DEFSYM (Qcomment_depth_hwm, "comment-depth-hwm");
> > > >> >     DEFVAR_LISP ("comment-depth-hwm", Vcomment_depth_hwm,
> > > >> >                  doc: /* Buffer position below which the 
> > > >> > `comment-depth' property is valid.  */);
> > > >> >     Vcomment_depth_hwm = make_number (1);
> > > >> >     Fmake_variable_buffer_local (Qcomment_depth_hwm);

> > > >> > This is clearly inadequate, of itself, to create a buffer local 
> > > >> > variable.

> > > >> In which way?

> > > > Until a `setq' (or the like) is done on the variable for a particular
> > > > buffer, C code reading/writing from Vcomment_depth_hwm reads and writes
> > > > the same piece of RAM that all other buffers do.

> > > That seems to work for deactivate-mark.

> > It's not working for comment-depth-hwm.

> > Perhaps somebody who understands this will chip in, here.  It could
> > really do with being in the Elisp manual.

> Is there still a problem?

No, I solved it by putting the variable into buffer.c, accessed via BVAR.
I'm still not sure if this is the canonical way to create a buffer local
variable accessed from C.

> If so, please describe the details, because I don't think I understand
> them well enough to try helping you.  (Yes, I've read what you wrote,
> but your description was too general and didn't include any code you
> tried to use, so I'm not sure what exactly did you try and how.)

The problem was assigning a value to Vcomment_depth_hwm in C didn't
create a buffer local copy of `comment-depth-hwm'.  Why should it?

I think the time it took me to find the solution is a problem.  A 2003
copy of the XEmacs internals manual was helpful.  ;-)

-- 
Alan Mackenzie (Nuremberg, Germany).