How do I find out, in a Lisp program, what is the current buffer?

How do I find out, in a Lisp program, what is the current buffer?

[Alan Mackenzie]

Hello, Emacs.

I want to find out what is the current buffer, and I mean the REAL
current buffer.

In particular, if that is the minibuffer, I want to be told it's the (or
a) minibuffer, not the buffer it was invoked from.  The function
current-buffer won't do this.

I want to use this for writing the current buffer into the new position
info in the doc string (bug #67455).  Evaluating a defun in the
minibuffer is different from evaluating one in *scratch*.

Is there some way I can get this information from Lisp, or do I have to
write a new C primitive?

Thanks in advance!

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: How do I find out, in a Lisp program, what is the current buffer?

[Eshel Yaron]

Hi Alan,

Alan Mackenzie <acm@muc.de> writes:

> I want to find out what is the current buffer, and I mean the REAL
> current buffer.
>
> In particular, if that is the minibuffer, I want to be told it's the (or
> a) minibuffer, not the buffer it was invoked from.  The function
> current-buffer won't do this.

My two cents: AFAICT, `current-buffer` does do this, in some sense at
least.  To see that, type `M-x (current-buffer) C-x C-e`, here Emacs
says "#<buffer *Minibuf-1*>".

Note that when you say `M-: (current-buffer) RET` the form is read in
the minibuffer, but it's evaluated in the original buffer.  That's why you
get the original buffer as the return value.
`M-: (current-buffer) C-x C-e` points at the minibuffer as expected OTOH.

> I want to use this for writing the current buffer into the new position
> info in the doc string (bug #67455).  Evaluating a defun in the
> minibuffer is different from evaluating one in *scratch*.
>
> Is there some way I can get this information from Lisp, or do I have to
> write a new C primitive?

IIUC, you need the buffer from which the form was read, not the one in
which it was evaluated.  Perhaps `read_internal_start` could store its
`stream` argument in some variable so if you immediately evaluate a form
with that variable it would tell you where it was read from.  It's just
a rough idea, though.


Good luck,

Eshel

Re: How do I find out, in a Lisp program, what is the current buffer?

[Alan Mackenzie]

Hello, Eshel.

Thanks for such a rapid reply!

On Tue, Dec 05, 2023 at 16:03:18 +0100, Eshel Yaron wrote:
> Hi Alan,

> Alan Mackenzie <acm@muc.de> writes:

> > I want to find out what is the current buffer, and I mean the REAL
> > current buffer.

> > In particular, if that is the minibuffer, I want to be told it's the (or
> > a) minibuffer, not the buffer it was invoked from.  The function
> > current-buffer won't do this.

> My two cents: AFAICT, `current-buffer` does do this, in some sense at
> least.  To see that, type `M-x (current-buffer) C-x C-e`, here Emacs
> says "#<buffer *Minibuf-1*>".

Ah.  Thanks for the clarification.

> Note that when you say `M-: (current-buffer) RET` the form is read in
> the minibuffer, but it's evaluated in the original buffer.  That's why you
> get the original buffer as the return value.
> `M-: (current-buffer) C-x C-e` points at the minibuffer as expected OTOH.

> > I want to use this for writing the current buffer into the new position
> > info in the doc string (bug #67455).  Evaluating a defun in the
> > minibuffer is different from evaluating one in *scratch*.

> > Is there some way I can get this information from Lisp, or do I have to
> > write a new C primitive?

> IIUC, you need the buffer from which the form was read, not the one in
> which it was evaluated.

Yes, indeed.  Thanks!  If I read a form from the minibuffer with M-:,
which then evaluates it, the current buffer for the evaluation is not
the minibuffer any more.

> Perhaps `read_internal_start` could store its `stream` argument in
> some variable so if you immediately evaluate a form with that variable
> it would tell you where it was read from.  It's just a rough idea,
> though.

It might be a good one, though.  I'll have a close look at lread.c, and
see whether or not I can slip that in.

> Good luck,

> Eshel

-- 
Alan Mackenzie (Nuremberg, Germany).