bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

[Augusto Stoffel]

If you call pp-macroexpand-last-sexp in a buffer with lexical binding,
the resulting *Pp Macroexpand Output* buffer has lexical binding turned
off.

The problem with this is that if you now go the *Pp Macroexpand Output*
buffer and call pp-macroexpand-last-sexp again, this second macro
expansion step may not reflect what would happen in the original buffer;
I don't recall an example, but some macros expand differently in the two
variable scoping modes.

bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

[Eli Zaretskii]

> From: Augusto Stoffel <arstoffel@gmail.com>
> Date: Tue, 02 Apr 2024 08:01:18 +0200
> 
> If you call pp-macroexpand-last-sexp in a buffer with lexical binding,
> the resulting *Pp Macroexpand Output* buffer has lexical binding turned
> off.
> 
> The problem with this is that if you now go the *Pp Macroexpand Output*
> buffer and call pp-macroexpand-last-sexp again, this second macro
> expansion step may not reflect what would happen in the original buffer;
> I don't recall an example, but some macros expand differently in the two
> variable scoping modes.

Adding Stefan.

bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

[Michael Heerdegen via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Augusto Stoffel <arstoffel@gmail.com> writes:

> If you call pp-macroexpand-last-sexp in a buffer with lexical binding,
> the resulting *Pp Macroexpand Output* buffer has lexical binding turned
> off.

Same as with any other displayed pp output buffer btw, or any temp
buffer.  All inherit the global lexical-binding -> nil binding.  Not
ideal indeed.

Michael.

bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

[Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> If you call pp-macroexpand-last-sexp in a buffer with lexical binding,
> the resulting *Pp Macroexpand Output* buffer has lexical binding turned off.
>
> The problem with this is that if you now go the *Pp Macroexpand Output*
> buffer and call pp-macroexpand-last-sexp again, this second macro
> expansion step may not reflect what would happen in the original buffer;
> I don't recall an example, but some macros expand differently in the two
> variable scoping modes.

Indeed.  `dolist` and `dotimes` were such examples (tho nowadays they
return the same result).  Most other examples I can think of signal
an error when used without lexical binding.  There are a few more subtle
examples where the macro tests `macroexp--dynamic-variable-p`.

Do you happen to have a patch to propagate the value of
`lexical-binding` to the new buffer?


        Stefan

bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

[Augusto Stoffel]

[-- Attachment #1: Type: text/plain, Size: 461 bytes --]

On Wed, 10 Apr 2024 at 17:39, Stefan Monnier wrote:

> Do you happen to have a patch to propagate the value of
> `lexical-binding` to the new buffer?

This would be the patch to propagate the lexical binding value.  (A less
magical alternative I guess would be to just always set it to t.  Who
needs to debug non-lexical code these days anyway?)

It doesn't cover the new `pp-insert-short-sexp'.  Here again setting
lexical-binding=t by default might be okay.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-Propagate-lexical-binding-value-to-pp-buffers.patch --]
[-- Type: text/x-patch, Size: 1281 bytes --]

From 8d81704d61153f13de516475f4e0ab53e8e94f11 Mon Sep 17 00:00:00 2001
From: Augusto Stoffel <arstoffel@gmail.com>
Date: Sun, 14 Apr 2024 09:36:05 +0200
Subject: [PATCH] Propagate 'lexical-binding' value to pp buffers

See bug#70137.

* lisp/emacs-lisp/pp.el (pp-display-expression): Set lexical-binding to
match the value in the calling buffer.
---
 lisp/emacs-lisp/pp.el | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lisp/emacs-lisp/pp.el b/lisp/emacs-lisp/pp.el
index d586fc59939..f89807c37be 100644
--- a/lisp/emacs-lisp/pp.el
+++ b/lisp/emacs-lisp/pp.el
@@ -327,7 +327,8 @@ pp-display-expression
 
 If a temporary buffer is needed for representation, it will be named
 after OUT-BUFFER-NAME."
-  (let* ((old-show-function temp-buffer-show-function)
+  (let* ((lexical lexical-binding)
+         (old-show-function temp-buffer-show-function)
 	 ;; Use this function to display the buffer.
 	 ;; This function either decides not to display it at all
 	 ;; or displays it in the usual way.
@@ -357,6 +358,7 @@ pp-display-expression
         (pp expression))
       (with-current-buffer standard-output
 	(emacs-lisp-mode)
+        (setq lexical-binding lexical)
 	(setq buffer-read-only nil)
         (setq-local font-lock-verbose nil)))))
 
-- 
2.44.0

bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

[Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors]

> This would be the patch to propagate the lexical binding value.

LGTM.  Can you install it, or do you want me to do it?

> (A less magical alternative I guess would be to just always set it to
> t.  Who needs to debug non-lexical code these days anyway?)

Not sure why you think it'd be less magical.  🙂

> It doesn't cover the new `pp-insert-short-sexp'.  Here again setting
> lexical-binding=t by default might be okay.

`pp-insert-short-sexp` should use `lisp-data-mode` rather than
`emacs-lisp-mode` anyway.


        Stefan

bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

[Augusto Stoffel]

On Wed, 17 Apr 2024 at 13:43, Stefan Monnier wrote:

> LGTM.  Can you install it, or do you want me to do it?

I don't have commit access to the repo and I absolutely don't mind the
status quo :-).

bug#70137: 30.0.50; pp-macroexpand-last-sexp buffers should turn lexical binding on

[Stefan Monnier via Bug reports for GNU Emacs, the Swiss army knife of text editors]

Thanks, pushed to `master`, closing,


        Stefan