byte-compile-preprocess+cconv seem to mutate self evaluating forms in expanded macros

byte-compile-preprocess+cconv seem to mutate self evaluating forms in expanded macros

[Vibhav Pant]

[-- Attachment #1: Type: text/plain, Size: 1543 bytes --]

While working on scratch/comp-static-data, which adds support for
native compiling self evaluating forms in Lisp code as static consts in
the eln (with immutability enforced additionally by `CHECK_IMPURE`), I
noticed that compiling minor-mode definitions would result in a
`pure_write_error`. This arises as cconv-closure-convert, called from
`byte-compile-preprocess` would call `setcar` on a self evaluating
interactive form as part of the minor-mode's function body.

A hacky way to reproduce this would be to stuff the following snippet
into loadup.el (purecopy ensures modifying the list triggers an error):

```
(load "emacs-lisp/bytecomp")
(setq sample-interactive-spec
      (purecopy '(interactive
                  (list (if current-prefix-arg
                            (prefix-numeric-value 
                             current-prefix-arg)
                          'toggle)))))

(defmacro define-purecopied-func ()
  `(defun foo-bar (arg)
     ,sample-interactive-spec))

(let ((byte-compile-debug t))
  (byte-compile '(define-purecopied-func)))
```
(Alternatively, just building scratch/comp-static-data will also
reproduce these errors)

Because modifying self-evaluating forms is technically undefined
behaviour as per the Elisp reference manual
(https://www.gnu.org/software/emacs/manual/html_node/elisp/Mutability.html),
shouldn't we be use 'copy-tree' after macro expansion in these cases?


-- 
Vibhav Pant
vibhavp@gmail.com
GPG: 7ED1 D48C 513C A024 BE3A  785F E3FB 28CB 6AB5 9598

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: byte-compile-preprocess+cconv seem to mutate self evaluating forms in expanded macros

[Stefan Monnier]

Hi,

> While working on scratch/comp-static-data, which adds support for
> native compiling self evaluating forms in Lisp code as static consts in
> the eln (with immutability enforced additionally by `CHECK_IMPURE`), I
> noticed that compiling minor-mode definitions would result in a
> `pure_write_error`. This arises as cconv-closure-convert, called from
> `byte-compile-preprocess` would call `setcar` on a self evaluating
> interactive form as part of the minor-mode's function body.

Sounds like a bug.  Can you `M-x report-emacs-bug` (and put me in
`X-Debbugs-Cc:`) so we get a bug#nb for it?


        Stefan

Re: byte-compile-preprocess+cconv seem to mutate self evaluating forms in expanded macros

[Vibhav Pant]

[-- Attachment #1: Type: text/plain, Size: 1073 bytes --]

On Fri, 2023-01-20 at 08:03 -0500, Stefan Monnier wrote:
> Hi,
> 
> > While working on scratch/comp-static-data, which adds support for
> > native compiling self evaluating forms in Lisp code as static
> > consts in
> > the eln (with immutability enforced additionally by
> > `CHECK_IMPURE`), I
> > noticed that compiling minor-mode definitions would result in a
> > `pure_write_error`. This arises as cconv-closure-convert, called
> > from
> > `byte-compile-preprocess` would call `setcar` on a self evaluating
> > interactive form as part of the minor-mode's function body.
> 
> Sounds like a bug.  Can you `M-x report-emacs-bug` (and put me in
> `X-Debbugs-Cc:`) so we get a bug#nb for it?
> 
> 
>         Stefan
> 

Hi Stefan,

The error doesn't come up on master, because we don't enforce the
immutability of self evaluating forms. Becuase the same files get
compiled just fine on master, does it still count as a bug?

Thanks,
Vibhav

-- 
Vibhav Pant
vibhavp@gmail.com
GPG: 7ED1 D48C 513C A024 BE3A  785F E3FB 28CB 6AB5 9598

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: byte-compile-preprocess+cconv seem to mutate self evaluating forms in expanded macros

[Stefan Monnier]

> The error doesn't come up on master, because we don't enforce the
> immutability of self evaluating forms. Becuase the same files get
> compiled just fine on master, does it still count as a bug?

Yes.


        Stefan

Re: byte-compile-preprocess+cconv seem to mutate self evaluating forms in expanded macros

[Vibhav Pant]

[-- Attachment #1: Type: text/plain, Size: 427 bytes --]

On Fri, 2023-01-20 at 13:21 -0500, Stefan Monnier wrote:
> > The error doesn't come up on master, because we don't enforce the
> > immutability of self evaluating forms. Becuase the same files get
> > compiled just fine on master, does it still count as a bug?
> 
> Yes.
> 
> 
>         Stefan
> 

Done, thanks.

-- 
Vibhav Pant
vibhavp@gmail.com
GPG: 7ED1 D48C 513C A024 BE3A  785F E3FB 28CB 6AB5 9598

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 488 bytes --]