GCC warning in minibuf.c

GCC warning in minibuf.c

[Eli Zaretskii]

With today's CVS, GCC 3.4.2 issues the following warning:

  minibuf.c: In function `Fminibuffer_completion_help':
  minibuf.c:2578: warning: passing arg 2 of `internal_with_output_to_temp_buffer' from incompatible pointer type

This happens because internal_with_output_to_temp_buffer is declared
as follows:

  Lisp_Object
  internal_with_output_to_temp_buffer (bufname, function, args)
       const char *bufname;
       Lisp_Object (*function) P_ ((Lisp_Object));
       Lisp_Object args;

Here, the 2nd argument `function' is declared to accept a single
Lisp_Object argument.  However, minibuf.c calls
internal_with_output_to_temp_buffer as follows:

  internal_with_output_to_temp_buffer ("*Completions*",
				       Fdisplay_completion_list,
				       Fsort (completions, Qstring_lessp));

and Fdisplay_completion_list is declared to accept 2 Lisp_Object
arguments.

I think this is a bogus warning, but I'm not sure what would be the
best way to shut up the compiler.  One way to do that is to modify the
above declaration of internal_with_output_to_temp_buffer like this:

  Lisp_Object
  internal_with_output_to_temp_buffer (bufname, function, args)
       const char *bufname;
       Lisp_Object (*function) ();
       Lisp_Object args;

However, the current full-prototype declaration was introduced (in
1998 by Andreas) instead of the incomplete one, so it sounds like a
step backwards.  OTOH, Andreas wanted to fix warnings under
the -Wimplicit option, which we don't use in compiling Emacs.

Suggestions?

Re: GCC warning in minibuf.c

[Andreas Schwab]

Eli Zaretskii <eliz@gnu.org> writes:

> Here, the 2nd argument `function' is declared to accept a single
> Lisp_Object argument.  However, minibuf.c calls
> internal_with_output_to_temp_buffer as follows:
>
>   internal_with_output_to_temp_buffer ("*Completions*",
> 				       Fdisplay_completion_list,
> 				       Fsort (completions, Qstring_lessp));
>
> and Fdisplay_completion_list is declared to accept 2 Lisp_Object
> arguments.

Then Fdisplay_completion_list is definitely the wrong function to pass to
internal_with_output_to_temp_buffer.

> I think this is a bogus warning,

I don't agree.  Fdisplay_completion_list will be called with only a single
parameter and will receive garbage in the second one.  A sure way to make
Emacs crash.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: GCC warning in minibuf.c

[Eli Zaretskii]

> From: Andreas Schwab <schwab@suse.de>
> Cc: emacs-devel@gnu.org
> Date: Thu, 20 Oct 2005 15:07:47 +0200
> 
> > I think this is a bogus warning,
> 
> I don't agree.  Fdisplay_completion_list will be called with only a single
> parameter and will receive garbage in the second one.  A sure way to make
> Emacs crash.

Perhaps I misunderstand what the code does, but it looks like
internal_with_output_to_temp_buffer invokes the `function' argument
with a list of arguments, which seems to suggest that any number of
arguments will do.

Re: GCC warning in minibuf.c

[Andreas Schwab]

Eli Zaretskii <eliz@gnu.org> writes:

> Perhaps I misunderstand what the code does, but it looks like
> internal_with_output_to_temp_buffer invokes the `function' argument
> with a list of arguments,

???

  val = (*function) (args);

Looks very much like passing a single argument.

Andreas.

-- 
Andreas Schwab, SuSE Labs, schwab@suse.de
SuSE Linux Products GmbH, Maxfeldstraße 5, 90409 Nürnberg, Germany
Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: GCC warning in minibuf.c

[Kim F. Storm]

Andreas Schwab <schwab@suse.de> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>> Perhaps I misunderstand what the code does, but it looks like
>> internal_with_output_to_temp_buffer invokes the `function' argument
>> with a list of arguments,
>
> ???
>
>   val = (*function) (args);
>
> Looks very much like passing a single argument.
>

Indeed. I have installed a fix.  Thanks.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk