bug#46824: 28.0.50; nativecomp crash with cl-block/cl-defun

[Pip Cet]

On Tue, Mar 2, 2021 at 1:28 PM Eli Zaretskii <eliz@gnu.org> wrote:
> > From: Pip Cet <pipcet@gmail.com>
> > Date: Tue, 2 Mar 2021 09:16:12 +0000
> > Cc: 46824@debbugs.gnu.org, Aaron Jensen <aaronjensen@gmail.com>
> >
> > +  gcc_jit_param *params[] =
> > +    {
> > +      gcc_jit_context_new_param (comp.ctxt, NULL, comp.void_ptr_type, 
> > "buf"),
> > +    };
>
> Nitpicking: the braces here should be in column zero.

How about we compromise and put them in column two rather than four or zero? :-)

(I hope that's what you meant).

I was going to apologize for apparently messing up my cc-mode setup
here, but emacs -Q also insists on the incorrect brace placement. We
might want to fix that somehow for those Emacs developers who use this
"Emacs" thing.

> > +  /* Don't call setjmp through a function pointer (Bug#46824) */
> > +  return gcc_jit_context_new_call (comp.ctxt,
> > +                                NULL,
> > +                                gcc_jit_context_new_function
> > +                                (comp.ctxt, NULL, 
> > GCC_JIT_FUNCTION_IMPORTED,
> > +                                 comp.int_type, STR (SETJMP_NAME),
> > +                                 ARRAYELTS (params), params,
> > +                                 false),
> > +                                1,
> > +                                args);
>
> Another nit: in cases like this one I find this style more readable:
>
>   return gcc_jit_context_new_call (comp.ctxt, NULL,
>                                    gcc_jit_context_new_function (comp.ctxt,
>                                                                  NULL,
>                                                                  
> GCC_JIT_FUNCTION_IMPORTED,
>                                                                  ...
>
> IOW, avoid having the function's name and its arguments on separate
> lines, as I find that harder to read.
>
> Or maybe even this:
>
>   gcc_jit_rvalue *val =
>     gcc_jit_context_new_call (comp.ctxt, NULL,
>                               gcc_jit_context_new_function (comp.ctxt, NULL,
>                                                             
> GCC_JIT_FUNCTION_IMPORTED,
>                                                             ...),
>                               1, args);
>   return val;

I've decided on the attached, if it's okay.

> Thanks.

Thank you!

0001-Don-t-call-_setjmp-through-a-function-pointer-Bug-46.patch
Description: Text Data