Re: A patch to constify gcc.c (Really, summarizing remaining warnings)

[Zack Weinberg <zack at rabi dot columbia dot edu>]

On Thu, 11 Mar 1999 11:08:45 -0500 (EST), "Kaveh R. Ghazi" wrote:
>	I attempted to figure this out myself and I poked around one
>of the places where the warning occurs.  One triggering site is in
>cse.c:simplify_relational_operation, where it calls setjmp().  The
>particular code in question does this:
>
> >   else if (GET_CODE (op0) == CONST_DOUBLE && GET_CODE (op1) == CONST_DOUBLE
> >            && GET_MODE_CLASS (GET_MODE (op0)) == MODE_FLOAT)
> >     {
> >       REAL_VALUE_TYPE d0, d1;
> >       jmp_buf handler;
> >  
> >       if (setjmp (handler))
> >         return 0;
> >  
> >       set_float_handler (handler);
> >       REAL_VALUE_FROM_CONST_DOUBLE (d0, op0);
> >       REAL_VALUE_FROM_CONST_DOUBLE (d1, op1);
> >       equal = REAL_VALUES_EQUAL (d0, d1);
> >       op0lt = op0ltu = REAL_VALUES_LESS (d0, d1);
> >       op1lt = op1ltu = REAL_VALUES_LESS (d1, d0);
> >       set_float_handler (NULL_PTR);
> >     }

I bet something like this will have the same effect as rth's suggestion:

    {
      jmp_buf handler;
 
      if (setjmp (handler))
        return 0;
 
      set_float_handler (handler);
      {
        REAL_VALUE_TYPE d0, d1;

        REAL_VALUE_FROM_CONST_DOUBLE (d0, op0);
        REAL_VALUE_FROM_CONST_DOUBLE (d1, op1);
        equal = REAL_VALUES_EQUAL (d0, d1);
        op0lt = op0ltu = REAL_VALUES_LESS (d0, d1);
        op1lt = op1ltu = REAL_VALUES_LESS (d1, d0);
      }
      set_float_handler (NULL_PTR);
    }

d0 and d1 are out of scope of the longjmp target, so it should be
clear to gcc that we don't care about the values if we jump.

zw