Re: gcc 2.95.2 PPC code generation bug(New version!)

[Ralph Schmidt <laire at basis dot owl dot de>]

You wrote:
> Sorry for the wrong From: header in my last email, I mistakenly deleted my 
> configuration while switching to KDE2.
> 
> 
> On Fri, 28 Apr 2000, Ralph Schmidt wrote:
> > > The patch by Mark fixed the problem for the specific test case
> > but it doesn't fix the problem in a bigger scope.
> >
> > Please take my first bugreport and look at the generated code
> > to see where the problem is.
> 
> But your first bugreport was buggy itself and I told you that. You have to 
> change your macros to produce valid statement expressions. Your original 
> macros produced sth like that:
> 
> int result = ({
> struct test mytest;
> ({
> int testi;
> testi = func (&mytest);
> testi;
> })
> });
> 
> You expect TESTI to be assigned to RESULT, but I think this is not covered
> by the definition of statement expressions in the GCC manual (at least not
> to my understanding). I believe you have to rewrite your macros to produce
> code like that:
> 

Ok..i will do that but besides not being my design(I only continued using a
proven design for my playfield) it works perfectly with 68k and with PPC for
*everything* besides the array case where it randomly fails.
So i wouldn't blaim this problem on the "maybe" wrong { level.


> int result = ({
> struct test mytest;
> int testi;
> testi = func (&mytest);
> testi;
> });
> 
> I don't see a reason for the extra level of braces anyway.
> 

I agree with you...but well..i just kept using a year old concept.


> Franz.
> 
> 
---
Ralph Schmidt,laire@popmail.owl.de(private),NextMail welcome