Re: Compiling MiNT with gcc 2.5.8

[cherborth@semprini.waterloo-rdp.on.ca (Chris Herborth)]

In <9411141032.AA18059@math.uni-muenster.de>, Ulrich Kuehn writes:
>Hello everybody!
>
>There has been the problem of a hanging AES 4.1, when mint was compiled
>with gcc 2.5.8, but not when it was made with 2.3.3.
>
>Now I found out, which file is causing the problems: it is proc.c
>If everything is made with gcc 2.5.8, except proc.c is done with 2.3.3,
>the AES still works the normal way.... but beware of a full 2.5.8
>version!

If you compile proc.c with absolutely no optimisation (ie, remove -O2
and -f<anything>), does it work?

>The big question now is: WHY????

2.5.x (x < 8) had a _lot_ of problems with 68k, and plenty of optimiser
problems, too.  There must be another one that nobody's noticed (since
almost nobody other than us makes 68000 code with gcc).

-- 
-------------------======================================-------------------
 |||  Jaguar 64-bit                    cherborth@semprini.waterloo-rdp.on.ca
/ | \ DO THE MATH                                             Chris Herborth