Mail Archives: pgcc/1999/07/21/09:39:44
> anything _more_ specific than "that did not work"?
ops !
>
>> on an amd k6-2 . there is no problem with pgcc-1.1.3 that it accepts the
>> code
>
>And still miscompiles it quite as often.
>
>> is it a problem of gcc-2.95 ? or is the code just broken ?
>
>Just broken. However, declaring them as dummy outputs instead of as
>clobbers would make it legal.
ok . i got the answer from the gcc-2.95 docs - localized register
spilling . thanx.
eRHan Bilgili
- Raw text -