[M3devel] I386_DARWIN GetPC, rep prefix?

[Tony Hosking]

Hmm.  Not sure I did test that.  I don't have any 10.4 systems around  
anymore.

On 11 May 2009, at 08:33, Jay wrote:

> Tony, to get RTSignalC.c to compile on Darwin (not MacOSX) 8.0.1  
> (aka 10.4).
>
>
>
> jdar1:/dev2/cm3/m3-libs/m3core/src/runtime/POSIX jay$ uname -a
> Darwin jdar1.local 8.0.1 Darwin Kernel Version 8.0.1: Fri Apr 29  
> 12:18:40 PDT 2005; root:xnu-792.obj
> /RELEASE_I386 x86 i386
> jdar1:/dev2/cm3/m3-libs/m3core/src/runtime/POSIX jay$
>
>
> Probably that's a not very interesting target, but I don't yet have  
> "real" x86/AMD64/Apple/Darwin hardware (just PPC/ARM).
>
>
> I had to:
>
>
> ===================================================================
> RCS file: /usr/cvs/cm3/m3-libs/m3core/src/runtime/POSIX/RTSignalC.c,v
> retrieving revision 1.11
> diff -r1.11 RTSignalC.c
> 71c71
> <       pc = Context->uc_mcontext->ss.eip;
> ---
>>      pc = Context->uc_mcontext->sc.sc_eip;
>
>
> Did you test ss.eip?
> And presumably sc.sc_eip doesn't work for you? Unless they have  
> #defines for compat (on such short names..)?
> Maybe this changed from 10.4 to 10.5?
> I just want to change over the Unix/*.i3 files.
> I can leave my change uncommited till I get a real MacOSX>=10.5  
> system.
>
>
> I also get problems even in native builds with "rep" prefixes using  
> the "builtin" tools (as).
> I'll change cm3cg to 586, or 486, or 386, whatever it takes, to  
> workaround that.
> Also not commit that.
> Or maybe build a newer "cctools".
>
>
> - Jay