At 12:13 PM 10/9/2001 -0400, Michael Maraist wrote:
>The key was &&lbl which returns the physical address of the a jumpable
>label. This is highly gcc-specific, but given that we can use
>#ifdef's I don't see a problem with it. :)
No #ifdefs here. Instead the contents of the DO_OP macro should be
dynamically generated at configure time. (Which is OK, since we need to eat
all the .ops files at compile time anyway)
The results look really interesting, though. I'm tempted to see how it
works with potentially 16-bit opcodes. (Which is where noops would come in
handy as padding to guarantee the 32-bit inlined integers were properly
aligned)
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
