[llvm-commits] [llvm] r43844 - in /llvm/trunk/lib/Target/X86: X86RegisterInfo.cpp X86RegisterInfo.td

[Evan Cheng]

On Nov 7, 2007, at 2:01 PM, Anton Korobeynikov wrote:

> Dale,
>
>> Complete conditionalization of Dwarf reg numbers.
> This looks pretty hackish. The numbers used before were definitely
> correct for x86-32/linux. I dunno about darwin, but it looks like
> "turning everything on darwin to be default implementation" is not a
> good approach. :)

The numbers were wrong from Darwin and x86-64. I think the approach  
is fine. But pleasejust remove the numbers from X86RegisterInfo.td  
since they are wrong for some of the targets. It's misleading to  
leave them in.

>
> And, if you need to conditionalize depending on target, table approach
> (like in i386.h) will be definitely better, otherwise we'll end with
> spaghetti-like code, when all supported targets will be in place.
>

That's good advice. Is that easy to do, Dale?

Evan
> -- 
> With best regards, Anton Korobeynikov.
>
> Faculty of Mathematics & Mechanics, Saint Petersburg State University.
>
>
> _______________________________________________
> llvm-commits mailing list
> llvm-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits