[LLVMdev] Floating point instructions patch

[Chris Lattner]

On Fri, 29 Apr 2005, Morten Ofstad wrote:
> New patch here -- it's not been tested yet because we're having some problems 
> with the application (I can't create new VM programs at the moment), but it 
> compiles OK ;-) Please look over it and see if there are some more changes 
> you'd like me to make before you can commit it...

Looks great.  I committed your patches here:
http://mail.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20050425/025913.html
...
http://mail.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20050425/025927.html

I made the following changes:
1. The legalize stuff wasn't quite right.
2. I told all non-x86 targets that FSIN/FCOS/FSQRT were not legal for
    them yet.
3. I added a regression test here:
     Regression/CodeGen/Generic/intrinsics.ll

Overall the patch looked great, thanks a lot!

-Chris

-- 
http://nondot.org/sabre/
http://llvm.cs.uiuc.edu/