[PATCH] Reassociate x + -0.1234 * y into x - 0.1234 * y

[Erik Verbruggen]

Hi Chad,

I merged in your patch. I also got rid of the "only do it for
non-fast-math". It was bogus, and only there to hide the fact that I
didn't set/copy any math flags to the newly created binop.

http://reviews.llvm.org/D4904

Files:
  lib/Target/README.txt
  lib/Transforms/Scalar/Reassociate.cpp
  test/Transforms/Reassociate/fast-basictest.ll
  test/Transforms/Reassociate/liftsign.ll
-------------- next part --------------
A non-text attachment was scrubbed...
Name: D4904.12664.patch
Type: text/x-patch
Size: 8056 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20140819/8fc8fa80/attachment.bin>