Mark negative-zero-to-int conversion as Inexact,
authorDale Johannesen <dalej@apple.com>
Tue, 7 Oct 2008 00:40:01 +0000 (00:40 +0000)
committerDale Johannesen <dalej@apple.com>
Tue, 7 Oct 2008 00:40:01 +0000 (00:40 +0000)
commite4a424590675cf32b729420665626774b01b114d
tree5627bd144f1706ff2b9890435a148a21dacdf8ee
parentb5cc34d098149a9606dacb6d421471d787253907
Mark negative-zero-to-int conversion as Inexact,
since -0.0 can't be represented as an int.
This prevents llvm from reducing -0.0 to a
load-and-convert of int 0.  Fixes
ieee.exp/mzero[2356].c in gcc testsuite.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@57224 91177308-0d34-0410-b5e6-96231b3b80d8
lib/Support/APFloat.cpp