[llvm] r210719 - [FastIsel][X86] Add support for lowering the first 8 floating-point arguments.

[Juergen Ributzka]

I am going to revert it.

On Jun 12, 2014, at 12:11 PM, Marshall Clow <mclow.lists at gmail.com> wrote:

> 
> On Jun 11, 2014, at 4:10 PM, Juergen Ributzka <juergen at apple.com> wrote:
> 
>> Author: ributzka
>> Date: Wed Jun 11 18:10:58 2014
>> New Revision: 210719
>> 
>> URL: http://llvm.org/viewvc/llvm-project?rev=210719&view=rev
>> Log:
>> [FastIsel][X86] Add support for lowering the first 8 floating-point arguments.
> 
> I’m seeing a lot (~50) of failures in the libc++ test suite with this change. (Mac OS X)
> I’m testing the cases in libcxx/test/localization/locales/locale.convienience/conversions/conversion.buffer.
> 
> Using a compiler built from 210718, these tests pass.
> Using a compiler built from 210719, these tests segfault.
> 
> — Marshall
> 
> 
>