Hi Chandler, this looks like a great start. Like Aaron says, it's missing more stuff, but it reduces the number of errors when compiling ms headers with clang's builtin headers considerably. Can you land this? (with s/inclued/include/ in 2 places and s/te rely/we rely/)
Nico On Mon, Feb 20, 2012 at 12:07 AM, Chandler Carruth <[email protected]> wrote: > On Mon, Feb 20, 2012 at 12:06 AM, Chandler Carruth <[email protected]> > wrote: >> >> On Sun, Feb 19, 2012 at 8:04 PM, Aaron Ballman <[email protected]> >> wrote: >>> >>> Here's what we've figured out so far. >>> >>> Several platform SDK header files include intrin.h from Visual Studio. >>> This file declares all of the intrinsics used by MSVC, including >>> chip-specific intrinsics (MMX, SSE, etc) as well as >>> platform-independent intrinsics (Interlocked***, etc). >>> >>> A full list of intrinsics is documented on MSDN at >>> <http://msdn.microsoft.com/en-us/library/26td21ds.aspx> and >>> <http://msdn.microsoft.com/en-US/library/w5405h95(v=vs.100).aspx> >>> >>> We need to come up with our own intrin.h that supports Microsoft's so >>> that we can compile Win32 applications (and Microsoft's STL >>> implementation). >> >> >> Here is an initial cut at adding an MSVC compatible 'intrin.h' builtin >> header. I've left a FIXME to fill in the MS-specific intrinsics. This should >> at least provide a place to begin fleshing out the intrinsics. > > > Just to emphasize, this is completely untested. =] Just finished typing it > up w/o all the weird include_next logic. > > _______________________________________________ > cfe-commits mailing list > [email protected] > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits > _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
