On Wed, Oct 6, 2010 at 1:42 PM, Chris Lattner <[email protected]> wrote: > > On Oct 6, 2010, at 4:39 AM, Sebastian Redl wrote: > >> On 05.10.2010 19:01, Matt Beaumont-Gay wrote: >>> See attached. This removes a dependency on errno.h and improves >>> compatibility with GCC's implementation. >>> >> Do we want an unconditional dependency on the existence of >> posix_memalign? What about Win32 platforms? > > I'm not familiar with windows, does it have posix_memalign replacement? If > so, it's probably best to #ifdef _WIN32 in the header. If not, it seems > reasonable for the old implementation can come back #ifdef _WIN32. > > -Chris
http://msdn.microsoft.com/en-us/library/8z34s9c6.aspx - Michael Spencer _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
