On Fri, 4 Dec 2020, Ozkan Sezer wrote:
On 12/4/20, Martin Storsjö <[email protected]> wrote:
The headers have a number of brittle workarounds, all trying to make
the "#define _mm_malloc _aligned_malloc" redirect in malloc.h work
properly.
Is this https://sourceforge.net/p/mingw-w64/bugs/192/
(after 10 years) ?
Yes, it seems like that. In C, a missing declaration is merely a warning,
but in C++ it's fatal, and it became an issue in libc++ recently after
https://reviews.llvm.org/D91379.
The prime issue itself looks like it'd be possible to fix by just removing
the "#define _MM_MALLOC_H_INCLUDED" in intrin.h, but the main issue would
remain - depending on header include order, you'd either get things
redirecting to GCC's renamed _mm_malloc function, or the real
_aligned_malloc, which I guess can be fatal depending if things are
allocated and freed by different object files - so I thought about trying
to get rid of the inconsistency at the same time.
// Martin
_______________________________________________
Mingw-w64-public mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/mingw-w64-public
