See ziglang/fetch-them-macos-headers#2 for more details. The path
forward looks like one of the following:
* Ony provide headers for the oldest supported macOS (currently 10.13
but soon to be bumped to 10.14).
* Provide headers for multiple versions, and select based on the Zig
target OS version range minimum.
* Don't try to provide macOS headers.
If we don't tackle the version problem, we would have to re-introduce
the ability to detect and depend on native system headers if we wanted
to support C/C++ code that used newer OS definitions.
This patch also adds support for `#include <mach/mach_time.h>`.
Also related: #5236
these new files are generated by recursively including:
sysexits.h
mach/mach.h
sys/attr.h
sys/mount.h
crt_externs.h
execinfo.h
all of which are depended on by LLVM libraries.
This reverts commit bb9c3118ed5fdc16b8e2d9882375005c2a62d0cc, reversing
changes made to 7015d84e0ca6f02fede45621571084df98dda712.
This is missing quite a few headers
previously zig would look for all the .defs even when not needed,
wasting time. also, we only had samsrv definitions for some architectures.
remove cross compiling support for this.
Clang pretends to be gcc 4.4 and that causes some re-definition errors.
The problem has been reported to the upstream some time ago but nothing
was done about it.
From: Szabolcs Nagy <szabolcs.nagy@arm.com>
armv8 removed the coprocessor instructions other than cp14, so
on an armv8 system the related hwcaps should never be set.
new llvm complains about the use of coprocessor instructions in
armv8-a mode (even though they are never executed at runtime),
so ifdef them out when musl is built for armv8.
<dalias> i think the patch looks ok
It was using __GNUC__ < 8 to check if _xgetbv intrinsic is available.
Since we always use zig cc with this header, we know we have clang 9,
which does have this intrinsic.
