A gcc compiled on musl does not define __gnu_linux__, it defines __linux__.
Only on glibc does __gnu_linux__ get defined, but both define __linux__, so
we should check for that instead.
With this patch, libffi works perfectly, and passes its testsuite entirely
on musl libc systems.
An earlier patch added --disable-docs, but went too far, making it
impossible to build the docs.
It turns out that Automake seemingly has a bug preventing the
conditional build of an info file. So, this patch works around the
bug by putting the info_TEXINFOS rule into a new doc/Makefile.am.
Tested by building with and without --disable-docs and looking for the
existence of doc/libffi.info.
Right now it concatenates it with the existing options and then appends
it to that list, fix it to simply append it as is, same as it is done
with the other variables.
Tested by running the following command which includes gcc options:
$ make check RUNTESTFLAGS="--tool_opts '-Werror'"
Without this patch, all the tests fail. With it, the test succeed.
Inspecting the logs shows that -Werror was indeed used when compiling
the test sources.
The commit fb25cd0 went a bit too far and removed safeseh
when -clang-cl was passed, but that's only needed in x86-64
which is already handled by the -m64 flag.
I discovered this when building Firefox x86 with clang-cl.
In some environments, installing the library to GCC's
multi-os-directory is not perferable. This patch adds a switch
to configure "--disable-multi-os-directory" which will disable
that check, typically falling back to ${libdir} unless
cross-compiling.
Original patch was written by Stewart Brodie, and was found at
https://sourceware.org/ml/libffi-discuss/2013/msg00144.html
I've just updated it to work with the current version.
ppc32 starts using the stack for integer arg passing when we run out
of integer arg passing registers. Similarly, we start using the stack
for floating point args when we run out of floating point registers.
The decision on where an integer arg goes does not depend on number of
floating point args, nor does the decision on where a floating point
arg goes depend on number of integer args. Alignment of stack args
also simply depends on number of stack args.
This patch untangles the horrible mess we had, with intarg_count being
wrongly used to count both integer args and stack words.
* src/powerpc/ffi_sysv.c (ffi_prep_cif_sysv_core): Count fprs,
gprs, and stack words separately.
(ffi_prep_args_SYSV): Similarly.
