Fix build on musl libc with clang.

[ac000]

As reported by @andypost on GitHub, if you try to build Unit on a system that uses musl libc (such as Alpine Linux) with clang then you get the following

clang -c -pipe -fPIC -fvisibility=hidden -O -W -Wall -Wextra -Wno-unused-parameter -Wwrite-strings -fstrict-aliasing -Wstrict-overflow=5 -Wmissing-prototypes -Werror -g -I src -I build/include \ \ \ -o build/src/nxt_socketpair.o \ -MMD -MF build/src/nxt_socketpair.dep -MT build/src/nxt_socketpair.o \ src/nxt_socketpair.c In file included from src/nxt_socketpair.c:8: src/nxt_socket_msg.h:138:17: error: comparison of integers of different signs: 'unsigned long' and 'long' [-Werror,-Wsign-compare] cmsg = CMSG_NXTHDR(&msg, cmsg)) ^~~~~~~ /usr/include/sys/socket.h:358:44: note: expanded from macro 'CMSG_NXTHDR' __CMSG_LEN(cmsg) + sizeof(struct cmsghdr) >= __MHDR_END(mhdr) - (unsigned char *)(cmsg) \

In file included from src/nxt_socketpair.c:8:
src/nxt_socket_msg.h:177:17: error: comparison of integers of different signs: 'unsigned long' and 'long' [-Werror,-Wsign-compare]
         cmsg = CMSG_NXTHDR(&msg, cmsg))
                ^~~~~~~~~~~~~~~~~~~~~~~
/usr/include/sys/socket.h:358:44: note: expanded from macro 'CMSG_NXTHDR'
        __CMSG_LEN(cmsg) + sizeof(struct cmsghdr) >= __MHDR_END(mhdr) - (unsigned char *)(cmsg) \
        ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ^  ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2 errors generated.
make: *** [build/Makefile:261: build/src/nxt_socketpair.o] Error 1

GCC works fine, it seems to have some smarts so that it doesn't give warnings on system header files.

This seems to be a long standing issue with musl libc (bad casting in the CMSG_NXTHDR macro) and the workaround employed by several projects is to disable the -Wsign-compare clang warning for the code in question.

So, that's what we do. We wrap the CMSG_NXTHDR macro in a function, so we can use the pre-processor in it to selectively disable the warning.

[ac000]

Added missing 'Closes' tag.

[andypost]

Thank you, posted review to https://github.com/nginx/unit/issues/936#issuecomment-1708938752 that fixed build on 4 arches

but it still fail on arm(7/hf/64) and riscv64 because of atomics - there's pipelines for clang 15 and 16 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/51157/pipelines

[ac000]

Thank you, posted review to #936 (comment) that fixed build on 4 arches

but it still fail on arm(7/hf/64) and riscv64 because of atomics - there's pipelines for clang 15 and 16 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/51157/pipelines

To avoid any confusion. This is not related to this patch and is being tracked in a separate issue.

[hongzhidao]

LGTM. The only question is if we need to add a BB test or not. @thresheek @ac000 @lcrilly.

[ac000]

Do we want a BB with say, Alpine Linux + clang?

Apart from the extra resources and possible extra time to complete a run, I don't think having more clang coverage would be bad thing in general. AFAICT the only builders that currently use clang are the FreeBSD ones and the code analyser.

[hongzhidao]

Do we want a BB with say, Alpine Linux + clang?

Yep, that's what I mean.

[ac000]

Rebase onto master