Makefile: set _POSIX_C_SOURCE

[ffontaine]

Set -D_POSIX_C_SOURCE=200112L to avoid the following build failure with uclibc-ng:

anet.c: In function ‘anetTcpGenericConnect’:
anet.c:153:21: error: storage size of ‘gai_hints’ isn’t known
  153 |     struct addrinfo gai_hints;
      |                     ^~~~~~~~~

In file included from dump1090.h:83,
                 from dump1090.c:50:
compat/compat.h:46:3: error: conflicting types for ‘clockid_t’
   46 | } clockid_t;
      |   ^~~~~~~~~
In file included from /home/fabrice/buildroot/output/host/or1k-buildroot-linux-uclibc/sysroot/usr/include/sys/types.h:132,
                 from /home/fabrice/buildroot/output/host/or1k-buildroot-linux-uclibc/sysroot/usr/include/stdlib.h:198,
                 from dump1090.h:66,
                 from dump1090.c:50:
/home/fabrice/buildroot/output/host/or1k-buildroot-linux-uclibc/sysroot/usr/include/time.h:91:21: note: previous declaration of ‘clockid_t’ was here
   91 | typedef __clockid_t clockid_t;
      |                     ^~~~~~~~~

net_io.c: In function ‘writeJsonToFile’:
net_io.c:2122:18: error: ‘PATH_MAX’ undeclared (first use in this function)
 2122 |     char pathbuf[PATH_MAX];
      |                  ^~~~~~~~

Signed-off-by: Fabrice Fontaine [email protected]

[mutability]

The manpage I have here says:

           Since glibc 2.22: _POSIX_C_SOURCE >= 200112L
           Glibc 2.21 and earlier: _POSIX_C_SOURCE

CONFORMING TO
       POSIX.1-2001, POSIX.1-2008.  The getaddrinfo() function is documented in RFC 2553.

why does uclibc-ng need _GNU_SOURCE here, isn't struct addrinfo POSIX?

[mutability]

For clockid_t, do you know why the current test fails on uclibc-ng? (maybe CLOCK_REALTIME is not a macro there?). Should we maybe test for _POSIX_TIMERS instead (or as well as) testing for a CLOCK_REALTIME macro?

[mutability]

Missing PATH_MAX must mean something like your _DEFAULT_SOURCE doesn't turn on POSIX, that's really basic POSIX and we're including <limits.h>

[mutability]

I suspect that most of these problems are better solved by defining an appropriate _POSIX_C_SOURCE, rather than _GNU_SOURCE

[ffontaine]

Indeed, this is a better solution, I updated the PR

[mutability]

Merged, thanks!