[MPlayer-cvslog] r38016 - trunk/configure

al subversion at mplayerhq.hu
Tue Jan 9 23:47:14 EET 2018


Author: al
Date: Tue Jan  9 23:47:13 2018
New Revision: 38016

Log:
configure: Fix detection for inet_aton function

Unbreak the build with internal FFmpeg. The misdetection leads to a
build failure with modern FFmpeg because in that case FFmpeg defines
a static function inet_aton itself.

The program source to test for inet_aton runs into a compile error
because the missing declaration of inet_aton that is treated as
an error by at least GCC with the given CFLAGS. Fix that by
defining _BSD_SOURCE before inclusion of inet/arpa.h .

The headers here indicate that _BSD_SOURCE is deprecated. So fixing
this in configure and/or at least in the MPlayer source files,
that actually use inet_aton, would be a good thing. Those would be:

  stream/librtsp/rtsp_rtp.c
  stream/tcp.c
  stream/udp.c
  udp_sync.c

Maybe like it was done in FFmpeg commit af1818276ef271af98 .

Modified:
   trunk/configure

Modified: trunk/configure
==============================================================================
--- trunk/configure	Mon Jan  8 13:44:45 2018	(r38015)
+++ trunk/configure	Tue Jan  9 23:47:13 2018	(r38016)
@@ -3718,7 +3718,7 @@ echocheck "inet_aton()"
 def_inet_aton='#define HAVE_INET_ATON 0'
 inet_aton=no
 for ld_tmp in "$ld_sock" "$ld_sock -lresolv" ; do
-  statement_check arpa/inet.h 'inet_aton(0, 0)' $ld_tmp && inet_aton=yes && break
+  define_statement_check _BSD_SOURCE arpa/inet.h 'inet_aton(0, 0)' $ld_tmp && inet_aton=yes && break
 done
 if test $inet_aton = yes ; then
   test "$ld_tmp" && res_comment="using $ld_tmp"


More information about the MPlayer-cvslog mailing list