Petr Salinger wrote: > found 497033 3.1~rc3-2 > thanks > > Hi, > > the builds fails with > > gcc -DHAVE_CONFIG_H -I. -I../.. -I../../libgst -I../../lib-src -Wall -g > -O2 -Wall -Wdeclaration-after-statement -Wno-format -Wpointer-arith > -Wno-pointer-sign -Wwrite-strings -Wno-strict-aliasing -Wno-switch > -fno-gcse -fstrict-aliasing -MT sockets.lo -MD -MP -MF .deps/sockets.Tpo > -c sockets.c -fPIC -DPIC -o .libs/sockets.o > sockets.c: In function 'fix_sockaddr': > sockets.c:125: error: size of array 'verify' is negative > sockets.c:125: warning: unused variable 'verify' > make[4]: *** [sockets.lo] Error 1 > > See > http://buildd.debian-ports.org/build.php?&pkg=gnu-smalltalk > > It looks like yet another cherry-picking is needed, > or current post rc3 upstream snapshot (i.e. 3.1~rc3+20080909-1).
No, the cherry-picking was fine. But you stumbled into exactly a safety net that I had added to check that the correct codepath was picked on BSD systems. It's strange that you have checking for struct sockaddr.sa_len... yes but for some reason it executes code under #ifndef HAVE_STRUCT_SOCKADDR_SA_LEN ... #endif Can you check the config.h file to see whether HAVE_STRUCT_SOCKADDR_SA_LEN was defined and, if not, investigate why? In my Darwin system I have #define HAVE_STRUCT_SOCKADDR_SA_LEN 1 in config.h. Paolo -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]