Executing on host: /mnt/gnu/gcc/objdir-test/gcc/gnatmake --GCC=/mnt/gnu/gcc/objdir-test/gcc/xgcc --GNATBIND=/mnt/gnu/gcc/objdir-test/gcc/gnatbind --GNATLINK=/mnt/gnu/gcc/objdir-test/gcc/gnatlink -cargs -B/mnt/gnu/gcc/objdir-test/gcc -largs --GCC=/mnt/gnu/gcc/objdir-test/gcc/xgcc\ -B/mnt/gnu/gcc/objdir-test/gcc\ -margs --RTS=/mnt/gnu/gcc/objdir-test/hppa2.0w-hp-hpux11.11/./libada -q -f /mnt/gnu/gc c/gcc/gcc/testsuite/gnat.dg/socket1.adb -lm -o ./socket1.exe (timeout = 300) Executing on host: /mnt/gnu/gcc/objdir-test/gcc/gnatclean -c -q -n ./socket1 ( timeout = 300) ./socket1.ali ./socket1.oPASS: gnat.dg/socket1.adb (test for excess errors)Setting LD_LIBRARY_PATH to :/mnt/gnu/gcc/objdir/gcc::/mnt/gnu/gcc/objdir/gccraised GNAT.SOCKETS.HOST_ERROR : [0] Unknown errorFAIL: gnat.dg/socket1.adb execution test Revision 176070 was ok.
Can you confirm that this also comes from Martin's patch on the 4.6 branch?
Based on Martin's comment in the other PR, I wonder whether this isn't rather a temporary failure because of a glitch on the host. Can you reproduce it?
On 14-Jul-11, at 4:40 AM, ebotcazou at gcc dot gnu.org wrote: > Can you confirm that this also comes from Martin's patch on the 4.6 > branch? No, it doesn't come from Martin's change. Rather, it appears to come from a change to resolv.conf that I made last weekend. There was a problem with accessing on of NRC's nameservers and I changed the lookup order. It appears going back to previous configuration fixes things. Will recheck. Dave -- John David Anglin dave.anglin@bell.net
Caused by incorrect name server configuration.