First of all we get the error that Ludo sent in the previous message if we leave nscd enabled.
Now the problem is in nscd/connection.c at struct database_dyn dbs[lastdb]. The macro PTHREAD_RWLOCK_WRITER_NONRECURSIVE_INITIALIZER_NP is used, but it's nowhere to be found in libpthread as one would expect. Normally there should be a definition in libpthread/include/pthread/pthread.h as is the case of PTHREAD_MUTEX_INITIALIZER. Should PTHREAD_RWLOCK_WRITER_NONRECURSIVE_INITIALIZER_NP be replaced with __PTHREAD_RWLOCK_INITIALIZER which is already declared? If I do this, there is no error. Second question, can nscd be enabled with hurd in the first place? Because even when I replace the macro, I get the message > gai.c:41:22: fatal error: check_pf.c: No such file or directory > #include <check_pf.c> > ^ > compilation terminated. > ../o-iterator.mk:9: recipe for target > '/tmp/nix-build-glibc-hurd-cross-i686-pc-gnu-2.18.drv-5/build/nscd/gai.o' > failed check_pf.c can be found in ./sysdeps/unix/sysv/linux/check_pf.c and ./inet/check_pf.c so I shouldn't be getting that error message. And lastly , Samuel please push the patch Ludovic sent here https://lists.gnu.org/archive/html/bug-hurd/2014-06/msg00108.html about nscd in the tschwinge/Roger_Whittaker branch. It removes conflicting ‘thread_info_t’ declaration between nscd and Mach. Thank you for your time. Manolis
