search for: num_token_sid

...pwards as a memory error, because it is being > called from getpwuid_alloc() which is a talloc variant. the api > doesn't allow us to distinguish either form of error. > Later, there is this code (in libgpo) > new_token = create_local_nt_token(mem_ctx, &object_sid, false, > num_token_sids, token_sids); > ADS_ERROR_HAVE_NO_MEMORY(new_token); > where the failure of create_local_nt_token() is simply assumed to be a > memory failure. This pretty much destroys any finess in lower level > error handling... > > Now, the reason getpwuid was failing was that the nsswitch...

Well, I had a look at the python bindings and the error handling was in shambles. I fixed libgpo/pygpo.c and I now get: gpos = ads.get_gpo_list(creds.get_username()) SystemError: Failed to get machine token for 'DC01$' (CN=DC01,OU=Domain Controllers,DC=rvx,DC=is) (the failing call is status = gp_get_machine_token(self->ads_ptr, frame, dn,

...This is then propagated upwards as a memory error, because it is being called from getpwuid_alloc() which is a talloc variant. the api doesn't allow us to distinguish either form of error. Later, there is this code (in libgpo) new_token = create_local_nt_token(mem_ctx, &object_sid, false, num_token_sids, token_sids); ADS_ERROR_HAVE_NO_MEMORY(new_token); where the failure of create_local_nt_token() is simply assumed to be a memory failure. This pretty much destroys any finess in lower level error handling... Now, the reason getpwuid was failing was that the nsswitch.conf wasn't set up on the...