1N/A chdir 't' if -d 't'; 1N/A # Check to make sure we're really using NIS. 1N/A # If there's no group line, assume it default to compat. 1N/A print "# Doesn't look like you're using NIS in ". 1N/A# By now the GR filehandle should be open and full of juicy group entries. 1N/A# Go through at most this many groups. 1N/A# (note that the first entry has been read away by now) 1N/A # LIMIT -1 so that groups with no users don't fall off 1N/A # In principle we could whine if @s != 4 but do we know enough 1N/A # of group file formats everywhere? 1N/A # Protect against one-to-many and many-to-one mappings. 1N/A # NOTE: group names *CAN* contain whitespace. 1N/A # what about different orders of members? 1N/A# Do not compare passwords: think shadow passwords. 1N/A# Not that group passwords are used much but better not assume anything. 1N/A# The failure of op/grent test is not necessarily serious. 1N/A# It may fail due to local group administration conventions. 1N/A# If you are for example using both NIS and local groups, 1N/A# test failure is possible. Any distributed group scheme 1N/A# can cause such failures. 1N/A# What the grent test is doing is that it compares the $max first 1N/A# with the results of getgrgid() and getgrnam() call. If it finds no 1N/A# matches at all, it suspects something is wrong. 1N/A print "#\t (not necessarily serious: run t/op/grent.t by itself)\n"; 1N/A# Test both the scalar and list contexts.