# New Ticket Created by Nicholas Clark # Please include the string: [perl #128831] # in the subject line of all future correspondence about this issue. # <URL: https://rt.perl.org/Ticket/Display.html?id=128831 >
$ ./perl6-m -Ilib t/spec/S16-io/eof.t 1..2 ok 1 - Regular file EOF was reached Use of Nil in string context in block at t/spec/S16-io/eof.t line 26 Failed to open "/proc/1/comm": in block at t/spec/S16-io/eof.t line 26 in block <unit> at t/spec/S16-io/eof.t line 24 # Looks like you planned 2 tests, but ran 1 It's assuming both that process 1 exists, and that /proc/1 for it contains a file comm. (I *think* that the assumption about process 1 is bad. Certainly on Solaris with Zones, the init process may not have pid 1. Perl 5 had a failing regression test until we removed the assumption that process 1 exists. I don't know if every Linux virtualisation system ensures that PID 1 exists. It would be safer to test the /proc entry for the current process ID) On this CentOS system there is a process 1, and its directory in /proc but it doesn't have a comm: $ ls /proc/1/ ls: cannot read symbolic link /proc/1/cwd: Permission denied ls: cannot read symbolic link /proc/1/root: Permission denied ls: cannot read symbolic link /proc/1/exe: Permission denied attr cpuset limits net root statm autogroup cwd loginuid numa_maps sched status auxv environ maps oom_adj schedstat syscall cgroup exe mem oom_score sessionid task clear_refs fd mountinfo oom_score_adj smaps wchan cmdline fdinfo mounts pagemap stack coredump_filter io mountstats personality stat I don't know what file would be safer. Maybe mem? Nicholas Clark
