pgrep -f init doesn't return anything on FreeBSD, but it does on Linux: $ uname -a FreeBSD freebsd-10-x64.localdomain 10.0-STABLE FreeBSD 10.0-STABLE #2 r270252+223d76d(stable/10): Wed Aug 20 20:55:18 PDT 2014 root@freebsd-10-x64.localdomain:/usr/obj/usr/src/sys/GENERIC-WITHOUT-WITNESS amd64 $ pgrep -f init $ pgrep -af init 75489 1 $ This is present on a custom version of 7.x, 10, and 11.
This was deliberately changed in SVN r192242, which excludes the pkill process itself and its ancestors, unless -a is given. Do you think a special case should be added here, or r192242 should be reverted? Note that the main sshd process may somewhat similarly be an ancestor (or not, even if the user is logged in via ssh). Why are you using pgrep on init anyway, given that its PID is always 1?
(In reply to Jilles Tjoelker from comment #1) > This was deliberately changed in SVN r192242, which excludes the pkill > process itself and its ancestors, unless -a is given. Do you think a special > case should be added here, or r192242 should be reverted? Note that the main > sshd process may somewhat similarly be an ancestor (or not, even if the user > is logged in via ssh). I agree with bryan about the change for pkill, but it would be nice if pgrep defaulted to the old behavior :/. > Why are you using pgrep on init anyway, given that its PID is always 1? Just for testing purposes because I was trying to prove that I was sane based on the way I was using the tool :).
(In reply to Garrett Cooper from comment #2) > (In reply to Jilles Tjoelker from comment #1) > > This was deliberately changed in SVN r192242, which excludes the pkill > > process itself and its ancestors, unless -a is given. Do you think a special > > case should be added here, or r192242 should be reverted? Note that the main > > sshd process may somewhat similarly be an ancestor (or not, even if the user > > is logged in via ssh). > > I agree with bryan about the change for pkill, but it would be nice if pgrep > defaulted to the old behavior :/. *brian
Is this still considered a bug? The behavior is still the same in 14-CURRENT.