pgrep -o not working properly

[jclulow]

On an OmniOS r151046 system, the -o option does not seem to interact well with at least the -l and -f options. For example, given this process tree:

$ ptree 16800
780    /usr/sbin/sshd
  16797  /usr/sbin/sshd -R
    16799  /usr/sbin/sshd -R
      16800  -bash
        16823  sleep 10000
        16824  sleep 20000

It would seem that pgrep -o works correctly on its own:

$ pgrep -o sleep
16823

And that pgrep -fl and -l work on their own as well:

$ pgrep -l sleep
16823 sleep
16824 sleep
$ pgrep -fl sleep
16823 sleep 10000
16824 sleep 20000

Things get weird when you mix them together:

$ pgrep -flo sleep
16823 sleep 20000

Note that it is the correct pid for -o, but the arguments are for a different process. If you get more specific, it's even weirder:

$ pgrep -flo sleep.10000
16823 sleep 20000
$ pgrep -flo sleep.20000
16824 sleep 20000

And it seems to happen for any process:

$ pgrep -Fl bash
16816 -bash
16800 -bash
$ pgrep -o bash
16800
$ pgrep -Flo bash
16800 sleep 20000

This is running on LTS bits:

$ cat /etc/os-release
NAME="OmniOS"
PRETTY_NAME="OmniOS Community Edition v11 r151046q"
CPE_NAME="cpe:/o:omniosce:omnios:11:151046:17"
ID=omnios
VERSION=r151046q
VERSION_ID=r151046q
BUILD_ID=151046.17.2023.08.24
HOME_URL="https://omnios.org/"
SUPPORT_URL="https://omnios.org/"
BUG_REPORT_URL="https://github.com/omniosorg/omnios-build/issues/new"

[danmcd]

pgrep was updated initially as part of the LX port. There were some additional pulls from SmartOS, and at least in pgrep.c a backout. Perhaps there's some suffering caused by that?

[citrus-it]

pgrep was updated initially as part of the LX port. There were some additional pulls from SmartOS, and at least in pgrep.c a backout. Perhaps there's some suffering caused by that?

It is part of OS-3513 ptools should see more process arguments - SmartOS needs the linked fix too.