core: don't log job status message in case job was effectively NOP (#3199)
authorMichal Sekletar <msekletar@users.noreply.github.com>
Mon, 16 May 2016 15:24:51 +0000 (17:24 +0200)
committerZbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Mon, 16 May 2016 15:24:51 +0000 (11:24 -0400)
commit833f92ad39beca0e954e91e5764ffc83f8d0c1cf
tree2228b1ca68dd68e3cbf939465d6d49a4eb075f7a
parent306578e51853f0a18597b1d976ab402ff7a91a95
core: don't log job status message in case job was effectively NOP (#3199)

We currently generate log message about unit being started even when
unit was started already and job didn't do anything. This is because job
was requested explicitly and hence became anchor job of the transaction
thus we could not eliminate it. That is fine but, let's not pollute
journal with useless log messages.

$ systemctl start systemd-resolved
$ systemctl start systemd-resolved
$ systemctl start systemd-resolved

Current state:
$ journalctl -u systemd-resolved | grep Started

May 05 15:31:42 rawhide systemd[1]: Started Network Name Resolution.
May 05 15:31:59 rawhide systemd[1]: Started Network Name Resolution.
May 05 15:32:01 rawhide systemd[1]: Started Network Name Resolution.

After patch applied:
$ journalctl -u systemd-resolved | grep Started

May 05 16:42:12 rawhide systemd[1]: Started Network Name Resolution.

Fixes #1723
src/core/dbus-job.c
src/core/job.c
src/core/job.h
src/core/manager.c
src/core/transaction.c
src/core/unit.c