Package: git-daemon-run
Version: 1:1.7.4.1-4
Severity: minor
Just stumbled on some confusing output:
# update-service --remove /etc/sv/git-daemon
Service git-daemon removed, the service daemon received the TERM and CONT
signals.
# dpkg-reconfigure git-daemon-run
Service git-daemon added.
fail: git-daemon: runsv not running
# sv stat /etc/sv/git-daemon
run: /etc/sv/git-daemon: (pid 9732) 7s; run: log: (pid 9731) 7s
The same happens on upgrade if the admin has removed
/etc/sv/git-daemon.
* What are the correct semantics? Is postinst right to add back the
service after the admin removed it? (Yes, I think so.)
* Would it make sense to notice this case and give a hint to the user?
For example,
# git-daemon service disabled?
if test -n "$2" &&
! update-service --check /etc/sv/git-daemon
then
echo >&2 'git-daemon-run/upgrade: re-enabling git-daemon
service'
fi
# enable git-daemon service
update-service --add /etc/sv/git-daemon
* Is the output from "sv -v term git-daemon" ('fail: git-daemon:
runsv not running') useful to the user? We are ignoring the exit
status; it could make sense to send output from sv -v term to
/dev/null, too. But if failures are expected, I wonder why
postinst uses -v in the first place.
Hmm. I'm not too comfortable with this story yet.
--
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]
