Bug 367819 – net-analyzer/fail2ban-0.8.4-r3: initscript not completely openrc compatible

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Robert Trace 2011-05-17 21:05:03 UTC]

Initscript uses svc_(start|stop).  These functions don't appear to be supported
in openrc.

Reproducible: Always

[Comment 1 Robert Trace 2011-05-17 21:06:30 UTC]

Created attachment 273719 [details, diff]
Remove svc_(start|stop) and make openrc compatible

[Comment 2 Robert Trace 2011-05-17 21:07:49 UTC]

Created attachment 273721 [details, diff]
fail2ban-0.8.4-r3.ebuild add openrc support patch

[Comment 3 Markos Chandras (RETIRED) 2011-05-18 09:01:49 UTC]

So you mean that the restart() function does not properly restart the service? It does nothing? Cause for me it works just fine. But I will have a look anyway

[Comment 4 Robert Trace 2011-05-18 17:52:43 UTC]

(In reply to comment #3)
> So you mean that the restart() function does not properly restart the service?
> It does nothing? Cause for me it works just fine. But I will have a look anyway

Openrc hardcodes the restart() function and ignores anything in the script.  So, yes, a "/etc/init.d/fail2ban restart" will work, but it's not executing the code in the initscript.  Since svc_stop/svc_start no longer exist, there's no way a restart should ever work.  The fact that it's working means that openrc is ignoring it.

Is this wrong/broken?  Functionally, not really, but since the restart() function in the script isn't doing anything (if it ever does, then it _is_ broken, since svc_stop won't work), it should be removed.  This would also clean up a misleading example of how the init system works.

[Comment 5 Markos Chandras (RETIRED) 2011-05-18 18:05:38 UTC]

You are right. I fixed the code but I didn't revbump the ebuild since it didn't affect the functionality of the package. Thanks for reporting