[Nagiosplug-devel] runcmd, common.h, and popen

[Andreas Ericsson]

sean finney wrote:
> hey andreas,
> 
> i've applied your patches after a bit of massaging for things that have
> changed in cvs since the submission, and have a new diff file of all
> this together.
> 

Sorry, but I run off my own CVS currently. The patches grew burdensome. 
I haven't touched anything since I implemented the new API though, so 
it's most likely identical code insofar as that goes.

It will however make patch submission quite a pain (indentation is a bit 
different and there are other fixes as well). I'll think of something 
though.

> however, some of the plugins seem to no longer work after using the
> np_runcmd framework.  this is what i've tested:
> 
> successful
> 
> check_dig
> check_dns
> check_load
> check_nagios
> check_ping
> check_procs
> check_swap
> 
> unsuccessful
> 
> check_by_ssh (segfaulted if i gave a remote cmd that didn't exist)

I'll look into it. Did you happen to have a backtrace of the core-dump?

> check_fping (always issues a warning)

This happens with urlize too. I'm not sure why, but I'll have a dig at it.

> check_snmp (doesn't seem to return data at all)

This I have fixed (more or less).

> check_users (doesn't return the correct number of users)
> 

What system was this on, and how does it differ (+-1 or wildly incoherent)?

> untested
> 
> check_game
> check_hpjd
> 
> attached is a diff of your combined patches against our current
> cvs tree, if you or anyone else could help look into these problem
> or untested plugins, i'd be appreciative.
> 

Will do.

-- 
Andreas Ericsson                   andreas.ericsson at op5.se
OP5 AB                             www.op5.se
Lead Developer