[Nagiosplug-devel] check_udp

Andreas Ericsson ae at op5.se
Fri Mar 24 02:46:10 CET 2006


Ton Voon wrote:
> 
> On 23 Mar 2006, at 23:29, sean finney wrote:
> 
>> On Thu, Mar 23, 2006 at 03:46:01PM -0500, sean finney wrote:
>>
>>> or something might trigger a response, but i think it would be
>>> better (cleaner and more appropriate) to just have check_udp
>>> complain and die if it has nothing to send/receive.
>>
>>
>> attached is a patch to check_tcp that does what i just described.
>> i haven't comitted it only because it's not clear that there
>> was an agreement on the matter.
> 
> 
> The patch looks good, and makes more sense to me now. Out of  interest, 
> are there any udp services that would give a response on a  connect 
> without the client sending some data?
> 

Not that I know of.

> 
> I'm trying to commit my changes but SF cvs seems to be out at the  moment.
> 

Migrate to git and I'll host it for you at ghost.op5.se (no, I won't 
keep nagging. ;) )

>> but in any case i think we can safely discard check_udp.c, right?
> 
> 
> I'm keen on dropping it, but I've also just realised that there is a  
> check_udp2 which links to check_tcp. Should we be cautious or or drop  
> check_udp.c, deleting check_udp2 and putting a suitable statement  in 
> ./CHANGES?
> 

Keep linking check_udp2 to check_tcp, but mark it deprecated and 
obsolete. Remove it in the release after next. Larger projects (glibc, 
gcc, ...) have four states (deprecated, obsolescent, obsolete, removed) 
with one minor release for each cycle and a short-circuit at major 
releases (I think a major release counts as two cycles for glibc). I 
don't think we need to be that cautious.

-- 
Andreas Ericsson                   andreas.ericsson at op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231




More information about the Devel mailing list