Re: timestamp monitoring + code simplification patch

[Jan-Henrik Haukeland]

Martin Pala <address@hidden> writes:

> Present "timestamp" statement is presently only optional extension of
> current "check process_name ..." stuff. I think it should stay as it
> is, while it is process object property - it indigates in adition to
> protocol tests another sort of failure, that is regarded to the given
> process. If error is detected, it will use common process property
> (such as mailinglist, start/stop methods, etc.)

Okay, I can see the rationale for this, no problem :) 

> It could be useful to provide this (and other feautures as Jan
> described bellow, such as size, deleted, etc.) as standalone tests,
> so it will allow to check just that file, directory or device and
> etc. (in the case it is not regarded to any process).

It could be interesting, yes. 

> I'm +1 for (Jan's) described syntax. 

Thanks! 

> We can add new major syntax to allow watch another
> standalone/independent objects. 

Exactly, and building future checks on the CHECK statement will look
more consistent, IMHO.

> If there is someone, who can go for it, i'll be happy, or i can
> start on it next week :)

Unfortunately I'm swamped with other work right now, but maybe you can
break a deal with Christian and/or Rory?

-- 
Jan-Henrik Haukeland