Re: statement changes suggestions

[Martin Pala]

Jan-Henrik Haukeland wrote:

> Martin Pala <address@hidden> writes:
>
>  
>
> > What about to move FAILED behind the target specification, as it is
> > in other tests 
> >    
>
> When I implemented this now, I found out that it somewhat looks better
> to use "failed" first because the port statement may be a very long
> statement, so using "failed" at the end may make it disappear. 
>
> Besides using "failed" first also clearly indicate that the if
> statement is a binary statement, that is; either it failed or it
> worked.
>
> Anyway, FAILED is a noise keyword so you may place it anywhere you
> like :-) But in the documentation I use this style based on the above
> argumentation.
>
>
> BTW. Please, Martin, don't let me stop you from check in your changes
> to monit.pod :-) I have checked in the small part I have added now.
> I'm mostly thinking philosophically about the monit.pod now and I will
> start to work with it and try to improve stuff when we are almost
> ready with the 4.0 release. That is when all the cards are on the
> table and we only have testing left.
>
>  
OK :) I'm very busy this week (we are starting new system and prepare 
migration) - i'll prepare documentation tomorrow evening i think.

I though about permission test modification related to checksum test 
modification:

Present syntax:

PERM(ISSION) octal_mode

New Syntax:

[IF] PERM(ISSION) octal_mode [FAILED] [[THEN]  action]

So you can use shortcut, for example: 'perm 755', which will behave same 
way like checksum (i.e. send alert) or you can specify long form in the 
case that you need custom action (such as stop, execute, etc.)

Similar way will be uid and gid test implemented.

Because of mentioned project, i'm not able to guarantee the progress - i 
can work on monit in very small windows this week + weekend.

Martin