Proposed 1.0 semantics specification

[Casey Dahlin]

On 06/17/2009 06:06 PM, Matthias Urlichs wrote:
> Hi,
>
> Casey Dahlin:
>    
>> The way its worded in my example the check would happen 1 minute after
>> startup, and never again.
>>      
>
> Alternately you could trigger it on "system_normal is in state 'stopped'
> for more than 1 minute" (which can't be specified yet IIRC, but that
> should be fixable), on the theory that a reboot or restart shouldn't
> take that long (you don't want a barrage of state emails every time you
> restart things when upgrading, for instance).
>
>    

True. There's a lot of sophistication that could be added. Its something 
to look into as the system takes shape.

>> An alternative would be a "settled" event,
>> which would occur every time the state machine stops transitioning for a
>> moment.
>>      
>
> "Send an event when there are no events" sounds like an internal
> inconsistency waiting to happen. :-/
>
>    
There's lots of right ways to do it. It depends on how we implement 
certain things though.

--CJD
-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.ubuntu.com/archives/upstart-devel/attachments/20090617/50b7916c/attachment.htm