Good to hear from you Jimmy. Well, if both you and Ismaell thinks that 
initng-scripts is the way to go i guess that's the way it's gonna be as 
you are the main devs. Hopefully those people that seems to be a little 
pissed before will come back when it works smoothly. :D And i can write 
a script that transform .i to .s when the API is finished.

And I sat here thinking before about what was the most important things 
in an initsystem for me and two things stands out:
1. Dependency checking (fast boot).
2. Easy to write initscripts.
(Daemon restarting when failure is also nice but I can live without it.)
So hopefully it will not get too complicated. I mean, .i files are very 
easy to understand, at least compared to just shellbased ones.

Jimmy Wennlund wrote:
> The new system for scripts is really a step forward, that meens that we
> can move logic away from initng, and reduce a lot of code, and at the
> same time make initng more dynamic, with this way init scripts can be
> written in c/bash/tcl/perl or other language.
>
> Maby its possible to make/write an .i -> .s convertor so the move will
> go smoth.
>
> Ismael, keep up the good work.
>
> /Jimmy
>
> lör 2007-03-17 klockan 00:56 -0300 skrev Ismael Luceno:
>   
>> Iens Persson escribió:
>>     
>>> Hello everyone.
>>>
>>> Just wanna ask if we could agree on going forward with initng-scripts or 
>>> if we should stay with initng-ifiles. I'm happy with either one but I 
>>> (and probably the rest of the devs and users) need to know how the heck 
>>> we should do. And about some other things here are my opinions:
>>>
>>>       
>> We will start working on the .s files after the 0.6.10 release :).
>>
>>     
>>> builtin shell: NO, not ever. This new service_file stuff seems to 
>>> support all scripting languages and it's hard to support them all in one 
>>> shell. :P
>>>       
>> This will never happen, but we may develop our own super-fast shell
>> *in a distant future*.
>>
>>     
>>> *inetd functionality: No, not now but may be reconsidered in the future.
>>> cron functionality: Hmm, not such a bad idea. This one could fit pretty 
>>> nicely in Initng.
>>>       
>> Both may be implemented *after* the 0.7.0 release.
>>
>>     
>
>   

-- 
_______________________________________________
Initng mailing list
[email protected]
http://jw.dyndns.org/mailman/listinfo/initng

Reply via email to