Init : someone could comment this ?

[Casey Dahlin]

Bill Nottingham wrote:
> Nicolas Mailhot (nicolas.mailhot at laposte.net) said: 
>   
>> Le Mar 8 janvier 2008 23:45, Ian Burrell a écrit :
>>
>>     
>>> I think it is important that we preserve the ability to use sysv-style
>>> init scripts.  They are part of LSB.
>>>       
>> Sort-of. The LSB description is ambiguous, as evidenced by the many
>> questions posted when we decided to migrate our existing pre-LSB
>> scripts to the LSB variant. And I'm far from sure that other
>> distributions answered those questions the same way.
>>     
>
> Well, they're not ambiguous in that they specify a script or command
> that accepts certain stop/start/whatever arguments.
>
> To put it a different way, the chances of a solution that doesn't at
> least support SysV scripts in some compatibility way going into RHEL
> is pretty much zero, at least for the first RHEL release that it would
> exist in. And (Red Hat hat on), Fedora should generally not try and break
> its downstream releases too much. Doesn't mean that the system itself
> should be using this mode for any of its scripts, though.
>
> Bill
>
>   
I'm not sure I see the point of this discussion. There hasn't been a 
single solution proposed to this problem which necessarily breaks sysV 
compatibility at runtime. Even a largely python-centric base of init 
scripts could clearly offer the ability to run shell scripts as a 
backward compatibility feature.

Achieving a certain legacy level of support for the old sysV 
packages/scripts is trivial under any model. We need to decide on 
preferred methods. How do we want things to work? Adding in support for 
them working in other ways, particularly in the sysV way, is trivial.

--CJD