Initscripts and LSB compliance
Michal Marciniszyn
mmarcini at redhat.com
Fri Mar 30 14:50:56 UTC 2007
Bill Nottingham wrote:
> Patrice Dumas (pertusus at free.fr) said:
>
>>> Our first step should be to produce guidelines (we have some for RHEL,
>>> but they are not obeyed), then force the developers to obey that. It is
>>> no big deal, but having all scripts behaving correctly and in some sense
>>> the standard way is definitely good think.
>>>
>> I completely agree. Having glanced through the specification there is
>> one point that doesn't seems to be desirable, it is the script naming
>> scheme which seems ugly to me:
>> http://refspecs.freestandards.org/LSB_3.1.0/LSB-Core-generic/LSB-Core-generic/scrptnames.html
>> Although it could be a SHOULD item that upstream is contacted to
>> register to the lanana.
>>
>
> System init scripts are not required to follow the LSB standards. I suspect
> that following them for something like return codes should be fine, but
> renaming them just leads to trouble, and should be avoided.
>
> Bill
>
>
I totally agree, my main point was avoiding %conf in the script (which
is the part of the policy AFAIK) and correct return codes and status
call behavior.
Michal
More information about the fedora-devel-list
mailing list