Darren Kenny wrote:
> Kyle McDonald wrote:
>
>> David.Comay at sun.com wrote:
>>
>>> Gary,
>>>
>>>
>>>
>>>> And we should question every addition to /etc/*. New /etc/* files
>>>> need justification as specified in the ARC SMF policy.
>>>>
>>>>
>>> I can certain appreciate the policy with respect to the bits of
>>> software that are represent of core part of Solaris but I don't agree
>>> it should apply generally with the large number of FOSS components that
>>> are being merely aggregated from upstream communities. It's important
>>> that the familiarity of these components across multiple OS platforms
>>> be maintained in *most* cases. There are, of course, exceptions here -
>>> putting executables in /etc seems wrong in all cases - but for the most
>>> part, I don't see an issue with new components delivering files under
>>> /etc unless they're interfering with other components (/etc/security
>>> might fall into this case but that didn't apply with this particular
>>> case.)
>>>
>>>
>> I'm not sure I'd agree.
>>
>> While for things in home directories might need this to enable homedirs
>> to be shared between more than one platform, I don't think the same
>> applies to things local to the system.
>>
>> Personally, I'd like to see these things put through the Sun/Solaris
>> engineering thought process. Too often in the OSS world things get done
>> without much consideration to how they fit into the larger scheme of
>> things. I'd hate to think we just pulling this stuff in without giving
>> it the same consideration we give to things that originate in Solaris.
>>
>> -Kyle
>>
>>
>
> I feel that we're digressing here, and moving away from the main topic, but I
> just can't help providing my *personal opinion* here:
>
> I personally feel that this is where there is the need for
> unstable/experimental package repositories (similar to blastwave, but where
> it's located on opensolaris.org) so that such OSS s/w can be available
> without being in the core of OpenSolaris/Solaris - sometimes we just want to
> make something available to users (developers) rather than have full
> integration.
>
> Over time, if there is sufficient interest (through measuring installs of the
> packages or requests from customers) in bringing this into Solaris then we can
> pursue the effort of stabilising things and doing the full Solaris/Sun
> engineering thought process.
>
>
I can definitely aggree that this is an ideal solution. I can't think of
anything else that might have the same 'best of both worlds' comprimise.
IPS seems to hold more an more promise everyday. I can't wait to see it
in main stream Solaris.
> Unfortunately we tend to pick up on things from the FOSS communities late in
> their development cycle when it's late to influence sufficiently to be able to
> have good integration into Solaris.
>
>
I know its somewhat of a chicken and egg problem. It's hard to know what
FOSS will be in demeand till it's more well known and in use, but by
then it's harder to make changes.
Still in the interest of improving things, I'd like to see us not
consider things the FOSS world does as untouchable. Especially when it
comes to enabling muilti architecture, multi OS home directories and
installs.
It's weird. Pre linux GNU software used to have some mechanism for
combining platform indepentdant files, and separating platfor specific
files (I'm thinking of emacs here mainly, but perl might have also had
something for this.) It seesm like since the linux explosion, that
consideration has gone out the window.
But as you said. We digress too much. :)
-Kyle
> Thanks,
>
> Darren.
>
