Bart Smaalders wrote:
> Jordan Brown wrote:
>> A developer who is considering moving an application to IPS and who 
>> encounters such a situation has three realistic options today:
>>
>> 1)  Hack up a replacement for scripting.
>> 2)  Leave trash all over the system on uninstall.
>> 3)  Stop considering moving to IPS.
> 
> We're still developing and porting Nevada to use IPS.  If we compromise
> the design today to facilitate your early use, we've got those hacks
> around forever.
> 
> Let's try and do this correctly, please.

See, the key disconnect here is that you think that it's *possible* to 
identify and fix all of the plug-in architectures that are not 
packaging-friendly.  I don't think it's possible.  I think that new ones 
are created every day, even inside Sun, and that there's no way that the 
ones outside Sun are going to get fixed in any predictable and reliable 
way.  Even for the existing ones inside Sun, getting resources allocated 
to do the work is not easy, and there are plenty of cases where I would 
call the probability zero.  Heck, even *identifying* all of them is tough.

I don't think requiring zero scripting is reasonable.  The goal should 
be to get as close to zero as possible, and to keep what remains under 
as tight control as possible.  Eliminating scripting against non-running 
systems is an enormous first step.  For those hopefully few cases where 
scripting is necessary, I think you can either provide a couple of 
standard hooks (post install, post remove) or standard patterns for 
using the existing mechanisms, or you can just sit back and watch what 
hacks people come up with to solve their problems.
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss

Reply via email to