Shawn Walker wrote:
Brock Pytlik wrote:
Implementation plans:
I think we've thought that user intent should be done by whoever the
first consumer ends up being. Looking at the above proposal though, I
think it's a large enough amount of work that it is reasonable to do
it separately. Further, the sooner we start tracking user intent, the
more we'll be able to do once consumers start appearing.
Within this work, I think there are further divisions that can be
reasonably made:
1) add intent tracking from this moment forward and get the storage
infrastructure in place and work with the install team so that newly
installed systems either have empty user intent or the desired user
intent
2) Add the historical inference of user intent as well as the default
info for packages.
3) Give packagers the ability to control the propagation of user intent
4) Give users the ability to change the intent for a package
5) If for expedience, the decision in step 1 was to have the
installer provide empty user intent, then fix that so appropriate
intent is delivered.
Those are my current plans and thoughts. I'd appreciate any comments
or suggestions.
One last item I failed to mention:
Before this work is done, image versioning need to be implemented.
The changes you've proposed within are essentially a change in image
format and that means that older clients attempting to manipulate an
image created by a newer client won't work quite right or will cause
loss of this additional information.
I don't think I agree that this work should be stalled by image
versioning. Clients without versioning information will plow ahead
regardless and there's nothing we can do about it. Essentially, all this
would effect is that the set of clients which exist post versioning put
back and pre-user-intent would function correctly. If this goes back
first, that set doesn't exist and thus (by fiat) functions correctly.
If others agree that this should stall on that, I'll wait for such a
scheme to integrate, but I don't believe it's necessary to do so. I
think user-intent is a clear win for a large class of users based on
what it will allow us to improve in terms of UI. (I intend to pick up
the group package uninstall once this goes back for examples.) I think
the class of users using multiple versions of pkg to manage is a tiny
set of users. Given that I don't see this introducing a significant
issue for those users, I'm not inclined to hold this work up.
As such, it would be great to see a proposal for an image versioning
scheme and a dependency on that. If it can go back first, that will
help ensure that older clients aren't used to manipulate images
created by newer clients. Obviously we can't do anything about
clients that don't have versioning checks, but it is a step in the
right direction.
And possibly that change could go back to the sustaining gate as well
to ensure there aren't problems going forward?
I don't have feelings on this one way or the other.
Cheers,
Brock
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
