From the GUI standpoint, we currently list the disabled Publishers in
situ which makes a lot more sense to the user. We do not gray them out
bu that's a minor tweak we can do to emphasize their offline nature.
When they are re-enabled the order should be unaffected, as is the
current behavior.
The ordering is important as well for the search order, so when we are
searching across all Publishers users will see results coming back in
the order reflected by the Publisher priority list, though disabled
publishers will not be searched. Disabling stickiness and re-enabling it
when the publisher is re-enabled as Danek suggested also makes sense to
me. Though perhaps this is something that could be overridden if you
want to freeze a package and its dependencies from a specific Publisher
that might be going away.
JR
Bart Smaalders wrote:
Dan Price wrote:
On Mon 05 Oct 2009 at 03:28PM, Bart Smaalders wrote:
I voted for always adding them at the end, as it is simple
and non-surprising.
I think if you think of it in terms of a GUI-- which would
likely render disabled publishers as "greyed out"-- then this policy is
more surprising than when seen from the command line perspective.
Another way to think about the disabled pubs is that perhaps we should
render them when listing publishers from the CLI (by default), rather
than having them vanish from the CLI's output (by default, yes I know
there is -a).
-dp
Ok; if we change the behavior to always show disabled publishers,
then restoring disable publishers in their place in the search
order works for me.
Should disabled publishers be made non-sticky, or should we
just leave them alone? I'm now thinking the latter is better
if we're going for least surprise.
- Bart
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
