Thanks for tracking this down Joanie, I am using Metacity. I have filed
a separate bug to track this.
12485 Webinstall sub dialogs not getting focus
We are reparenting the dialogs using a call to
gtk.Window.set_transient_for(). Compiz does not seem to be respecting
this. Will look into it.
JR
repository.py:
def __set_modal_and_transient(top_window, parent_window = None):
if parent_window:
top_window.set_transient_for(parent_window)
top_window.set_modal(True)
Joanmarie Diggs wrote:
John, perchance do you use Metacity rather than Compiz? The reason I ask
is that I'm still seeing the unfocused dialog issue and was trying
different things to work out why you were failing to see it. If I switch
to Metacity the problem goes away.
While this is apparently a window manager issue, I think it would be
good if a solution could be found. I'd think this even if Metacity were
going to be the default window manager. But I heard somewhere that
OpenSolaris is switching over to Compiz by default.
Take care.
--joanie
On Tue, 2009-11-03 at 15:48 +0000, jmr wrote:
Padraig - thanks for the comments, in fixing this also noticed a few
Webinstall UI issues that I have cleaned up (#12433):
http://cr.opensolaris.org/~jmr/pm_12000_webinstall_disabled_3Nov_333pm/
12000 Using Webinstall with a disabled publisher is cumbersome
12290 Proceed button disabled after being invoked in WebInstall
12433 Webinstall UI cleanup
JR
Padraig O'Briain wrote:
Some comments on changes to webinstall.py:
Can self.repo_gui be None at line 348?
Check added.
Is update_package_list called in main (GUI) thread or another thread?
If not the GUI thread I am concerned about calling gtk.main_iteration().
Removed.
Padraig
On 11/03/09 08:22, John Rice wrote:
Thanks Joanie and Shawn - so we will leave this for now as is.
Hopefully the performance issue will improve in the future. I will
check on the focus issue before submitting.
Shawn the reason we are doing the disable -> enable -> disable
sequence is that the user has disabled the Publisher for a reason,
but we still want them to be able to install packages from the
Publisher and leave the system in the same state before they started
the p5i install.
This covers the use case of someone periodically installing a
package from a Publisher they use infrequently and have no desire to
see in PM listings or in search results.
JR
Joanmarie Diggs wrote:
On Mon, 2009-11-02 at 15:04 -0600, Shawn Walker wrote:
John Rice wrote:
Good question Joanie, I think this is required to trigger the
catalog updates. We are just operating on a Publisher object when
we set it to disabled, for the update to take effect I believe you
need to trigger the update, but Shawn would know best.
Yes, you need the update for the image to correctly reflect what
packages are available.
Well.... Then, at least hypothetically, couldn't pkg do a mini-update
rather than a full update, given that:
* an update was just done moments prior, and thus the image
presumably correctly reflects what packages are available
* the only thing which has changed since that update is the
installation
of a finite set of known packages
* the publisher in question is being re-disabled
* the user cannot interact with that publisher until it becomes
re-enabled (right?), at which point a full update will occur
Anyhoo.... :-)
Regardless of the above, this new functionality is cool. As it is.
Thanks again John.
--joanie
_______________________________________________
pkg-discuss mailing list
pkg-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
_______________________________________________
pkg-discuss mailing list
pkg-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
