KDE Panel freezing when HDMI connected

[Duncan]

Anne Wilson posted on Sun, 11 Sep 2011 15:24:36 +0100 as excerpted:

> On Sunday 11 Sep 2011 Tim Edwards wrote:
>> On Sunday, September 11, 2011 2:22 PM, "Alex Schuster"
>> 
>> <wonko at wonkology.org> wrote:
>> > Tim Edwards writes:

>> > > My theory is it looks like the Pyweather widget is hanging trying 
to
>> > > get data sometimes, and this in turn hangs plasma-desktop. Which, 
if
>> > > true, is IMHO a huge design fault in plasma-desktop.

>> > Yes, that's the problem with plasma-desktoüp, it's single threaded, 
and
>> > if one plasmoid hangs, whole plasma hangs. I often had such trouble 
when
>> > using plasmoids I downloaded fron the net.
>> > 
>> > I also believe it's a huge design fault, but there _are_ indeed 
reasons
>> > for doing this [but] I don't find the link right now

>> I guess there's not much point, the problem solved/design flaw worked
>> around by removing plasmoids that are possibly unstable. It does kind 
of
>> limit the usefulness of the whole plasma system though - anything which
>> might possibly hang should not be used. This particular machine is a
>> netbook so I expect to use it in various places where internet
>> connectivity is bad or non-existent.
>> 
>> It sucks [...] - no wonder there's so many posts on
>> forums about 'KDE freezing' or 'KDE hanging'.
>> 
> I wonder if it depends on the design of the specific plasmoid.  I'm not 
sure 
> whether I've used my netbook for a long enough period without 
connection to 
> guarantee it, but I think it's true that my yawp plasmoid simply shows 
the 
> cached information when it can't connect, so I'm guessing that yawp has 
a 
> time-out and pyweather doesn't.

It does depend on the plasmoid.  See the other subthread, the previous 
thread between Wonko and me discussing it, and the link in the other 
subthread to ASegio's comment discussing it on a blog.

His point is that there's no other way to paint (with any kind of 
performance) on a (presumably) qt canvas, but having all painters in a 
single thread.

My point, Wonko's, the person who wrote that blog, Tim's in this thread, 
etc, is that, if qt-canvas as a technology doesn't allow performant 
robustness, then qt-canvas as a technology was the wrong technology 
choice for an app which has as a primary bullet point how extensible it 
is, by "ordinary" people without expert-level cooperative-multitasking 
skills.  These freezes simply shouldn't be happening in an app where such 
extensibility is touted as a primary feature, and if they are, the wrong 
choices were made SOMEWHERE.

Meanwhile, it should be noted that it's possible to run a plasmoid in a 
separate window and app, if necessary, instead of in plasma itself, thus 
eliminating the problem since all it takes down is that separate window/
app.  It's not as integrated, but for plasmoids that have this sort of 
issues, it DOES allow them to be run.

See the plasmoidviewer and plasma-windowed commands (run in konsole with 
--help to get a short description and parameter help) for two such 
solutions.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman

___________________________________________________
This message is from the kde mailing list.
Account management:  https://mail.kde.org/mailman/listinfo/kde.
Archives: http://lists.kde.org/.
More info: http://www.kde.org/faq.html.