Open GoogleCodeExporter opened 8 years ago
Sorry for double post, now i successfully hook into and get the xml,
and the next step is about the MENU_EVENT,
its format is simple path string (like 123:/xx/xxx/xxxx), not xml ?
And about keyboard shortcut handle,
The shortcut seems always working(handled by the active window), so i didn't
need to
register global shortcut according to the XML's accel field ?
Btw, when client change the menu context, are you sure there's now performance
problem,
i see pixmap marshal, and whole menu need to be modified, not the changed menu
entry ...
Original comment by Zealot....@gmail.com
on 24 Mar 2009 at 1:34
>its format is simple path string (like 123:/xx/xxx/xxxx), not xml ?
Yes. Something like an XPath?
>
>And about keyboard shortcut handle,
>The shortcut seems always working(handled by the active window), so i didn't
need to
>register global shortcut according to the XML's accel field ?
Supposingly so.
>Btw, when client change the menu context, are you sure there's now performance
>problem,
>i see pixmap marshal, and whole menu need to be modified, not the changed menu
entry ...
On GTK there was no performance issue. The only issue we have is issue 365.
The number of xml introspections can be reduced by adding a timer or so.
Original comment by rainwood...@gmail.com
on 24 Mar 2009 at 6:48
I've implemented the specification basically (exclude pixmap demarshal, keyboard
shortcut),
take a look at:
http://qiacat.blogspot.com/2009/03/gnome-global-menu-in-kde4.html
Original comment by Zealot....@gmail.com
on 25 Mar 2009 at 6:14
@comment1:
I bet neither does Thomas have heard about your GSOC project.
As your hardwork has already accomplished gtk->KDE, the v2 spec will
concentrate on
QT->GNOME.
@comment3:
Good work! When do you plan to publish the code repository?
Original comment by rainwood...@gmail.com
on 25 Mar 2009 at 3:00
Find a way to make things happen style independently,
now it's working based on Thomas's patch in his QStylePlugin,
code is simple,PropertyNotify handled ...
the code is there.
browse:
http://websvn.kde.org/trunk/playground/base/plasma/globalmenu/
check out:
svn co svn://anonsvn.kde.org/home/kde/trunk/playground/base/plasma/globalmenu/
Original comment by Zealot....@gmail.com
on 25 Mar 2009 at 11:42
I didn't understand how you construct a path from the id. But if it works then
it works.
Original comment by rainwood...@gmail.com
on 25 Mar 2009 at 11:57
the only ways to get this style "independent" are:
- patch Qt4 and make the Trolls accept the patch
- use KStyle (and thus all styles inheriting it can make use of it)
- link out the patching code from bespin into a mini lib (it's designed to be
encapsulated) and so allow other styles to link it (styles are the only plugins
loaded
by all Qt4 apps for sure)
- just extend KDEs kmenubar (what i personally don't like, as it would not
allow plain
Qt apps to make use of this)
i assume the fastest solution would be the minilib (the usage in other style is
really
simple and it allows us to play on the private backend without having Qt/KDE or
using
styles to care about), while the best one was to have integrated Qt support
(but the
spec should be entirely finished then)
so i'll just do, and promote it on kde-look - then we'll see what's gonna happen
Original comment by thomas.l...@gmail.com
on 26 Mar 2009 at 12:06
@comment 3
recursively parse the xml, pass the parent's id as an argument.
@comment 7
I personally prefer make KStyle linked with that minilib.
Seems above methods except patch Qt can't make Qt programs which use QGtkStyle
have
global menu ??
Original comment by Zealot....@gmail.com
on 26 Mar 2009 at 4:15
Seems above methods except patch Qt can't make Qt programs which use QGtkStyle
have
global menu ??
yes, and this would of course be a little mean, because the QGtkStyle is likely
the
one used on gnome desktops...
the good news is however that i /may/ know another way that will
- not require changes to Qt or KDE (ever)
- not require changes to any existing style plugin (ever)
- work regardless of the style you use
- betray my bespin pet... hey wait - that's not a good one ;-)
it remains however some kind of hack -> Qt should have support for this as soon
as we
might have an official fdo protocol
i'll try this first and push it to playground - iff it works.
(if not, i didn't make a fool out of myself by posting idiotic ideas around)
Original comment by thomas.l...@gmail.com
on 26 Mar 2009 at 12:24
basically works ;-)
there may still turn out problems with specific apps (that we'll have to
workaround
then) but the general approach is ok
i'll now try to drag in the tab transition code as well, maybe the amarok hacks
and
make it transparent (if possible, not sure on this) so the user doesn't have to
do
anything but installing.
btw (OT), i got /one/ mail from some "rob sheepmaker" because of GSoC.
i told him to ask around here, as i though the project wasn't accepted (i know
/nothing/ about GSoC beyond it exists and every year about the same time a lot
of guys
with more or less half baked ideas popup and pollute mailing lists ;-)
so did he call? is this in GSoC? is it one of you?
Original comment by thomas.l...@gmail.com
on 26 Mar 2009 at 3:10
GSoc rejected us. We are now providing ideas into accepted upstream
organizations.
Refer to Issue 331.
I am not sure what 'the tab transition code' means. To summarize you've found a
way
to write a QT-plugin for global menu clients, right? That will be a very good
news.
Original comment by rainwood...@gmail.com
on 26 Mar 2009 at 3:31
bespin allows to crossfade between tabs (+some other animated transitions)
there're some more hacks (e.g. for amarok) that have been demanded to "work in
other
styles as well, please" that i could add to this plugin (being optional, of
course)
and yes, we can now inject this into any Qt (see: Qt, like "cute") app w/o
further
assistance from any 3rd party side
Original comment by thomas.l...@gmail.com
on 26 Mar 2009 at 3:39
Issue 381 has been merged into this issue.
Original comment by rainwood...@gmail.com
on 28 Mar 2009 at 11:13
NOTABUG
Original comment by rainwood...@gmail.com
on 28 Mar 2009 at 11:13
Tag this on 0.8 to match up with the roadmap.
@thomas. We gonna hurry up on the v2 spec. I'll be able to concentrate on the
topic
in late April.
Original comment by rainwood...@gmail.com
on 29 Mar 2009 at 5:10
Original comment by pierre.s...@gmail.com
on 13 Apr 2009 at 4:03
well, after errr... some "lag" i implemented support in the XBar plasmoid to
cover the timeframe before canonical will push dbusmenu as std...
so actually this could be closed
(oh, and sorry for the silence from my side)
Original comment by thomas.l...@gmail.com
on 27 Aug 2010 at 5:26
Original issue reported on code.google.com by
Zealot....@gmail.com
on 23 Mar 2009 at 2:33