artifacts in e.g. amarok, or dolphin (to be confirmed).

CCBUG: 290879

DBus services might require more than one type of inhibition, and all of them must be released when this occurs.

Signed-off-by: Dario Freddi <drf@kde.org>

CCMAIL:gaf.stephanos@gmail.com
REVIEW:103712

BUG: 291690

BUG:288827

(cherry picked from commit 14ca7a22)

NVIDIA Binary Blob does not report the type of the screen and instead it
always report "default". Basically this patchs prevent the
checkInhibition code to always inhibit when used with the  binary blob.

BUG: 289760

This checkInhibition code is designed to inhibit events ONLY when used
in a laptop and it allows to inhibit suspension and other events that
are triggered when the lid  is closed.

So, to avoid affecting desktop pc's and other form factors we are
checking if a lid is present.

Something more global will have to be add in order to support other form
factors such Active devices.

This check is completely broken in NVIDIA binary blob since they are
announcing that they support up to 1.3 though in reallity the driver
only supports XRandR 1.1.

This may fix though weird behaviour with other crappy drivers like the
FLGRX blob.

TabBox is always declarative view.

BUG: 289945
REVIEW: 103687

Don't listen to repetitive motif hints regarding the window border and prefer rules over them anyway

BUG: 291312
REVIEW: 103681
FIXED-IN: 4.8

Withdraw partial maximization state when resizing partially maximized client but moveResizeMaximizedWindows() isn't true

REVIEW: 103683
BUG: 290990
FIXED-IN: 4.8

The effect did not check whether the ShaderManager is valid causing
a null pointer access when trying to use the shader.

Additionally this change moves the Shader init into the close window
slot as close window referrenced windows without checking whether the
effect would work. If the effect would not work each closed window
would be referrenced without any chance to being unreffed again as this
code is in a block checking whether the effect is valid.

BUG: 291390
FIXED-IN: 4.8.0

This is a condensed version of Martin's patch that fixes a high cpu
usage in KWin and X. It seems to be due to a window being created and
deleted at almost the same time, such that the fade effect never quits.
The bug was reliably reproducable with starting Amarok.

CCBUG: 290025
CCBUG: 288948

Patch by macau

Signed-off-by: Dario Freddi <drf@kde.org>

abrahams, you'll reopen if this does not fix it when you'll update to 4.8, thans for your cooperation
FIXEDIN: 4.8.0
BUG=290368
REVIEWBOARD:103645

call updateStruts when setting the value outside of editing with the control bar (e.g. from scripting)

a little messy as it's a private slot in Applet and so is prone to breakage. something to consider in improving the config handling in libplasma2
also requires 4ab2139 from kdelibs as this contains a crash fix this exposed

Up to now, the ok mark is shown right after starting
the mounting/unmounting procedure.
It should be instead shown after the procedure is done.
I will make sure this happens, but, for now,
do not show the ok mark as it may induce
users to unplug devices before flushing has completed, causing
data loss.

CCBUG: 290965

CCBUG: 290965

(cherry picked from commit ab8a0f3c)

(cherry picked from commit d2183a40)

(cherry picked from commit c4b8c14a)