for the area behind the slide, that is

Change-Id: Ie020f43a81e5eaa257c20d0a7b6feffce1614205
Reviewed-on: https://gerrit.libreoffice.org/16466Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Philippe Jung <phil.jung@free.fr>
(cherry picked from commit 1c481fe6)

Change-Id: Ia8c080ef50e1ddbfce17b5c5d357a240edea46f2
(cherry picked from commit b1a9aba3)

Impress has tools Color that is wrapped in this primitive, then later
drawinglayer::processor2d::VclPixelProcessor2D::processBasePrimitive2D()
converts it back to tools Color. Problem is that the primitive uses basegfx
BColor, so the alpha channel is lost.

Add member and API to survive this roundtrip.

Change-Id: I940e60f6e352022306abac3223636d19dd859355
(cherry picked from commit 9824134a)

Which happens e.g. during rendering tiles:
OutputDevice::BlendBitmapWithAlpha() produces an RGB alpha
(255,255,255), then tries to set a pixel in the 8bit alpha channel using
it, so when BitmapReadAccess::SetPixelFor_8BIT_PAL() tries to get the
color index, we try to get it from a color that is not indexed.

Let's assume that when the color is not indexed, it's always gray, so it
doesn't matter what color we pick for the alpha mask needs.

Change-Id: I325c1d70514fd176fdc9cc39683b444447adf07f
(cherry picked from commit 4caaa09d)

Change-Id: Ie018a878eb7d7ef14a80a6b86020c114ff14da88
(cherry picked from commit 4fe010cc)

Change-Id: I65c31995c02a644aa436aecd065255fab38045e4
(cherry picked from commit 1d3b6133)

For tiled rendering this has the effect of not invalidating the
previous slide when switching slides. After switching slides, an
invalidation occurs caused by showing the new slide, and calling
'GetPart' before 'SwitchPage' finishes returned the old part number.

Change-Id: I1cafd0e51cd39be3a80d0559ae3051238b8df744
Reviewed-on: https://gerrit.libreoffice.org/17562Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
Tested-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
(cherry picked from commit 45576ea3)

Change-Id: Ie84f41ab8ad032de5a31e4f566ec5b2358adcf37
(cherry picked from commit 13a609e2)

Change-Id: I88c5c341b88a6ea62730e5d3dd1962365c531f70
(cherry picked from commit 8e27c688)

Change-Id: I20acd44f7a81471982ba96ad3894a9124e035c5f
(cherry picked from commit d355207b)

Change-Id: I9fa710288729b904f2f1c5b3c575212ef21c8a79
(cherry picked from commit edb9ef41)

(cherry picked from commit 14a9a7d5)

Conflicts:
	desktop/source/lib/init.cxx

Change-Id: I345a6cab57cc594614e348700bb3c1c9a0673989

Change-Id: I3b1f4e11f2495e5ccb41f85802f243c0190695ee
(cherry picked from commit 9b9f2ad9)

Change-Id: I4e00679547997cfb3dafe603b908f055011a3b30
(cherry picked from commit 32f419fe)

Also:

- let the unit test set the global LOK flag, as sw code now depends on
  that
- in framework, don't return early after emitting the LOK status
  indicator callback, otherwise CppunitTest_sw_tiledrendering shows how
  sw LOK callbacks are missing

Change-Id: I0c4ac12f2ef5118d29afd131676bcb27d5db7746
(cherry picked from commit 1a83f30e)

(cherry picked from commit af29b2b5)

Conflicts:
	libreofficekit/qa/gtktiledviewer/gtktiledviewer.cxx

Change-Id: Ia71e80c521cdc6a8e4df52e51e063a44a0dd53df

(cherry picked from commit 43b2103f)

Conflicts:
	libreofficekit/qa/gtktiledviewer/gtktiledviewer.cxx
	libreofficekit/source/gtk/lokdocview.cxx

Change-Id: Ic48adaf038e8fbcc86a94b5e351d2f963fcfcd16

Change-Id: Ie33f8b2f8d2a65fbdb7f4f3caaeee8f0f046cba5
(cherry picked from commit de244205)

Change-Id: Ib4c1caede94d3eac372a3605c430bbbd353755f3
(cherry picked from commit 5c2811d6)

Change-Id: I9dd2934eaa7b479460c3a2dc7d2d51ce5c767a9f
(cherry picked from commit b20de6c5)

With this, a postKeyEvent() in one view properly results in one
LOK_CALLBACK_INVALIDATE_TILES per view, at least in Writer.

Change-Id: Ia0a9a00ea5a98c38f3d399208b9ef028f3036f79
(cherry picked from commit b8a54763)

Change-Id: Id839dc076824e69fe07386c83bf21fc4c7ce2b8e
(cherry picked from commit b0f04d1b)

Change-Id: Iad0b2ee419327daf478f3ddda2378effe0184067
(cherry picked from commit 17042210)

This is similar to the existing LOK callback, the difference is that the
existing one assumes there is only one SfxViewShell instance at the same
time.

This newer callback is precisely per-view, so model notifications can
invoke all view callbacks, while view notifications can invoke only the
callback of the relevant view.

This is just the framework, all actual client code has to be still
ported over (and then the existing callback can be removed).

(cherry picked from commit c74ccac7)

Conflicts:
	sfx2/source/view/viewimp.hxx

Change-Id: I3d8f27740c69fcf6ffbbdce12db2ea088321493d

Change-Id: I9cf3a4887774657e0e205114688b088fc11c6822
(cherry picked from commit 714257af)

getTextSelection() is still to be done, but has to be moved to the
widget first.

(cherry picked from commit d3446bc8)

Conflicts:
	libreofficekit/source/gtk/lokdocview.cxx

Change-Id: I780a31f1dbce38b2b7398bfdacbbc1c40881229d

Change-Id: I5dd5f0cfb8b34621b18a35be682249e042d9b173
(cherry picked from commit 42844c2a)

(cherry picked from commit c9ebc92a)

Conflicts:
	libreofficekit/source/gtk/lokdocview.cxx

Change-Id: Iafb08cb44a96dbc03d12367d21b4120063bd0222

Just to be consistent, as all other view-related member functions are
there, too.

No real impact, as only the unit test uses this so far, and it always
works with a single document.

Change-Id: I46f1ed8265ab95017986ab45c1b510e961192241
(cherry picked from commit a04b31c9)

Also in SfxLokHelper::setView() set the current view shell directly,
GetFocus() in VCL may be a NOP for hidden windows.

With this, the Writer layout dump shows that two Gtk windows can have
different cursor positions correctly.

(cherry picked from commit f2e55ea1)

Conflicts:
	libreofficekit/source/gtk/lokdocview.cxx

Change-Id: I81890c1d8ad7972f1194db3d5f2e9d8a39fc2f87

Allows getting rid of vcl::ITiledRenderable::getCurrentViewShell(),
which would do the same, just not implemented outside Writer.

Change-Id: Id26ceca560fb9002dc2d5c740c411b9c4a149523
(cherry picked from commit 8cb60944)

Change-Id: Ic3bce8f01d7e048e853c063c4bce1255845c60d0
(cherry picked from commit 46588c42)

(cherry picked from commit 855e1bac)

Conflicts:
	libreofficekit/qa/gtktiledviewer/gtktiledviewer.cxx

Change-Id: Ibbbaaa1c148cfd4578ad37e0f99b62ab885a1a83

With this finally the number of GTK+ windows always match the number
returned by SfxLokHelper::getViews().

(cherry picked from commit b3498e9f)

Conflicts:
	libreofficekit/source/gtk/lokdocview.cxx

Change-Id: Ia45bef7dea86b80cfac00e2ad7c1a16d7f5b507b

GTK+ calls should be made from the main thread.

(cherry picked from commit e6e5c248)

Conflicts:
	libreofficekit/source/gtk/lokdocview.cxx

Change-Id: Idcfa46d427d6e35fc544246a691bafc72f75a74c

(cherry picked from commit c1d72145)

Conflicts:
	libreofficekit/qa/gtktiledviewer/gtktiledviewer.cxx

Change-Id: I9dcb5871c231a49bcd65a1187df6c3c81f92bd55

(cherry picked from commit 8d8a8e1a)

Conflicts:
	libreofficekit/qa/gtktiledviewer/gtktiledviewer.cxx

Change-Id: Iac803f03267859c11f0f9090346b09b90c115b76

(cherry picked from commit 2b69ce33)

Conflicts:
	libreofficekit/qa/gtktiledviewer/gtktiledviewer.cxx

Change-Id: Ib7ccd2192247805205b65e5e8da67bcc826f4d31

(cherry picked from commit 4fd59356)

Conflicts:
	include/LibreOfficeKit/LibreOfficeKitGtk.h
	libreofficekit/source/gtk/lokdocview.cxx

Change-Id: I2ff57c4be11cd3908bc951ebf572a2c02e2c4585

With this, all previously global data is now stored per-window in the
g_aWindows map.

(cherry picked from commit d51e3f78)

Conflicts:
	libreofficekit/qa/gtktiledviewer/gtktiledviewer.cxx

Change-Id: I45b4449ee7d516106ea0b039d1af97db49edf759