Apparently we did not have the pInputWin living in the background; but
now we have.  There is no reason to treat it in an 'else' branch, these
checks should be separate either way.

Change-Id: Ibb8d92fb1e2803942460d2063847917d082fcb2e
Reviewed-on: https://gerrit.libreoffice.org/71863Reviewed-by: Aron Budea <aron.budea@collabora.com>
Tested-by: Aron Budea <aron.budea@collabora.com>

This limits duplicate cursor invalidation removal to same
view only, which seems to have been left in error.

We should now remove any old cursor invalidations (for
the same view) and only keep the latest (current) one.

Change-Id: Ie2323d0c5fcf3977576a1bdc098c95351a5753e0
Reviewed-on: https://gerrit.libreoffice.org/71846Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Change-Id: I0db9ec275cc95184ceb2cdbce8ae5343a10582c1
Reviewed-on: https://gerrit.libreoffice.org/71862Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Change-Id: I9b8e060c2c7655565b95004d82bf50ada2ed0d08
Reviewed-on: https://gerrit.libreoffice.org/71845Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Change-Id: I4879d65285e01cf2fd99233d226f4201997b4dd7
Reviewed-on: https://gerrit.libreoffice.org/71844Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Change-Id: I3a3bd1fb6922e435599f604328f558be60594729
Reviewed-on: https://gerrit.libreoffice.org/71843Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Change-Id: I7f8695789e1223df7bbb235fa6081ed91502ab99
Reviewed-on: https://gerrit.libreoffice.org/71838Reviewed-by: Andras Timar <andras.timar@collabora.com>
Tested-by: Andras Timar <andras.timar@collabora.com>

Rather than emitting asynchronously at idle.

Change-Id: I6c72e9fad0b5587941e3a4a4d17d331a0d889942
Reviewed-on: https://gerrit.libreoffice.org/71770Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Setting a custom thread name is a terrible idea for online unit tests.

Change-Id: I1f5843255ab59072f806587ffcf655958380fe8e
Reviewed-on: https://gerrit.libreoffice.org/71761Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

lo_startmain drops the SolarMutex without taking it, interestingly.

Change-Id: I2280e2fa6acb8739f970983032db72a8241d1c4f
Reviewed-on: https://gerrit.libreoffice.org/71742Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Change-Id: I5d8e95e2d4cfc8d94552d31ff36f2a10271e0b46

Fixes CVE-2019-7317.

Change-Id: I3374f5cbd6552e2c1569d63ee680d0c1d9389621
Reviewed-on: https://gerrit.libreoffice.org/71663
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>
(cherry picked from commit e2abbc00)

Change-Id: I8d846bc1f794cf16fd45486fd83ef87f43fd52c7
Reviewed-on: https://gerrit.libreoffice.org/71715Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Change-Id: Iae59e7fa2c412c8991af661d63a4bf1c17bcc6b9
Reviewed-on: https://gerrit.libreoffice.org/71719Reviewed-by: Marco Cecchetti <mrcekets@gmail.com>
Tested-by: Marco Cecchetti <mrcekets@gmail.com>

Change-Id: Id386575029edf5a75f91b6831bbc7fe49741c06e

Change-Id: Id9f8f07bfdb1060f0c3d7ed6f8526a6bfdd84eb4
Reviewed-on: https://gerrit.libreoffice.org/71687
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Reviewed-on: https://gerrit.libreoffice.org/71696Tested-by: Mike Kaganski <mike.kaganski@collabora.com>

During redaction of Calc sheets

Change-Id: If2a4f71c77a687a74b8f823152b76d2a7e2e21da
Reviewed-on: https://gerrit.libreoffice.org/71658Reviewed-by: Muhammet Kara <muhammet.kara@collabora.com>
Reviewed-by: Andras Timar <andras.timar@collabora.com>
Tested-by: Andras Timar <andras.timar@collabora.com>

Use a function to consistently check if gImpl is valid before assigning last
exception message.

V595 The 'gImpl' pointer was utilized before it was verified against nullptr.
     Check lines: 2673, 2683.

Change-Id: I18fc0d632c84f37b05e9d1c3c83689e6eaa6f66c
Reviewed-on: https://gerrit.libreoffice.org/71488
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Reviewed-on: https://gerrit.libreoffice.org/71528Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Tested-by: Michael Meeks <michael.meeks@collabora.com>

Helps building the Online before the appropriate patch gets merged.

Change-Id: I9983d546bb0cb7fb472806622a0a10e3a4d3ab87
Reviewed-on: https://gerrit.libreoffice.org/71664Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Tested-by: Michael Meeks <michael.meeks@collabora.com>

This fixes the gtktiledviewer on some distros/gtk
versions that require an activation function to be
provided.

Change-Id: I95fb3ed4e92e27d9c725ca38b4cb5a4f8af51771
Reviewed-on: https://gerrit.libreoffice.org/71588Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

This re-enables the sidebars for LOK, which were
disabled to avoid overheads.

Also, fixes a regression where editing stopped
when loading a new view because the sidebar window
creation killed the edit view.

Revert "Avoid various desktop GUI elements when used from LibreOfficeKit"
This reverts commit c0b70ec3.

Change-Id: I06b06a990f05c06b1889fa89b698dff6f494e09c
Reviewed-on: https://gerrit.libreoffice.org/71165Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

The reasoning is somewhat complex:

void Window::ImplInvalidateFrameRegion( const vcl::Region* pRegion, InvalidateFlags nFlags )

sets the mnPaintFlags on the mpWindowImpl - and then queues an idle
paint. This paint in LOK mode does ~nothing - since all rendering is
tiled, although amazingly it does emit eg. selection callbacks.

However the paint flag - changes the behavior of Window::Update() to
force a complete window invalidate. This happens, but only rarely -
when a key-event manages to get into the mainloop before the idle
paint handler arrives and does nothing (except clear the paint flags).

So - don't do these big invalidations we don't need to in lok mode,
unless it is for dialogs - which presumably Pranav wanted fixed by
625087b5.

Change-Id: I88dda34b8d8bba9c89296d883ad9169fe49a7c5e
Reviewed-on: https://gerrit.libreoffice.org/71396Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

Change-Id: I1893d52df505bc43428c37a624ca05c569ba1bc0
Reviewed-on: https://gerrit.libreoffice.org/70958
Tested-by: Jenkins
Reviewed-by: Henry Castro <hcastro@collabora.com>
Reviewed-on: https://gerrit.libreoffice.org/71369Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

... so that 2017-07-10T09:11:02.999999... becomes 2017-07-10T09:11:03,
not 2017-07-10T09:11:02. The latter created duplicated items in pivot
table cache previously.

TODO: check what to do if the times are actually different by 100 ns?
What Excel does then? Should we increase cache item precision?

Change-Id: I622d1c784ee9fddf6b387bec2d8af87bae5668ba
Reviewed-on: https://gerrit.libreoffice.org/71610
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Reviewed-on: https://gerrit.libreoffice.org/71618Tested-by: Mike Kaganski <mike.kaganski@collabora.com>

Failed on Windows with

  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(925): error C2079: 'aDragParameters' uses undefined class 'std::array<int,5>'
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(927): error C3312: no callable 'begin' function found for type 'int'
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(927): error C3312: no callable 'end' function found for type 'int'
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(931): error C2065: 'rParam': undeclared identifier
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(935): error C2109: subscript requires array or pointer type
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(936): error C2109: subscript requires array or pointer type
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(937): error C2109: subscript requires array or pointer type
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(938): error C2109: subscript requires array or pointer type
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(941): error C2109: subscript requires array or pointer type
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(944): error C2109: subscript requires array or pointer type
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(945): error C2109: subscript requires array or pointer type
  C:/lo/src/core2/svx/source/svdraw/svdmrkv.cxx(930): error C2043: illegal break
  make[1]: *** [C:/lo/src/core2/solenv/gbuild/LinkTarget.mk:293: C:/lo/src/core2/workdir/CxxObject/svx/source/svdraw/svdmrkv.o] Error 2

Change-Id: Iaf2762a43a7eefce582bb187dd6978d5b28564e5
Reviewed-on: https://gerrit.libreoffice.org/71616Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>

As the patch comment explains, the properties
deck needs to be rendered with height long
enough to avoid scrolling, but others would
be greedy in using up all the available height,
so we render them in more restricted space.

Change-Id: I361e653239e22a7be0e5d75505cdcde18f2367d3

By resizing the deck we ensure that if the contents do not fit
in the existing height, they will after we expand to the minimum
height. Also, the minimum height is now the correct one.

This prevents seeing vertical scrollbar in LOK, which is very
slow and unfriendly, instead renders the full sidebar and lets
the client scroll, if necessary.

Change-Id: Idc2a622a65074b7afb26263624f2df0b5968f40d

And reformat.

Change-Id: If8a3dd4d16dfd51f8006d2d5d47951db30390fd8

Change-Id: I29d4eb38c72db50061266f0cf0ace31bab690315

Change-Id: I35b174c3a5e302ce52ee4063fa71d47feffab624

This avoids scrolling within the sidebar and makes
it possible to scroll the rendred sidebar image
in the client, which is more responsive and lighter.

Change-Id: I0713c699ba82bfc5fe503c9351402c67aef4494e

Change-Id: I0559a2a4ae041bc61cf3a2d5979656a7941f4b02

Change-Id: I210df42fec4404ed9629c9b1c24c78b680212214

The ViewShell, which represents the view in question and therefore
which view gets the window notifications, is incorrect when
the sidebar is first created upon creating/attaching a new view.

This leads us to a workaround to make sure that we publish
notifications to the correct view.

We also have to hide the sidebar instead of closing because
the workaround wouldn't work when re-creating the sidebar
on an existing view. See comments in code.

Change-Id: I9d3be901688291b04d634b68e1e20c7add77381f

Change-Id: I67babe9b340bafd196dedb2918d6a24bfc9c0ad5

These aren't fully functional, so not very useful in LOK.

Change-Id: If6330b1f7bf911bbb9edad00aaea7c4d37671746

Change-Id: I589fce5ee207da0e0ac113b7ccc1ed1edd06c5da

And enable sidebar routing only while we have a
valid SidebarController instance (i.e. not disposed).

Change-Id: I665f934f762c0ce9efdce5489ec9babe006376a8

Change-Id: Idb361164db9ea8681b45fe99608d2cb5690db161

Change-Id: Ia8f197d07ba4b6a089571e9ca23b4697d04e7734