This fixes the black rectangle instead of actual image content in the
bugdoc.

An alternative would be to handle this in the PNG import, but commit
66dbd4da (tdf#113918: Workaround: Load
1bpp indexed PNG as 8bpp indexed Bitmap, 2017-11-20) already tried
something similar and failed, leading to the revert in commit
25cd8436 (tdf#115297: alternative fix
for displaying 1bit images, 2018-02-27).

The test is especially useful with SAL_FORCEGL=1 SAL_USE_VCLPLUGIN=gen
specified on Linux, so the GL codepath is triggered.

Change-Id: Ia9d049d09dce2ac34826ee427f74616a96a35c88
Reviewed-on: https://gerrit.libreoffice.org/68925Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins

Change-Id: Ib08cca8faf2822e5a62f745c4213cf0ca2cc09d7
Reviewed-on: https://gerrit.libreoffice.org/68923
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

HarfBuzz 2.x has native support for AAT fonts which is, according to Chrome
developers, sgnificantly faster that HarfBuzz Core Text integration.

Change-Id: I4d5e861a1958402a6e3ccb720b10f40828c3db6a
Reviewed-on: https://gerrit.libreoffice.org/68919
Tested-by: Jenkins
Reviewed-by: Khaled Hosny <khaledhosny@eglug.org>

std::set::contains from C++20 is not here yet :-)

Change-Id: I50b5db2d44cb0effa64ab89a16205145883ce374
Reviewed-on: https://gerrit.libreoffice.org/68922
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>

The commit is based on the patch in https://bz.apache.org/ooo/show_bug.cgi?id=120066.

Change-Id: Ia4ad536331954679940104adf5d6178228c90854
Reviewed-on: https://gerrit.libreoffice.org/68814Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
Tested-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>

This is a test on a random file to see if all compilers accept this.

Change-Id: I412aef57730d07c8e6e2710fee71d5eded326d48
Reviewed-on: https://gerrit.libreoffice.org/68896
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>

CppunitTest_sd_import_tests actually fails with services.rdb
and CppunitTest_sd_misc_tests crashes :(

Change-Id: If11b658b139d85a2563768a11dd6979b721f3a92
Reviewed-on: https://gerrit.libreoffice.org/68918
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>

Change-Id: Ic8cd9babe3c196cdd66524f39a78940c9a029d2e
Reviewed-on: https://gerrit.libreoffice.org/68917
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

This should be more maintainable.

Change-Id: I60ee68cd79f7a0627bcaf8fd34d3036124ea0f9e
Reviewed-on: https://gerrit.libreoffice.org/68914
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>

Change-Id: Id39d322661f7537e8999acafee655c0cc16a78c1
Reviewed-on: https://gerrit.libreoffice.org/68911
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>

In a specific case, Err:504 was emitted by COUNTIF when its 1st
parameter was #REF!. This value was written into pivot cache definition
when exporting to XLSX. Clearly Excel only expected propagated #REF!
there, emitted error to user, and dropped the pivot table.

This problem made XLSX pivot table cache generated by Calc unusable
by Excel, despite ECMA-376 Part 1, 18.10.1.27 tells that the error
value is application-dependent.

The change makes COUNTIF to propagate errors received in its 1st
parameter, instead of owerwriting them with own Err:504. This will
likely need to be extended to other functions, if similar problems
would arise.

Change-Id: I459fa163a0e3e5f2c658631d66e096f4d46b1ace
Reviewed-on: https://gerrit.libreoffice.org/68868
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>

Change-Id: I77beadee964f08f87f2fe7cc0daef5cb91151b72
Reviewed-on: https://gerrit.libreoffice.org/68912
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: Ie9a6b5be0729e70525be660b88e7d47e65a72f66
Reviewed-on: https://gerrit.libreoffice.org/68909
Tested-by: Jenkins
Reviewed-by: Tor Lillqvist <tml@collabora.com>

Reviewed-on: https://gerrit.libreoffice.org/68861Reviewed-by: Andras Timar <andras.timar@collabora.com>
Tested-by: Andras Timar <andras.timar@collabora.com>
(cherry picked from commit 172f58ca)

Change-Id: I26c54ad52ab53802dc368b0bfcbde84affa46fdd
Reviewed-on: https://gerrit.libreoffice.org/68897
Tested-by: Jenkins
Reviewed-by: Andras Timar <andras.timar@collabora.com>

* Update helpcontent2 from branch 'master'
  - tdf#53524 (compl) Minor tweaks in Help page.
    
    Added ref to icons
    minor updates on tags
    
    Change-Id: Ia9812a5928554284a5a2fb5416983888e2437439
    Reviewed-on: https://gerrit.libreoffice.org/68915
    Tested-by: Jenkins
Reviewed-by: Olivier Hallot <olivier.hallot@libreoffice.org>
    

warn:vcl.schedule:22358:22358:include/vcl/task.hxx:107: Priority will just change after next schedule!

Just don't set the priority if it's already set to the correct value, so
we can avoid a warning each time the visible cursor is shown or hidden.

Change-Id: Ie11164db0af4cfba06a620c9b2426cb903af3887
Reviewed-on: https://gerrit.libreoffice.org/68907Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins

* Update helpcontent2 from branch 'master'
  - tdf#53524 help-submodule - grid control Dialog editor
    
    Change-Id: Ie6918fdc9eb0cfed7042362cb1ad223e908d6d8a
    Reviewed-on: https://gerrit.libreoffice.org/68895
    Tested-by: Jenkins
Reviewed-by: Olivier Hallot <olivier.hallot@libreoffice.org>
    

Change-Id: Ia0d2a883a3bd3aca96fe4c8f72f44c694837e568

... in SplitNode.

The problem is that first the new anchor position is set in the
SwFrameFormat, then SwDrawContact::DisconnectFromLayout() is called
(implicitly from its SwClientNotify()).

This then cause the a11y wrapper to be disposed and an event to be sent,
which then ATKListener::notifyEvent() immediately processes by
retrieving all of the children of the anchor SwTextFrame.

At this point, we get an assert from SwTextFrame::MapModelToView,
because the layout anchor frame is still the old one, but the model
format already has the new position, so the frame can't map the
anchor position.

Avoid this by explicitly disconnecting from the layout before setting
the anchor on the SwFrameFormat.

Change-Id: Iba8960729dd041e13de4963d1b2ab6b223c8a427
Reviewed-on: https://gerrit.libreoffice.org/68880
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>

* Update helpcontent2 from branch 'master'
  - Donation button and sticky header for Help
    
    Task: https://redmine.documentfoundation.org/issues/2815
    
    Header was made sticky on non-mobile by using position: sticky.
    This was the only way to make it work with the grid without
    carving magic numbers into stone. IE11 ignores it, but so what?
    
    Donation text was made shorter.
    
    The Google search and the donation button were wrapped in a
    container to make it easy to haul them all over the place.
    
    Lots of tweaks were made to the grid areas and the styles of
    the bottom elements to make everything fit nicely in the new
    markup order. The result was tested on IE11 to make sure it
    degrades gracefully.
    
    The Displayarea element had a fixed width (my old mistake),
    which made the min-width: 960px layout look very uncool.
    
    Change-Id: I131a1e82290bc3381c938acb4dc9dbc408fb2ba7
    Reviewed-on: https://gerrit.libreoffice.org/68890
    Tested-by: Jenkins
Reviewed-by: Olivier Hallot <olivier.hallot@libreoffice.org>
Tested-by: Olivier Hallot <olivier.hallot@libreoffice.org>
    

Noticed this on console:
cpp: <macro>:1 C:/BLP/core/scp2/source/calc/registryitem_calc.scp:1261 \
Bad token \text/x-ms-iqy produced by ##

See https://bugs.documentfoundation.org/show_bug.cgi?id=111344#c9

Change-Id: If6b5b4ae90c1b0fb812a5e2cd87d17fc688d21c6
Reviewed-on: https://gerrit.libreoffice.org/68885Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
Tested-by: Julien Nabet <serval2412@yahoo.fr>

Change-Id: I42fc97c37099906efe0ee626a44d13194eca25c3
Reviewed-on: https://gerrit.libreoffice.org/68902
Tested-by: Jenkins
Reviewed-by: andreas_kainz <kainz.a@gmail.com>

In SwXMLImport::SetConfigurationSettings() there is a list of settings
that will be ignored/not loaded form the file if IsLoadUserSettings()
is false, which is presumably not the default.

The beginning of this list is obviously settings that can be set via
Tools->Options, but the end is mostly a bunch of Word-compat layout
settings that cannot be set in the UI.

Let's assume some cargo-culting happened there (this function is tricky
and underdocumented anyway), and replace the end of the list with those
settings that have been added to Tools->Options->Writer->Compatibility
in the last years, since those can be called "user settings".

Change-Id: I6ea4a5da4033ca3512d71e4123b47d46e20bf596
Reviewed-on: https://gerrit.libreoffice.org/68803
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>

SwXMLImport::SetConfigurationSettings() needs to handle only those
settings for which the default is different vs. old OOo/LO versions;
DisableOffPagePositioning however defaults to false, so we can just rely
on that default.  The inverted bDisableOffPagePositioning check is
wrong.

(regression from fe3d5766)

Change-Id: I300fa597f58b586b49089bb555a00a6923862abe
Reviewed-on: https://gerrit.libreoffice.org/68801
Tested-by: Jenkins
Reviewed-by: Michael Stahl <Michael.Stahl@cib.de>

Change-Id: I8871eb49e11d8c72c469230967c3c979091a0f9b
Reviewed-on: https://gerrit.libreoffice.org/68898
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>

Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.

Change-Id: I18dc216a59a5770456b20162c881c8726e2f44b2
Reviewed-on: https://gerrit.libreoffice.org/68788
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>

Change-Id: Id46bf2b293c4465f5c58f503c75904639d59f168
Reviewed-on: https://gerrit.libreoffice.org/68701
Tested-by: Jenkins
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>

Change-Id: I4b701814e40f1ca3ced81c619696d7c3e18dc765
Reviewed-on: https://gerrit.libreoffice.org/68894
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>

By implementing BTLR support in
SwTextFrame::SwitchVerticalToHorizontal() (Point version).

Cursor traveling now looks good: all of up/down/left/right direction
have the correct paragraph and character position for all the lrtb, tbrl
and btlr cases.

As a side effect this also fixes mouse click, where clicking above the
paragraph positioned the cursor at the bottom of the paragraph.

Explicitly add a test for the mouse case as well, given that I initially
planned to fix the keyboard part directly in SwCursor::UpDown(), where
the keyboard test would pass, but not the mouse one.

Change-Id: Iabeded3f03a64416cfcaf58e0438c4a1a793e662
Reviewed-on: https://gerrit.libreoffice.org/68886
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>

Change-Id: I7ccb8286e7cfa692f16f4969716e3f018a92c4c3
Reviewed-on: https://gerrit.libreoffice.org/68892
Tested-by: Jenkins
Reviewed-by: andreas_kainz <kainz.a@gmail.com>

Change-Id: I8791d0f2cc33813c6034b771fa1f699143c3cfac
Reviewed-on: https://gerrit.libreoffice.org/68877
Tested-by: Jenkins
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>

Change-Id: Ic6d8082251246f3ca8c00b090679a0e590b0a8d0
Reviewed-on: https://gerrit.libreoffice.org/68876
Tested-by: Jenkins
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>

detect static variables that can be made const.

Thanks to mike kaganski for suggesting this.

Here I introduce a new plugin feature - using markers
in nearby comments to disable the plugin for specific
vars.

Some of this stuff was old debugging code. I removed the stuff
that was older than 5 years.

Change-Id: I6ec7742a7fdadf28fd128b592fcdf6da8257585c
Reviewed-on: https://gerrit.libreoffice.org/68807
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Move XNamed Java tests to C++ for ScDataPilotItemObj.

Change-Id: I2fcffbd7e7329357ac30fa14716bf2bcf49c5666
Reviewed-on: https://gerrit.libreoffice.org/68891
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

Change-Id: I65fb4a444e7e072004adcc9f9ad7af77a3f9f919
Reviewed-on: https://gerrit.libreoffice.org/68888
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

Prefix variable nMaxFieldIndex with m_ to show it's a data member of the
class.

Change-Id: I207b3552f460b3433af6892abb6c4693e990d7bb
Reviewed-on: https://gerrit.libreoffice.org/68887
Tested-by: Jenkins
Reviewed-by: Jens Carl <j.carl43@gmx.de>

workaround baseclass NotifyStartDrag weirdness rather than risk changing it

Change-Id: I70b56a913e6192da9b3a57653924dcc4160352c0
Reviewed-on: https://gerrit.libreoffice.org/68879
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: I911b6822b1c631700d7f43630e5a6fbb21de1c1f
Reviewed-on: https://gerrit.libreoffice.org/68878
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Regression from commit b160db92
(tdf#96947 vcl opengl win: fix background of menu items w/ check/radio
marks, 2018-05-14), the problem was that GetMenuPopupMarkRegion() did an
unconditional downcast from ImplControlValue to MenupopupValue.  This
looked safe, but when the extracted function was used in
WinSalGraphics::drawNativeControl(), then that argument was not safe to
downcast.

Fix the problem by switching to safe downcast, and returning an empty
rectangle (instead of a bogus one) on error.

This way the old bug stays fixed but fontwork popup menu works again as
well.

Change-Id: I2b69b9f2e6823850892ac73df78dac5c2f04ec03
Reviewed-on: https://gerrit.libreoffice.org/68883Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins

Change-Id: I8385acf7ecd46b7c8b990abb688bd2ddddcb6bc5
Reviewed-on: https://gerrit.libreoffice.org/68867
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>