(cherry picked from commit 12fc5b85)
(cherry picked from commit 7587f8ba)

Change-Id: I1ad477b9db9ba960ae2c417799e5d91b63538d57
Reviewed-on: https://gerrit.libreoffice.org/20354Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

Change-Id: Idc9e8fd0c35de48a2366dca3850bcc4e69ef997f
(cherry picked from commit afc565bf)
(cherry picked from commit 46c5197f)
Reviewed-on: https://gerrit.libreoffice.org/20368Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

When an unloaded linked picture comes into the visible view
(including repainting a page), SwNoTextFrm::PaintPicture()
starts a thread to load it in the background using the
TriggerAsyncRetrieveInputStream() method of the graphic node.

To avoid to start a second thread on the same graphic node,
TriggerAsyncRetrieveInputStream() checks mpThreadConsumer,
the graphic node member variable for the possible thread object.

The problem is that when the thread finished and
SwGrfNode::UpdateLinkWithInputStream() reset mpThreadConsumer,
the graphic object of the graphic node is still in unloaded
state (its type is GRAPHIC_DEFAULT or GRAPHIC_NONE instead of
GRAPHIC_BITMAP or GRAPHIC_GDIMETAFILE) for a while, because
its modification is solved asynchronously after several
SvFileObject::GetData() calls. In the intermediate state
of the graphic object, with the high priority repaints of
the new scheduler, PaintPicture() could start new thread
to load the image again.

Using the new member variable SwGrfNode::mbUpdateLinkInProgress,
this patch will prevent the graphic node to start newer thread
unnecessarily.

Change-Id: I9433f0fa4613294103a00a3955fc2f35d8863b59
Reviewed-on: https://gerrit.libreoffice.org/19974Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Reviewed-by: Andras Timar <andras.timar@collabora.com>
Tested-by: Andras Timar <andras.timar@collabora.com>

Change-Id: I222bd8bb4c647c2c7d2a673b95e4c9be3234ef86
Reviewed-on: https://gerrit.libreoffice.org/20273Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>

Presumably that was the intent?

(cherry picked from commit a933990a)

Change-Id: Ibfca90041d6913827d2f329bab448afe4d725b46
Reviewed-on: https://gerrit.libreoffice.org/20289Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

...called from within the signal handler, so any bets are off

Change-Id: Iedb5c7bc8d08350e5f3e3118c6713f5c25b238b6
(cherry picked from commit 9b3ca276)
Reviewed-on: https://gerrit.libreoffice.org/20282Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

(cherry picked from commit 50386403)
(cherry picked from commit 6408b735)

Change-Id: I006c739a9dc876dd9ec83375f76fd1343b39557a
Reviewed-on: https://gerrit.libreoffice.org/20312Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

(cherry picked from commit 2ca429e7)

(cherry picked from commit 9401a065)

Change-Id: I3b8620a4d8bffeb5e0558c35cb3a063238c12ba5
Reviewed-on: https://gerrit.libreoffice.org/20311Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

But you can use SAL_OVERRIDE instead.

Change-Id: I2697593606ef3ff63f923e364d0ea7e5661b1b3a
Reviewed-on: https://gerrit.libreoffice.org/20297Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Because i'm an idiot i pushed a commit without looking at the jenkins
build failure...

Change-Id: I8e7a02d6c1d20c2231288ea050986064f1f814c8

This is the same problem that

commit 133e04fc
Author: Caolán McNamara <caolanm@redhat.com>
Date:   Wed Jun 17 09:23:32 2015 +0100

    Resolves: tdf#91880 Invalidate graphics when the gtk window is destroyed

    not just when the GtkSalFrame is dtored

tried to fix, but that just made it more unlikely to fail

(cherry picked from commit 26c32cfe)
(cherry picked from commit aab8bed7)

Change-Id: Icba750c787adb6cd5c5ed0874ef07e6201c4cf25
Reviewed-on: https://gerrit.libreoffice.org/20231Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

partially cherry-picked from the
commit 187445b2

Change-Id: I9e74fb41448c6be0b8daa6c0f8a0e207be0be6d6
Reviewed-on: https://gerrit.libreoffice.org/20250Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: jan iversen <jani@documentfoundation.org>
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>

Change-Id: Id33d9aacdf647b4b1eb414970089b1e157f1a86a

Change-Id: I723122069541bc5752164554e693268267969b1a
Reviewed-on: https://gerrit.libreoffice.org/20240Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com>
Tested-by: Norbert Thiebaud <nthiebaud@gmail.com>

Change-Id: I6f999f36d6ef57dadddd774e6ef60ceb9cb8033c
(cherry picked from commit 9affe65c)

Change-Id: I58d60ee1121fcd1bf84ef998ed9116386dc3c70f

Project: translations  6d28ea63432c6b5844f224f68276096be053c1a3

update translations for 5.0.4 rc1

and force-fix errors using pocheck

Change-Id: I4d6d82a8ed09f9eb5874108b27ab758455a13566

Regression from commit 2b78f2cd (rhbz#988516:
DOCX import: fix context stack when importing header/footer, 2014-03-05),
though that just made an existing Writer layout problem visible.

RTF/WW8/newer (drawingML) DOCX import doesn't have this problem, as those
import pictures as sw graphics, not draw ones.

<w10:wrap type="through"/> is normally mapped to our page wrap (as it uses
"through" in the "not only wrap around, but also in the holes of the shape, if
it has any" context, not in our "text should go through it, so no wrapping"
one), but for some reason in this case (most probably due to the extreme large
negative margins) Word handles the situation as our through, i.e. the text
should not go to the second page, as it would normally happen with a "Word
through" wrapping.

Work around the strange situation by ignoring the wrapping request for extreme
top margin values.

(cherry picked from commit 37b5f1ed)

Conflicts:
	oox/source/vml/vmlshape.cxx
	sw/qa/extras/ooxmlimport/ooxmlimport.cxx

Change-Id: I20555b1fa7a769e20c40a3a5ff3873807403e937
Reviewed-on: https://gerrit.libreoffice.org/20212Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

"(abc)|(def)" matches "def" with result offset sequences
{0,-1,0},{3,-1,3}

And thus the assert() calls introduced with
4cf1d290 and
ce91f3c1 were hit.

Change-Id: I571b6c7d47349a2cc7b1d1e34193b2865012a49f
(cherry picked from commit 4dd2d406)
Reviewed-on: https://gerrit.libreoffice.org/20174Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Mike Kaganski <mikekaganski@hotmail.com>

modeled on the quartz impl

(cherry picked from commit 742da70e)

Change-Id: I2f6776c14c7350954932df2bffcca36172f13473
Reviewed-on: https://gerrit.libreoffice.org/20083Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>

instead of assuming they are in their final position when the
top level dialog resizes. Cause there's a timer/idle involved
so the final position isn't known yet

(cherry picked from commit d4d4e00a)

Change-Id: I713e938b0e275f181514315cb4ce4531f36191ca

loplugin:vclwidgets

(cherry picked from commit 4882664e)

Change-Id: I3e5d88bdc2ecae6164c75da8ddf1b8cfe0ed325f
Reviewed-on: https://gerrit.libreoffice.org/19796Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>

CppunitTest_writerperfect_writer file libmwaw/pass/Acta_1.0.hqx uses the
font "Courier", and for unknown reasons we can't properly load that
font, because the PhysicalFontFamily::mpFirst for "courier" is null.

This causes OutputDevice::GetTextArray() to fail to create a SalLayout
and return early before initializing the passed pDXAry, which then
generates lots of DrMemory warnings.

Let's hope the callers are happy about an all-0 pDXAry.

Change-Id: I07b29a59660cf5cd060fd77da5d96021f9d8f9f5
(cherry picked from commit e6e409b6)
Reviewed-on: https://gerrit.libreoffice.org/20206Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>

Needs a bit more work, so lets play it safer and default off for
windows for now and leave it on for 5.1

This partially reverts...

commit 1bc91425

tdf#93662 - allow GL by default only on recent drivers on recent windows.

Feature for triple review for 5.0.2 - enables GL by default for some subset of
very recent GL drivers, on Windows 7+.

Change-Id: I7a12540afd9f1889db15810db448db466b35241d
Reviewed-on: https://gerrit.libreoffice.org/19950Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>

Unknown names make the formula fail
PythonTest_sw_python, especially set_expression is ok
Variables are taken into account

Thanks to Mike Kaganski for his help on this patch.

Change-Id: Ia6f9c54d90ce88138fd9c9df9422b34ce8223ca2
Reviewed-on: https://gerrit.libreoffice.org/20122Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Mike Kaganski <mikekaganski@hotmail.com>
(cherry picked from commit d667e321)
Reviewed-on: https://gerrit.libreoffice.org/20159Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

As far as I see, the surfaces created in the DX canvas are of
D3DFMT_A8R8G8B8 or D3DFMT_X8R8G8B8 format, which means that the bytes
(in memory order, on little-endian Windows) are B,G,R,A/X. So if the
desired destination wants R,G,B,A, we need to swap the blue and red
bytes.

Let's hope this doesn't break anything else...

Change-Id: I1b90d2cf95418f6557cac633ec6fce27599e8a61
Reviewed-on: https://gerrit.libreoffice.org/20161Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

The old code in MultiSalLayout::GetTextBreak() only makes sense if the
base level character width is zero where a fallback level has a
non-zero character width, and vice versa. But this is not the case for
Windows, at least not any more now when using UniscribeLayout and not
SimpleWinLayout.

This simple change fixes that: Only use the width from a fallback
level if the width at the base level is zero. Hopefully it does not
cause regressions for other documents or on other platforms. (But, I
repeat, I find it hard to believe that the intent of the code could
really have been to ever add two or more non-zero widths for the same
character from different levels of fallback and use that.)

Change-Id: Ic66c55db4b7463f9e04fcedec76f1c44f5e62e03
(cherry picked from commit a0d112a5)
Reviewed-on: https://gerrit.libreoffice.org/20172Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

"errorOccured" is an English spelling mistake, but is the
name of the event, which cannot / should not be changed for
backwards compatibility reasons.

Spelling was boguously "corrected" in:

  commit 60a9b517
  Author: Takeshi Abe <tabe@fixedpoint.jp>
  Date:   Sat Dec 4 12:56:38 2010 +0900

      Replace all occured, occurance etc.

This was leading to impossibility of assigning a macro
to that event, since it was not appearing in this dialog.

Change-Id: Ib0373501cba8af8f4abb1c62faddf8d65c926a7e
Reviewed-on: https://gerrit.libreoffice.org/20180Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Noto fonts correctly.

LanguageTag()::getLanguage() would always return "zh" despite
traditional Chinese were choose. hb_language_from_string()
requires iso639-3 language tag except for Chinese, so
we use use Bcp47() language tag for Chinese to differ zh-CN and
the others.

Change-Id: I6700a79a9c49a97d840b430def77741912935f72
Reviewed-on: https://gerrit.libreoffice.org/20027Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit 4424daa2)
Reviewed-on: https://gerrit.libreoffice.org/20204

1) Write endParaRPr so that PowerPoint display them properly.
2) Original design forbid properites with default value to be exprted, now fixed.
3) Automatic colors are written as white or black based on
   whether background is dark.

Note that tdf#77881,tdf#80520,tdf#89525 depend on this.

Change-Id: I255c16f35149b738be2daf2800b1c90389f2c7cf
Reviewed-on: https://gerrit.libreoffice.org/17472Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com>
Reviewed-on: https://gerrit.libreoffice.org/19862Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
Tested-by: Katarina Behrens <Katarina.Behrens@cib.de>

various uninitialized variables in CppunitTest_sd_filters_test hang-8.ppt

Change-Id: Id0b9f146a7ab8b5bb2b36a5c2a39d65ee52e122d
(cherry picked from commit 78407eea)
Reviewed-on: https://gerrit.libreoffice.org/20157Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

(cherry picked from commit b1e5983f)

Change-Id: I493b79e70fb8003ce215e602f9b4ce92afa50150
Reviewed-on: https://gerrit.libreoffice.org/20156Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>

With simple transliteration, TextSearch::searchForward used to use
whole string to perform the search, then started to create substring
to search. But it left the precautions from
commit c00601da by Eike Rathke:
searching for $ may actually return a result set pointing behind the
search string which it does with the ICU regex engine.
The precaution made it to skip reverse mapping if index was 0.

Commit 9aae521b by Michael Stahl
didn't consider the case when nStop is 0, and startPos > 0. Then it
tried to get offset[-1].

Anyway, using value of startPos in those conditions seems illogical.

Removed those precautions (and made assertions for that).
Fixed handling zero indexes.

Reviewed-on: https://gerrit.libreoffice.org/19840Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
(cherry picked from commit 4cf1d290)

remove comment that makes no sense, tdf#94810 follow-up

nStart should not become <0, not even when searching for $ end.

(cherry picked from commit 67ab2ce3)

similar to searchForward() use the correct offsets, tdf#94810 related

(cherry picked from commit ce91f3c1)

Change-Id: I2066abc51fff7fb7323bc7f6198bdea06439d4f3
79586471f1a7908372fa3ee044a6f66499b1790f
cb5a45acaad9029e0998594de8b3dc2e9211866e
Reviewed-on: https://gerrit.libreoffice.org/20136Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Mike Kaganski <mikekaganski@hotmail.com>

On one hand, neither our binary DOC import, nor Word maps the "TOC
Heading" style to something special, and that's how the DOCX import
added that property to some paragraphs in the document, moving the
as-char picture from the first to the second page.

OTOH, the DOCX export filter has a lcl_guessQFormat() function that
explicitly assumes that such a style name exists in Writer document
models, so again it doesn't make sense to handle this style name with
special care.

Change-Id: I3af548930f9683695fc3ad56b486e013f107d61a
(cherry picked from commit 70f3c595)
Reviewed-on: https://gerrit.libreoffice.org/20037Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: I479a3083f56631135a7f6d94e2864583a9b30a30
(cherry picked from commit 660e37ad)
Reviewed-on: https://gerrit.libreoffice.org/20101Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: I7896d1d064cd92fd45a6b32ac64f534f68213169
Reviewed-on: https://gerrit.libreoffice.org/20064Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
(cherry picked from commit 96ff6372)
Reviewed-on: https://gerrit.libreoffice.org/20126

Project: help  ab4c21fa199cdaa11862f49e201af73d67c7850b

Revert "tdf#87313 Update instructions to Firefox 42"

This reverts commit 1f6ca79f4270e03fe256129147752fd7e6c8ab54.

Project: help  1f6ca79f4270e03fe256129147752fd7e6c8ab54

tdf#87313 Update instructions to Firefox 42

And remove references to the dead Mozilla Suite.

It seems that number recognition in tables are not working properly
enter 10-10-10 and it should be converted to a date but it is not.
I tracked it down to the fix of bug fdo#32082. It looks like bSetNumFmt
was changed to false by mistake. Since then it has changed name to
bSetNumFormat. From what I can tell fdo#32082 still works after this
patch, but I might have missed some nuance of that bug report.

Added two tests, one for the bug mentioned above and one to check
that number recognition is working. At least with a simple date.

Reviewed-on: https://gerrit.libreoffice.org/19563Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
(cherry picked from commit aa334d55)

Change-Id: Id58849a223eb602054c66c7379cd56a68a93dea2
Reviewed-on: https://gerrit.libreoffice.org/20082Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

this fixes a11y issue, where widgets inside VclScrolledWindow weren't 
accessible by keyboard as Tab, Alt-Tab events respectively, never 
reacheed those
Change-Id: Ic73e24f8036b0f6129c3a4f5efb160f9089cb0b8
Reviewed-on: https://gerrit.libreoffice.org/19982Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

ODF export in xmloff does the same

Reviewed-on: https://gerrit.libreoffice.org/20051Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Andras Timar <andras.timar@collabora.com>
(cherry picked from commit 11b7a2c8)

Change-Id: I78ddbef3ac96f10f334c45a467dafb95bb9e3453
Reviewed-on: https://gerrit.libreoffice.org/20066Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>