Change-Id: Ia79bc8d35cf19438d177f31a8e0c8603102b4ef1
Reviewed-on: https://gerrit.libreoffice.org/35061Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>

Change-Id: I23b16792005d66588e3439eb6482adda96c45673
Reviewed-on: https://gerrit.libreoffice.org/35064Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>

Change-Id: I2be7113bfe86476acfd6550f1eed53eed83c9747
Reviewed-on: https://gerrit.libreoffice.org/35063Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
Tested-by: Julien Nabet <serval2412@yahoo.fr>

Change-Id: Ibefaed54c8fdce22fbae0116b4df5db4023ef7d6
Reviewed-on: https://gerrit.libreoffice.org/35057Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>

Visual Studio 2015 has a Community Edition. Use that. VS2017 has no
Express variant at all.

Change-Id: I23bcd10c543a7626b2d69d64306fa5eac74e3b27
Reviewed-on: https://gerrit.libreoffice.org/35009Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: David Ostrovsky <david@ostrovsky.org>

...which needs to be past at the end of the command line, after a /link option.
Looks like at least MSVC 2015 Update 3 silently ignores the garbage /D option
-DEBUG:fastlink, while clang-cl complains.

Fixes 96af392b "Use /debug:fastlink linker
option to improve link performance".

Change-Id: Ib679d4ce3ac4a7622ba7b066edbfe1872892137a
Reviewed-on: https://gerrit.libreoffice.org/35048Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: David Ostrovsky <david@ostrovsky.org>

Commit aa68c99d added some code using
std::condition_variable to comphelper.

Built with MSVC 2017, this causes many cppunittester.exe processes to
deadlock in ThreadPool::shutdown():

    maTasksChanged.notify_all();

This ultimately calls NtReleaseKeyedEvent(), which never returns.

The reason appears to be a bug in Windows 7, for which a "hotfix"[1] is
avaiable here, but it's apparently not distributed via Windows Update
so we likely can't rely on users or even developers having this installed.

However, the documentation of DllMain[2] and ExitProcess[3] indicates
that during shutdown, by the time global destructors are invoked
all threads other than the one that called ExitProcess have already
been terminated.

Returning from main() implicitly calls ExitProcess [4].

As it turns out the problem only happens for some CppUnitTests because
soffice.bin will call ThreadPool::shutdown() from Desktop::doShutdown()
while it is still safe.

[1] http://support.microsoft.com/kb/2582203
[2] https://msdn.microsoft.com/en-US/library/windows/desktop/ms682583(v=vs.85).aspx
[3] https://msdn.microsoft.com/en-us/library/windows/desktop/ms682658(v=vs.85).aspx
[4] https://blogs.msdn.microsoft.com/oldnewthing/20100827-00/?p=13023

Change-Id: I6137461ca7efe9a5fbe4f8f8478fb96de3570469
Reviewed-on: https://gerrit.libreoffice.org/35066Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

Same code which was added to the method
DrawingFragment::onEndElement() in
d178d7be

A shape's size and position inside a chart object
are primarily defined by <cdr:from> and <cdr:to> tags in
drawing.xml, but in the same file we also have
an <a:xfrm> tag which is used for the same thing
(defining size and postion) in general. Testing with
MSO it seems it ignores what values are set to <a:xfrm> and
uses only the <cdr:from> and <cdr:to> values.
The test document showing this problem must be created
an earlier version of MSO, becuase <a:xfrm> is set to some
random values. Editing the chart and saving it with MSO
2013 the <a:xfrm> values are set consistently with
<cdr:from> and <cdr:to> pair.

Change-Id: Icbaff78e5a6fcea36678c3b3755c97f57976b626
Reviewed-on: https://gerrit.libreoffice.org/35069Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>

Change-Id: I810241ad4624ae2ad4e42e2dbf4c226f622a35ad
Reviewed-on: https://gerrit.libreoffice.org/35070Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>

Might be more of them left, I found these by looking for "is save" or
"it save".

Many more potential misspellings of "Safe" as "Save" in identifiers
are left. Need to check individually which of them actually mean
"Safe", which mean "Saved" perhaps.

Change-Id: I5d3791ea2df2e1d485f65400404972b72af45aec

Change-Id: I6fab05c05c8f79607b75429f32be0497b54b3cb7
Reviewed-on: https://gerrit.libreoffice.org/35059Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>

Change-Id: I69f0144006e4ed226e8b158c0cc9f60dda0a9184
Reviewed-on: https://gerrit.libreoffice.org/35060Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Zdenek Crhonek <zcrhonek@gmail.com>

Change-Id: Ic54e808956e5cf4e8079942c0ff799f802cd4b6c
Reviewed-on: https://gerrit.libreoffice.org/35053Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

use OUStringLiteral to store format names.
just create ResMgr, no need to check if its gone.

Change-Id: Ie3e33bc70b975d8ee49f47dcf23737ca56652f1a
Reviewed-on: https://gerrit.libreoffice.org/35007Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Change-Id: I19d30a1ef4236dbe0a0053dad67a5ba7b1586fc8
Reviewed-on: https://gerrit.libreoffice.org/35054Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Johnny_M <klasse@partyheld.de>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>

Since this test round-trips well, move it to ooxmlexport

Change-Id: I95038e27a5779687152cca2bd31fe5719cac65b6
Reviewed-on: https://gerrit.libreoffice.org/35034Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Justin Luth <justin_luth@sil.org>

Change-Id: Ia78492ce4b30dc46004037448c613de1ff07c601
Reviewed-on: https://gerrit.libreoffice.org/35065Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>

Change-Id: I9014e068576d68613da6a70526dcb5b8e278e3f1
Reviewed-on: https://gerrit.libreoffice.org/35027Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Zdenek Crhonek <zcrhonek@gmail.com>

We want to always prefer Graphite shaping when supported by the font,
which is also what HarfBuzz does by default.

Change-Id: I6670fc03b8e6b3d7e07e1b8e0062880524da1655
Reviewed-on: https://gerrit.libreoffice.org/35046Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Khaled Hosny <khaledhosny@eglug.org>

roundsig( ,15) breaks some buildbots

Change-Id: I53715a698d20df78458056c3a90db8e88116e77c
Reviewed-on: https://gerrit.libreoffice.org/35049Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Zdenek Crhonek <zcrhonek@gmail.com>

Change-Id: Icba12a6218086079a7d5538f7263bccb1f2ee1b0

Change-Id: If03837dff5d7db6907f13c32ce05a74e4a0b7b55

Change-Id: I147b68882d99b8712625178e5206ea7c0f5f7401

Change-Id: I6f0c06491cf3c09d97b637d7456debd1e3decbcd

This replaces f764d67d with a better
fix, no point in littering the makefiles with manual dependencies when
we can just avoid the problem in one place with a tiny performance cost.

Change-Id: Ic7e91ee4c99deb12a9544cf691c0995de6ae2bad

our qa tests already include such, so this isn't right

Change-Id: I86fb02bf353fb5b228db915f04db6e36cfb40c51

Change-Id: I88e44e637e2ac93b4deed42a2d0f94b0a973c995

Change-Id: I2c7f607f05b07b7ec5453758910fdcb7f27ae697

Change-Id: I13eb327673af451cc81d4134ec8fedb33702c0ac

Change-Id: I07c13c354df1d7f523b08a39f2ed3f008051f7aa

Change-Id: I73814e38437f83f8aa7b87addc285e82f7357bea

Change-Id: I4040ebbf7b21905b950f4af36012bc83aebaa320

Change-Id: I750bee7241e36b5c5fe9075cb3ce606f2b9bbd5e
Reviewed-on: https://gerrit.libreoffice.org/35042Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Andras Timar <andras.timar@collabora.com>

- instead of keeping vcl resources in a rtl::Static
- these are shared owned by all EditEngines
- they let go of them when disposed
- thus, this should prevent VclPtrs to leak beyond dispose()

Change-Id: Ic6198c19d0ff9c09ecdea377d07807f08711d6a0
Reviewed-on: https://gerrit.libreoffice.org/34986Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

In an ideal world, by the time we reach DeInitVCL() there should be
neither non-null VclPtr nor EditEngine instances around. This fixes the
former for those VclPtrs owned by EditEngines, by disposing them when
SfxApplication is dying. The goal is -- for now -- to not have any
non-null VclPtrs around by DeInitVCL(), so we can e.g. assert() on that
to pick up regressions.

Should we one day have all EditEngines accounted for with proper smart
pointers that get torn down before DeInitVCL(), the Dispose() introduced
here might be removed again.

Change-Id: I47b17f7fd31fce2ac5fd6f6b146d356d7d07e677
Reviewed-on: https://gerrit.libreoffice.org/34985Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

Change-Id: I68d6cca1b359aa8fba42663bddb1107c31102415

Though only the closing 0-character and the following check excludes that,
dbgutil asserts.

Change-Id: Ife1299042a60f6f058c4cf58b406d1cc022786a7

Change-Id: I4f6e6e8e77cdabe593bca5719b6ef38aeecc5da7

to avoid super slow polygon math

Change-Id: I30ef0afe96d9863df27ef7a4e7f7e8e7af36f8c1

Visually, this unit test round-trips terribly.
During round-tripping, the 3 columns (related to textsections)
in the very first section get transferred into the pagedesc. Thus the
first section is "lost". For some unknown reason, another section
is added later on. So, cummulatively, the section count remained
the same, but whatever fixes these two situations will also affect
the section count.

However, it should work fine as an import only test.  There ARE
three sections in MSWord after all.

Change-Id: I1973977f5c5a45189896b10cacd757600e7684db
Reviewed-on: https://gerrit.libreoffice.org/35012Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Justin Luth <justin_luth@sil.org>