Change-Id: I210d0ded9cb38569b0831355522d8309cee05f56

Change-Id: I91d2e956b76919c56d14712789198c3c10470742

It will make implementing the background compilation of OpenCL kernels
easier, I hope. That needs data accessible through the ScFormulaCell.

Change-Id: I77ad5a94eb13551e6238d5f5285f1c916d7cb992

Change-Id: Icd851c17bab1f55cd16d5a48cc71ff2ac91daa2c

Change-Id: I9fbf848bd487e5ea49a383461f04e3c1678af607

Change-Id: I1a21ba33dd5c5c87938efaa9a779e2eab4f2778c

Introduce a new abstract class CompiledFormula, of which the
(existing) class BinaryKernel, that represents one OpenCL kernel, is a
concrete subclass.

Change-Id: I70fd26fa4bd3de026276727bf9fa8ed01bfb62c9

For single-file mail merge this skips the page format normalization,
which would be run during each document / page append.

Instead the normalization / checks are run for the whole document
at the end of the generation, changing this code from O(n2) to
O(n). This way each mail merge generated document uses the same
time, because the target document isn't completely traversed for
each page change.

With my test document this produces the same output, save some
changed internal document IDs.

This fixes fdo#56355.

Change-Id: Ia779705cfb107383341f946542d28434c986591a
Reviewed-on: https://gerrit.libreoffice.org/6733Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

Change-Id: Ic291f9cdc7413d125a79a1e79ef8215de32bce3b

Change-Id: I9cc6ae4fc5fc67837fcfc5fdea92baadd032996e

--> no shrink of master table, when follow table current is layouting its lower
table row

(cherry picked from commit c037ec6a)

Conflicts:
	sw/source/core/inc/tabfrm.hxx
	sw/source/core/layout/tabfrm.cxx

Change-Id: Ic1366b20e421d582977d44c6dd7dae3865f53ec4

Change-Id: Ia7ec6c4e9808a379e955a013264e1f92a7f6a358

Change-Id: I9a1709eb0d5bb81c6c90c4b9f90ff7bed74cda67

Change-Id: Ic31811c287e2dc2908243a228a340c3005077297

Change-Id: Iedd4075eadce9ca8fc41b279ea03c2679b01ec71

where the inner one deletes Shells that the outer one is still
processing. Push the candidates onto a stack and let inner
FlushImpl modify them to inform outer FlushImpl's that an entry
has been deleted

Change-Id: I1db8546d53e24cc96c72f2cd5cbec57b6cecaff5

Change-Id: I59d553f7b70955de18562046f1d7aa5fae6e9fb1

Change-Id: I82b093f7778a9d85016903cad1e39fc97f696cf2

Change-Id: Ibd8574dd74398c8b231059ca606edacf7f8005c4

Change-Id: I34bea8aaedec8df47874c5230c87455aa1ce6883

Fallout from f01580ce, vs2012 was
complaining about mismatching IE_VER / WINVER in boost.

Change-Id: I2092aba9fa64740eb796538855caaefae9bceab6

before the changes clewInit was called by fillOpenCLInfo. Now we need to
do that manually.

Put back the definition of WITH_RPM: otherwise a build
--with-pkgformat=archive fails since the scripts/install is not built.

Leave in the change to installer that should fix the original DEB bug.

This reverts commit 4f995b39.

Change-Id: I0cf4ba003237c66dfb7a0677d6a32a69f8203c6b

Change-Id: Idf9e7ddaf4b28c00222470460478631c37c79a6b

Change-Id: Ifd1ab147926ab4f1d34161840d17fdf1ab3f9810

Normally scoped_ptr is preferred for local heap objects, than direct
new and delete pair.

Change-Id: Ia8783bfe5f9b37a9bd58f08cce89bf85d88199d1

This change also fixes memory leaks.

Change-Id: I68a02a6de0b3d1f125e4c6cd164980a92e0fcbf0

Change-Id: I5f6f50e941031284fe519999a2fddede9753abea

Change-Id: Ife109b8a123564c5d8e43c793fc6c9de9b9f2090

Change-Id: I8af49ccf498cafb48f9b82fabc4910c754ba0f96

This reverts commit acebbee9.

I mistakenly converted OUString::equalsL calls to OUString::startsWith
calls.

... otherwise it will get inherited from Executables to StaticLibraries
and then mt.exe will be unhappy and complain.

Change-Id: Ic9cc32ea3d0fb5d1fe91c1e9efa354fdf30c8264

Complains with warning LNK4044 and ignores it.

Change-Id: Ie91369339d121a652ce51e9229c3ef3b1c39d214

This reverts commit 5003ab57.
It did not work, when both RPM and DEB were selected as targets.

Change-Id: I90c74f0d00804b80daeca2c69d1f12e094243a81

See f0a9ca24 "rhbz#887420 Implement 'block
untrusted referer links' feature" for details.  This adds some further /*TODO?*/
comments, and one known problem (marked /*TODO!*/) is that movies/sounds are not
blocked during a slideshow presentation.

Change-Id: Ib2d0c7e4f7b02c4bdec0d8a90cee5e7e1bee8325

Change-Id: I88fc77ea6c0b6a7c02d0b7eff64d8fdd3db6020e

This reverts commit 20f029c6.

The native X11 and gtk+ backends both have this functionality,
and both with a hard-coded 100, and the gtk+ one has had active
maintenance and testing. I'd like to leave that there, I suspect
it is used in some obscure multi-threaded corner-case where we
want to process events but only a few of them.

QAbstractEventDispatcher::processEvents always processes all
pending events, so we can ignore the the second Yield parameter
(bHandleAllCurrentEvents).

Change-Id: I49a70ed13d0215b89f48eed93d78eef2f083dc0d
Reviewed-on: https://gerrit.libreoffice.org/6686Reviewed-by: Thorsten Behrens <thb@documentfoundation.org>
Tested-by: Thorsten Behrens <thb@documentfoundation.org>

The Qt event loop may start a nested event loop, when checking for
clipboard and Drag'n'Drop events.

Previously this was handled by running this nested yield loop
inside the main glib loop using
  qApp->clipboard()->property( "useEventLoopWhenWaiting" );

But this results in nested paint events which crash LO:
  QWidget::repaint: Recursive repaint detected

To prevend yield mutex deadlocks, check for nested event loops
and always release the yield lock before starting the nested Yield
event loop.

This fixes fdo#69002.

Change-Id: I7e827abd3489783053ec7123372742a32555875d
Reviewed-on: https://gerrit.libreoffice.org/6685Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
Reviewed-by: Thorsten Behrens <thb@documentfoundation.org>
Tested-by: Thorsten Behrens <thb@documentfoundation.org>

If a KDE4FilePicker is opened via an remote UNO call, e.g. via
a RPC Java call, a JVM is started in a new thread to handle the
call.

This was creating the KDE4FilePicker in the non-UI JVM process,
which crashes LibreOffice.

Therefore we redirect the processing of all FilePicker calls to
the main thread.

This fixes fdo#71145.

Change-Id: If6ec2d205af5a883df35fddb44a12ac43e3560f0
Reviewed-on: https://gerrit.libreoffice.org/6683Tested-by: Michael Meeks <michael.meeks@collabora.com>
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>