Change-Id: I96d081f394b0b62d99ec1034bf5e99da9aedd9d9

Change-Id: I7a45a61465f71a1cc6f583b9ffc68157d19e51fa
Reviewed-on: https://gerrit.libreoffice.org/11386Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Otherwise I get sw/source/core/layout/anchoredobject.cxx:582:

  <SwAnchoredObject::GetObjRectWithSpaces> - cache for object
  rectangle inclusive spaces marked as valid, but it couldn't be.
  Missing invalidation of cache. Please inform OD.

Change-Id: I3f315b15fca6e2480c11183269be5583e6aea123
Reviewed-on: https://gerrit.libreoffice.org/10977Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

If the anchor is valid, check if its page is valid too,
before moving anchors on view change.

Change-Id: Idffdfdb786cb29a105c06d652e6ad7315abfb4c5
Reviewed-on: https://gerrit.libreoffice.org/10976Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

Previously there was a restriction for the range to contain at
least two nodes.

Change-Id: Iadf21d3928303c2ab8df0d93dabd969ed1ff7e09
Reviewed-on: https://gerrit.libreoffice.org/10972Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

There is still a difference, if run mail merge throught one or
the other function (MM wizard or UNO).

Change-Id: Ia372bd3edea4c8dfb58e045fc8fed2efcb2a4fde
Reviewed-on: https://gerrit.libreoffice.org/10971Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

It turns out that document compare / merge and master document
insert file are all the callers of SwFindDocShell() so move the fix-up
there...

Change-Id: I912a464c3acab8d0adb7b5b659d62f629a15cb70

Change-Id: Ice3d2bc4a575b324377273cdb655e06577f6a4a2

Change-Id: I7195bba02474ac1c00fce0a93f59be622a920686

Change-Id: I2ad27cb09031dfadaa628939c191ccc5d88ecffe

Change-Id: I66255e8bc37191e859c134eab7d7f8117cd9181a
Reviewed-on: https://gerrit.libreoffice.org/11380Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

Change-Id: Ie21c61d718d720ec6bac741cd17e807123f47129

In commit 88a874fc "convert SfxItemState constants to a proper enum"
I made some mistakes in converting bitwise logic to boolean logic.

I fixed one of those places in commit 7ad83656 "fix bitwise->logic
conversion in SfxItemState commit"

This commit fixes the other places where I converted bitwise to normal
boolean logic. I also validated that none of the existing code tries to
uses combinations of these enum values.

This commit also introduces an exception-throwing check in the one place
where the enum is explicitly cast to make sure that no combinations
sneak in.

Change-Id: I545f7d17b76c4fd999078867caec314e83ffe165
Signed-off-by: Stephan Bergmann <sbergman@redhat.com>

from ParagraphObj, by using boost::ptr_vector.

This is also a supposed fix for ParagraphObj::operator=(),
because ParagraphObj::ImplClear() now makes the vector empty
while its original code kept deleted pointers.

Change-Id: I7710eb33fc05bed55474ff00baaeb8e19cd79f3d
Reviewed-on: https://gerrit.libreoffice.org/11379Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

hmm, make CppunitTest_i18npool_test_breakiterator fails after all

This reverts commit 693f72384e081429cddb29415276cc4d07b05645.

Change-Id: I3339324ec99bbed56f81f2aa4e6cf54edcbec8e6

Change-Id: I88eeac06413fc3935cbbdb357ff8bf9acaa383a8

Change-Id: I64be7ca711dcd3ea7c4d5840a30f2f701d055e1b

Change-Id: Ic8751562e4dbfa5b6be903d87d369a750d305495

For the temporary document there is no layout so SwTxtNode::MakeFrm()
isn't called.

(regression from 6db39dbd)

Change-Id: I711d32a81519edd6840cc29f5cd9dbc57ef21982

Change-Id: Idcd7de83ff7e551b350db0bf80e2ac78baa67686
Reviewed-on: https://gerrit.libreoffice.org/11374Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

valgrind + bff

Change-Id: Ib3ed8a6e518c0686f8cbeaf021b9ca3a07005032

Change-Id: Id1c48d99f18736de1e9b0edf4b8a77becaa1c208

Change-Id: I90135b6d7780ba8a22fff725a965f0ea7e442707

Change-Id: I94fed6a9361f21457b3e631efffc0db833068aef
Reviewed-on: https://gerrit.libreoffice.org/11256Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

Change-Id: I274236939c6de4db811f8932025c619479bebbee

so that the global aLotusPatternPool map gets destroyed at a sane point in time
where calling delete on its patterns is a legal operation and not at some
random location in the dtor chain

Change-Id: I6f202c5345c6a227b5a680d70150bf32fb9450c7

Change-Id: Ie31c16318b09699e080484292d489a378e3a6dce

Change-Id: I15ead38cadbee5d092c96010c8e9a7253c2027cc

When the Xft.dpi setting is not found, and if only one screen is
available, fall back to detecting the screen DPI from the reported
resolution and monitor size.

Change-Id: I1481d94b5dc4072c1f8da4659a221dfc7971080b
Reviewed-on: https://gerrit.libreoffice.org/11377Reviewed-by: Chris Sherlock <chris.sherlock79@gmail.com>
Tested-by: Chris Sherlock <chris.sherlock79@gmail.com>

Move the validity check for the DPI (>= 50 and <= 500) into its own
(static) function, since we're going to use it repeatedly in the
fallback DPI detection code.

Change-Id: Id5d9b10be385d45549bbb57cfa14939cd51a0324
Reviewed-on: https://gerrit.libreoffice.org/11376Reviewed-by: Chris Sherlock <chris.sherlock79@gmail.com>
Tested-by: Chris Sherlock <chris.sherlock79@gmail.com>

Just like the debug documents in SwDBManager::MergeDocuments.

Change-Id: I275575853c80d4e19d6df5ba9d8dc689353d7c85
Reviewed-on: https://gerrit.libreoffice.org/10970Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

This adds a lot of SAL_INFO messages to SwDoc::CreateCopy and
SwDoc::Append.

Change-Id: Ie924e20698bec84482fcc5496175a9e29245729e
Reviewed-on: https://gerrit.libreoffice.org/10969Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

This drops all the specialized, workaround code from MM, introduced
to use the SwFEShell::Paste function and additionally merges
and renames SwDoc::Paste into SwDoc::Append.

There is still a little common codepath, therefore this adds
comments to always update both functions.

Change-Id: I704b3ef3257dd977dac95e16e25049ff8ade97ed
Reviewed-on: https://gerrit.libreoffice.org/10967Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>
Tested-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

Change-Id: I83b9c86685f72569c0df88231458183b1ae511c8

Change-Id: I6627363aea34570b5f7d118b1d64b6a2d93c19cb

Change-Id: I10e4a0aa23fe83fa0f2edd671b321fa4588c880f

Change-Id: I97d5a8dc85fe77c6bf0235425db0a7b9e42ff869

Change-Id: Ibf478abae5aa0574279e8b45c26b96ab66bfe203