Version 5.2.1.2, tag libreoffice-5.2.1.2

Change-Id: Iebf6d3cc430f6d511982c0e1bbdacf69148139cf

bump product version to 5.2.1.2

Change-Id: Ib26b39d9fedfe2e4056bcb9aea7f88d22b0e8353

libpng: upgrade to release 1.6.24

Change-Id: I5528522d866262655eadcfcf09c038e0c6097bb9
Reviewed-on: https://gerrit.libreoffice.org/28364
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
(cherry picked from commit 1efbdfcffff97ec0834274d13abe0f0d318b5f3c)
Reviewed-on: https://gerrit.libreoffice.org/28367
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
(cherry picked from commit f90be96ed400a5471d6c3a5cfa5087957803a9fe)
Reviewed-on: https://gerrit.libreoffice.org/28370
Tested-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>

Resolves: tdf#101460 use most recent poppler release

Change-Id: I660e1f13eb4b05fdb7257f422eca288c2d07cf4d
Reviewed-on: https://gerrit.libreoffice.org/28087
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit 10290000117b562e3f179c3d1073774dd54cdfee)
Reviewed-on: https://gerrit.libreoffice.org/28366
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
(cherry picked from commit 92986a1d516195ed73ad2036ee0647a54baf84b6)
Reviewed-on: https://gerrit.libreoffice.org/28369
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
Tested-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>

Resolves: tdf#101137 handle empty elements in matrix concatenation

Change-Id: I2de90e1cdfc4f215f6d7c644e0ab74372a4c2639
(cherry picked from commit 31d39c24e1ff2d3602e0212eeeecd8de88da6d38)
Reviewed-on: https://gerrit.libreoffice.org/28336
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
Tested-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>

Resolves: tdf#88953 byref args incorrectly passed in basic->c++

partial revert

of

commit 22b094f5d8e1e82375b135abd3a6f99a9a555244
Date:   Tue Jul 14 14:50:07 2015 +0200

    loplugin:unusedmethods basic

and partial revert of...

commit e2080e70fe8b085f18e868e46340454720fa94ca
Date:   Wed Jun 18 12:14:29 2014 +0200

    new compilerplugin returnbyref

and revert of...

commit 536051f8862203e0e115a5394a6379acd83cc8fe
Date:   Wed Jul 15 14:04:01 2015 +0200

    fix Windows build

    after commit 22b094f5d8e1e82375b135abd3a6f99a9a555244
    "loplugin:unusedmethods basic"

Change-Id: I612937334fdb75365080c98a9d4da5ed7ae647e3
(cherry picked from commit 4a647a04a7881964ce13b541399f89e4ab042ea8)
Reviewed-on: https://gerrit.libreoffice.org/28251
Reviewed-by: Eike Rathke <erack@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
Tested-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>

tdf#101661: Report XDG_{DOCUMENTS,TEMPLATES}_DIR only when it exists

...as had been done in the GConf-based backend prior to
3cf557c12d27f1b2250e69a543136da098112d80 "drop gconf integration as per ESC
decision".

Change-Id: I56f256aaec542e0a62fc30a37d5c982e93e77812
(cherry picked from commit 4278cac8935339dddfa5473490124d11061a0b38)
Reviewed-on: https://gerrit.libreoffice.org/28344
Reviewed-by: Eike Rathke <erack@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
Tested-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>

update credits

Change-Id: I20497f102f212a66525ca7be2771baef4b8b55a5
(cherry picked from commit 1b78ee6fc00924fd8d502dc92b82738a404cee34)
(cherry picked from commit 7b6affd1ceeed34a279b9b11895af52ae839c59d)

update emoji autocorrect files from po-files

Change-Id: I1a90e11daf4e1cc3e56984d340a0a43c84d7166d
(cherry picked from commit 53d0d98cc5351cfa4af3059c112db620acf667a2)

Updated core
Project: translations  75ff2af971a9c8cd1ebc00dcbd4723f05c82a358

update translations for 5.2.1 rc2

and force-fix errors using pocheck

Change-Id: I596fa1e8acc5be5e9ef1867b378b291a6853d935
(cherry picked from commit cc220e305045212ebb6311476ff3999d88fa6def)
(cherry picked from commit 42510aa08ecb272537241e7d4afaf6539c5c97b0)

Resolves: tdf#101347 conditional formatting super slow on exit

since...

commit 88a0c7d01b7dfd085a0569030f97cc7de0f0d106
Date:   Sat Mar 26 13:11:53 2016 +0100

    switch to a listener based cond format update, tdf#95437

I'm obviously assuming here that the notify doesn't matter
when the broadcasted is shutting down.

Change-Id: I0fef0ac1c1d8757199dcffb3c8ec1b05a73c1f17
Reviewed-on: https://gerrit.libreoffice.org/28006
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit cbab44f5eb27c8a45905167d2443f56b816b4a7b)
Reviewed-on: https://gerrit.libreoffice.org/28205
Tested-by: Jenkins <ci@libreoffice.org>
(cherry picked from commit 5c1812103a6c0cd29dd18e546d29ee683f675f6d)
Reviewed-on: https://gerrit.libreoffice.org/28238
Reviewed-by: Eike Rathke <erack@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>

Resolves: tdf#100901 crash pressing tab in r-o document with hyperlink

rather odd union in use here. Trying to call SwTextField::GetStart
on SwTextINetFormat blows up under visual studio.

Change-Id: Ic8145d7645bd6a68ef19e018311a4de6e6958bcb
(cherry picked from commit 3196e949bb23a33bdb8700dbe27782e0e6c8f1e6)
Reviewed-on: https://gerrit.libreoffice.org/28243
Reviewed-by: Eike Rathke <erack@redhat.com>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

XML number format: do not enquote space followed by minus, tdf#97837 follow-up

Which with
[$kr.-406] #.##0,00;[RED][$kr.-406] -#.##0,00
saved and reloaded gave
[$kr.-406] #.##0,00;[RED][$kr.-406]" -"#.##0,00
and thus an almost identical duplicated format with the built-in format.

Change-Id: Ie39b97576842bc29ac301d2fe64f7331f6c57fc4
(cherry picked from commit 4d993ae8f267e35f7c030861a92226c940bb46cc)
Reviewed-on: https://gerrit.libreoffice.org/28041
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@canonical.com>

tdf#100965: Restart on initialisation-time OpenCL crash

Add a flag to the OpenCLZone indicating whether we are performing the
first-start OpenCL functionality verification, so that if we run into
a crash that is caught by the VCL VCLExceptionSignal_impl() handler,
we terminate the process with the EXITHELPER_NORMAL_RESTART
status after first having disabled OpenCL use. The wrapper process will
then restart soffice.bin. This is for Windows only so far.

This matches what we do if OpenGL fails early during start of
LibreOffice.

Also, the enter() and leave() functions are not used anywhere

(cherry picked from commit 32881f01833dbcefd5600e1135dd8743178bfd96)
(cherry picked from commit b9898f03eb05411c508b1b02588812074d40417a)

Change-Id: Ibb9bf3a86b7521bf16728de2a118ad4323be674b
Reviewed-on: https://gerrit.libreoffice.org/28139
Reviewed-by: Eike Rathke <erack@redhat.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Add missing Loader$1.class

...that was introduced with abbf4777f29374025d576ef8daa3f6dcba02ddf5
"cid#1326844: DP: Use doPrivileged"

Change-Id: I8cd4d947b258313d4d171c5888490d1a860ebee7
(cherry picked from commit e175592a2732c34e16bcd9c19dc465e4974071c3)
Reviewed-on: https://gerrit.libreoffice.org/28195
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

GALAXY: Add show track changes bar small icon

Change-Id: I8166fc40e683899ab562c2c80c06336d8a7b7598
Reviewed-on: https://gerrit.libreoffice.org/28051
Reviewed-by: Yousuf Philips <philipz85@hotmail.com>
Tested-by: Yousuf Philips <philipz85@hotmail.com>
(cherry picked from commit 1fcb7eafd95ebed7f5532d86a8cd8cd89196883e)
Reviewed-on: https://gerrit.libreoffice.org/28053
Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
Tested-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
(cherry picked from commit c324c79b087c98e024a2ae130ce3993346797fee)
Reviewed-on: https://gerrit.libreoffice.org/28054
Reviewed-by: Eike Rathke <erack@redhat.com>
Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
Tested-by: Katarina Behrens <Katarina.Behrens@cib.de>

Revert "tdf#98837 - Fail loading a new document with read only property ..."

There was not really a good enough reason for the change, and it
causes unfortunate regressions in 3rd-party code that previously
worked fine. (It also caused two regressions in LO's own code, but
those were already fixed.)

This reverts commit 5334ff287c65f028753171c7b38da7ffbcebc1cc.

(cherry picked from commit 38a3743e0c5d52f9386f74097fd512d3133fbbe3)

Change-Id: Iec36345fa3663bc5d96ce3e74d334c7aa075ada2
Reviewed-on: https://gerrit.libreoffice.org/28036
Reviewed-by: Eike Rathke <erack@redhat.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jan Holesovsky <kendy@collabora.com>

bump product verion to 5.2.1.1.0+

Change-Id: I6df7667fc8f6f20dbd98371246d034e09960b94b

Branch libreoffice-5-2-1

This is 'libreoffice-5-2-1' - the stable branch for the 5.2.1 release.
Only very safe changes, reviewed by three people are allowed.

If you want to commit more complicated fix for the next 5.2.x release,
please use the 'libreoffice-5-2' branch.

If you want to build something cool, unstable, and risky, use master.

Change-Id: I73774938c95bbb11d9bcdd02ae72ddb5563ecff5

tdf#101385: don't try to get SharePoint folder outside of allowed range

for a given site.
This patch uses the fact that getFolderByServerRelativeUrl API returns
site's root folder if used with empty path ('').
Hopefully this isn't otherwise context-dependent, so that it doesn't
return other folders in some unknown circumstances.

More correct would be instead to use the site's root path directly,
like '/sites/mysite'. But I don't know a reliable way to filter the
API part from the base URI.

To make use of the corrected SharePoint repository root id, the patch
also redirects all cases where "/" object is queried, to session's
getRootFolder().

libcmis part of the patch will hopefully soon be merged upstream.

Change-Id: I0f3b9930748de3fe4763eb27721739f097b9892d
Reviewed-on: https://gerrit.libreoffice.org/27987
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: David Tardon <dtardon@redhat.com>
(cherry picked from commit 1b482aec4e7b6aec86fb20bcda555c4060a80332)
Reviewed-on: https://gerrit.libreoffice.org/28011
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>

Resolves: tdf#101309 crash on deleted SdrObject

when we're not recording undo, then SdrUndoDelObj just deletes
the argument its passed. And this code assumes that it is transferred
to the SdrUndoDelObj and it still exists.

Use the same pattern as ScPostIt::RemoveCaption for this situation

Invalid read of size 8
   at 0xB1713B6: SdrObject::GetOrdNum() const (svdobj.cxx:777)
   by 0x3C9E029C: ScDetectiveFunc::DeleteArrowsAt(short, int, bool) (detfunc.cxx:695)
   by 0x3C9E1D4C: ScDetectiveFunc::FindSuccLevel(short, int, short, int, unsigned short, unsigned short) (detfunc.cxx:1118)
   by 0x3C9E2405: ScDetectiveFunc::DeleteSucc(short, int) (detfunc.cxx:1207)
   by 0x3CF812BA: ScDocFunc::DetectiveRefresh(bool) (docfunc.cxx:480)
   by 0x3CFC9FAF: ScDocShell::DoHardRecalc(bool) (docsh4.cxx:1250)

 Address 0x50992d50 is 112 bytes inside a block of size 464 free'd
   at 0x4C2D22A: operator delete(void*) (vg_replace_malloc.c:576)
   by 0xB1D04BC: SdrPathObj::~SdrPathObj() (svdopath.cxx:1681)
   by 0xB1701F5: SdrObject::Free(SdrObject*&) (svdobj.cxx:394)
   by 0xB258666: SdrUndoObjList::~SdrUndoObjList() (svdundo.cxx:720)
   by 0xB258A39: SdrUndoRemoveObj::~SdrUndoRemoveObj() (svdundo.cxx:774)
   by 0xB25E29B: SdrUndoDelObj::~SdrUndoDelObj() (svdundo.hxx:298)
   by 0xB25E2B7: SdrUndoDelObj::~SdrUndoDelObj() (svdundo.hxx:298)
   by 0x3C7E5D0E: ScDrawLayer::AddCalcUndo(SdrUndoAction*) (drwlayer.cxx:1120)
   by 0x3C9E0238: ScDetectiveFunc::DeleteArrowsAt(short, int, bool) (detfunc.cxx:692)

Change-Id: Idc6d1f8e9ad8b203dac55630f8c100e74d3e017b
(cherry picked from commit d79a4f46926e81a6d0fe353feb3322673c1b509d)
Reviewed-on: https://gerrit.libreoffice.org/28009
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>

update credits

Change-Id: I004192541467afc5d8f31fe4c3c5f466b7b3dd17
(cherry picked from commit b17b4da3a713b935e7eb8cf23978689800d10f7f)

update emoji autocorrect files from po-files

(cherry picked from commit a363ac95bda1a4c3837bcd9dd29670bf0ab1853a)

Change-Id: Ie8af6c3ee0f655387e12967632f9bd31ace204ff

TANGO: Add show track changes bar small icon

Change-Id: Ia80da0b1ef394d19073e1f18ee22590010bb0257
Reviewed-on: https://gerrit.libreoffice.org/28002
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Yousuf Philips <philipz85@hotmail.com>
(cherry picked from commit ae3a78fa5f1a08a9bbfe233a2d6244a4d1fef254)
Reviewed-on: https://gerrit.libreoffice.org/28010
Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>

Updated core
Project: translations  c00038573d277bc17dfb3515b13b307a126e470c

update translations for 5.2.1 rc1

and force-fix errors using pocheck

Change-Id: I49f23c37a0f656727ba869e616e9d0b5679b9b15
(cherry picked from commit 0f9380734b89eb362e97e4e23bbf32d924b39132)

tdf#96607 'Save as' doesn't update global auto-recovery state

The auto-recovery service maintains a list of structures (one for each open
document) containing information needed to carry out the auto-save
functionality. One such piece of information is the location of the backup
file, stored in a struct member named 'OldTempURL'.  At every auto-save
interval, this list is iterated through and a function (implts_saveOneDoc)
is called during each iteration to save the current state of the associated
document.

The algorithm works as follows:
 1. A new backup file URL is chosen so as not to conflict with any already
    existing backup files in the backup directory.  This URL is based on the
    file name and incorporates a number (starting at 0) that is incremented
    until a name is chosen that doesn't conflict.

 2. The document is saved to this new backup file URL

 3. The previous backup file (indicated by its structure's 'OldTempURL') is
    deleted

 4. The new backup file URL is stored (in its structure's 'OldTempURL') for the
    next time the file needs to be saved.

Assuming you start with a new Writer doc and then make some changes, when it is
time to auto-save, the backup file name 'untitled_0.odt' (excluding path) will
be selected, the latest state of the open file will be written to that backup
file, and the full URL for the backup file will be saved into the struct
'OldTempURL' member.

The next time changes are made and an auto-save occurs, this algorithm will
result in the name 'untitled_1.odt' being selected, the file contents saved
into this new file, 'untitled_0.odt' being deleted, and the full URL for the
new backup file being saved in 'OldTempURL'.

The third time through results in 'untitled_0.odt' being selected (since this
file doesn't exist on disk), and subsequent iterations of auto-saving cause
the backup file name to alternate between the two aforementioned.

The problem occurs during a 'Save as' operation. When this happens, the backup
file is deleted (which is fine - it was just saved, and the next auto-save will
back it up) but 'OldTempURL' is not properly reset (see below for more info.)
During the next auto-save, 'untitled_0.odt' will be selected for the new backup
file name (since no file exists by this name), and one of two things will
happen (based on how many auto-saves have occurred):

 1. 'OldTempURL' points to 'untitled_1.odt', and the algorithm above continues
    to work correctly (at least in that it continues to backup file contents.)

 2. 'OldTempURL' points to 'untitled_0.odt', the name chosen for the new backup
    file.  In this case, the document contents will be saved to this file
    (step 2) but then the file will be deleted (step 3).  'OldTempURL' will
    maintain this URL from then on out, causing this case to be hit for all
    future auto-save intervals.

So, 50% of the time (30 minutes out of every hour) auto-save will stop backing
up file contents on a 'Save as'.

The function that handles the 'Save as' case (implts_markDocumentAsSaved)
clears 'OldTempURL' and sets other relavent struct members for a local variable
copy of the global struct, but doesn't copy them back. :(  These changes are
effectively lost when the function returns.

There are several other cases where this appears to be happening as well, but
more work is needed to determine whether this is actually the case:
 - implts_prepareSessionShutdown
 - implts_saveDocs, handling the 'dangerousDocs' and in a few other places
 - implts_openDocs
 - implts_resetHandleStates

Also, there is some JUnitTest code for auto-save, but it is currently disabled
(and fails to run successfully.) It'd be great to get these working again, or
to just write python equivalents. Implementing this would like take me a while,
though, so for now I just tested manually to ensure that this fixes the issue.

When I have some more time I'd like to work more on this, but I wanted to send
this patch in for now to address bug #96607.

This may also address bug #99890, since some of the struct members that don't
make it into the global state relate to the file name.  I haven't explicitly
tested this case, though.

Change-Id: Ic702d6f78e60c7cf828a1564ccca118dd45d152b
Reviewed-on: https://gerrit.libreoffice.org/25948
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: jan iversen <jani@documentfoundation.org>
Reviewed-on: https://gerrit.libreoffice.org/27919
Tested-by: jan iversen <jani@documentfoundation.org>
Reviewed-by: Aron Budea <baron@caesar.elte.hu>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

do not resolve system locale when queried, rhbz#1364406 related

When loading older documents that calculated a number format on the fly an
inherited but default format of a type could had been applied using the fixed
resolved locale instead of the default system locale, which then was stored
upon save and remained sticky. This because a format the formula depends on
already was applied using the resolved system locale.

http://bugs.documentfoundation.org/attachment.cgi?id=78559 of tdf#63267
exhibits the behavior on the hidden sheet 'Festwerte' when unprotected and
inspecting number formats in column A under Datum.

Change-Id: If23908f259458e988c5164cc5e268bfc9a6a6bcd
(cherry picked from commit 0d386267458b881f15e555186b52f7e2517ceca1)
Reviewed-on: https://gerrit.libreoffice.org/28004
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Resolves: rhbz#1364406 inherit the actual format index also for date and time

So summing [HH]:MM cells or calculating with dates uses the same format
in the result, not just the default format of a type.

This also fixes the apparently broken state stored by 4.4 (and earlier,
later?) where no type information was stored with the formula cell,
which may be just due to the old behavior of not applying the actual
format but determining it on the fly instead.

Change-Id: I14d0a7d07185bf5c77e0d7f6989a4a1d1a468d27
(cherry picked from commit f2e3de4dfcf10f9a59f8fc3f051c620fd50ef3c2)
Reviewed-on: https://gerrit.libreoffice.org/27916
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Resolves: tdf#100582 SetMatColsRows() when constructing matrix ScFormulaCell

... from an ScFormulaCellGroup token array, because ScFormulaResult that holds
the matrix dimensions is not cloned in that case as we don't clone from an
ScFormulaCell.

Change-Id: I13ab1b29db71ae1618580de995fe12ec423d4dc7
(cherry picked from commit c82a81bbda104ef08dd9e18725a09475b2d65183)
Reviewed-on: https://gerrit.libreoffice.org/27808
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

OFFICE-3708: package: recognize correct SHA256 URL

ODF 1.2 uses an incorrect URL to refer to SHA256, add support for the
correct W3C URL on import but continue to export the incorrect URL for
now.

Change-Id: I3135bcf989070d20f85f14702db07595f304e706
(cherry picked from commit 1015d35f2362953f415804476037d4f162eb49b5)
Reviewed-on: https://gerrit.libreoffice.org/27988
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#101213 drop use of CAIRO_OPERATOR_DIFFERENCE

for tdf#99446 and rhbz#1283420 there is a hackaround which ended up in 5.1.5,
which is not in 5.1.4, for corrupt glyphs under X. I can still reproduce the
problem if I drop the CAIRO_OPERATOR_DIFFERENCE usage here with master and
gtk2.

This alternative hackaround to force a read of the underlying surface works
just as well (help->license information is the reproducer) but reportedly
solves the performance regression.

(cherry picked from commit 705d7597480b2307d7e4929ce9386d80ce2a0f16)

Change-Id: Ie3c5b07409537a1734226b4ce034620351297e25
Reviewed-on: https://gerrit.libreoffice.org/27983
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#101359 getBookmarksCount includes more than aTableBookmarks

aTableBookmarks is just "BOOKMARK"s while getBookmarksCount() includes two
extra types. So cache the result of getBookmarksCount when filling
aTableBookmarks to compare if the count from the time of filling
aTableBookmarks is unchanged.

Change-Id: I69fedab613f23e4e2b30498e4620a370d92272e0
(cherry picked from commit cdb708291b59ac89b43c24154f0edc77f237eadd)
Reviewed-on: https://gerrit.libreoffice.org/27992
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

xmloff: forms import: convert relative xlink:href to absolute

There is currently only one place in the forms directory that converts
relative hyperlinks to absolute on import, in
OURLReferenceImport::handleAttribute(), but there are other elements
that have xlink:href attribute as well, such as form:form.

The export of form:form xlink:href does convert absolute to relative,
in exportTargetLocationAttribute(), but if the model URL is actually
already relative it will be converted to absolute instead, oddly enough.
This leads to different absolute href attributes, depending on the
directory where the file is exported, as can be observed with e.g.
ooo95698-1.odt and fdo40634-2.odt.

Let's apply a big hammer and make all "href" attributes absolute on
import.

Change-Id: I39d05707f3a8a899c7bbde8d9c0e2bc006c39e12
(cherry picked from commit c49b87ac140f3f2c79c8211f38cd86118022bdce)
Reviewed-on: https://gerrit.libreoffice.org/27901
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

framework: stop adding silly empty accelerator/current.xml files

... to ODF packages.  Somehow this is even skipped sometimes, but it's
much easier to just turn off the silliness than find out why.

Change-Id: Iff509dfd8325fd517e6434bcb56edbd06a3c27f1
(cherry picked from commit d76e3abe130007086099c62c5b425aaef82dc944)
Reviewed-on: https://gerrit.libreoffice.org/27900
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#97120 printing controls uses wrong font and font sizes

regression apparently since

commit 825b3df7f1d987021ec4a08ff8e7ed78e5772c97
Date:   Thu Oct 22 19:03:01 2015 +0200

    tdf#94138 fix printing of edit form fields

revert the GetDrawPixelFont part of that so the font is pulled
from the control and not the device its printed to, this makes
tdf#97120 and tdf#97120 work properly again

then revert

commit 6c41727484a04ab89005ffb052937dae5d7dc223
Date:   Tue Dec 1 17:44:23 2015 +0100

    tdf#94138 Use correct fonts for multiline edit when printing

because that replicates the original GetDrawPixelFont behaviour
so its not needed after the other revert.

Then, to solve the original tdf#94138, in the edit StateChanged handler call
ApplySettings(*this); like FixedText::StateChanged does to merge in the
controlfont setting to the underlying OutputDevice of the control, which
presumably is what is then retrieved from GetDrawPixelFont

Change-Id: I992a0e2011ffce7748d39f7f2bc49fbf6b8eaa79
(cherry picked from commit 5a5db03acc605a02c76c0f2977079b3dcf48de22)
Reviewed-on: https://gerrit.libreoffice.org/27879
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

no range based for loop when container is modified, tdf#100870, tdf#101320

This crash has been responsible for around 20% of the 5.2.0.4 crashes
with backtraces for the modification of the container similar to:

    sxlo.dll!SfxWorkWindow::SetChildWindowVisible_Impl(unsigned long lId, bool bEnabled, unsigned short nMode) Line 1834    C++
     sfxlo.dll!SfxDispatcher::Update_Impl_(bool bUIActive, bool bIsMDIApp, bool bIsIPOwner, SfxWorkWindow * pTaskWin) Line 1452    C++
     sfxlo.dll!SfxDispatcher::Update_Impl(bool bForce) Line 1326    C++
     sfxlo.dll!SfxApplication::SetViewFrame_Impl(SfxViewFrame * pFrame) Line 372    C++
     sfxlo.dll!SfxViewFrame::SetViewFrame(SfxViewFrame * pFrame) Line 3138    C++
     sfxlo.dll!SfxViewFrame::MakeActive_Impl(bool bGrabFocus) Line 1697    C++
     sfxlo.dll!SfxFrameWindow_Impl::Notify(NotifyEvent & rNEvt) Line 111    C++
     vcllo.dll!vcl::Window::CompatNotify(NotifyEvent & rNEvt) Line 3747    C++
     vcllo.dll!vcl::Window::Notify(NotifyEvent & rNEvt) Line 198    C++
     vcllo.dll!DockingWindow::Notify(NotifyEvent & rNEvt) Line 633    C++
     vcllo.dll!vcl::Window::CompatNotify(NotifyEvent & rNEvt) Line 3747    C++
     vcllo.dll!vcl::Window::Notify(NotifyEvent & rNEvt) Line 198    C++
     vcllo.dll!DockingWindow::Notify(NotifyEvent & rNEvt) Line 633    C++
     sfxlo.dll!SfxDockingWindow::Notify(NotifyEvent & rEvt) Line 1655    C++
     vcllo.dll!vcl::Window::CompatNotify(NotifyEvent & rNEvt) Line 3747    C++
     vcllo.dll!vcl::Window::Notify(NotifyEvent & rNEvt) Line 198    C++
     vcllo.dll!vcl::Window::CompatNotify(NotifyEvent & rNEvt) Line 3747    C++
     vcllo.dll!vcl::Window::Notify(NotifyEvent & rNEvt) Line 198    C++
     vcllo.dll!vcl::Window::CompatNotify(NotifyEvent & rNEvt) Line 3747    C++
     vcllo.dll!vcl::Window::Notify(NotifyEvent & rNEvt) Line 198    C++
     vcllo.dll!Control::Notify(NotifyEvent & rNEvt) Line 274    C++
     vcllo.dll!vcl::Window::CompatNotify(NotifyEvent & rNEvt) Line 3747    C++
     vcllo.dll!vcl::Window::GetFocus() Line 1842    C++
     svtlo.dll!SvTreeListBox::GetFocus() Line 2096    C++
     swlo.dll!SwContentTree::GetFocus() Line 2571    C++
     vcllo.dll!vcl::Window::CompatGetFocus() Line 3716    C++
     vcllo.dll!vcl::Window::ImplGrabFocus(GetFocusFlags nFlags) Line 384    C++
     vcllo.dll!vcl::Window::GrabFocus() Line 3004    C++
     swlo.dll!SwNavigationPI::SwNavigationPI(SfxBindings * _pBindings, SfxChildWindowContext * pCw, vcl::Window * pParent) Line 835    C++
     swlo.dll!VclPtr<SwNavigationPI>::Create<SfxBindings * __ptr64 & __ptr64,SwNavigationChild * __ptr64 const,vcl::Window * __ptr64 & __ptr64>(SfxBinsfdings * & <arg_0>, SwNavigationChild * const && <arg_1>, vcl::Window * & <arg_2>) Line 131    C++
     swlo.dll!SwNavigationChild::SwNavigationChild(vcl::Window * pParent, unsigned short nId, SfxBindings * _pBindings, SfxChildWinInfo * __formal) Line 1306    C++
     swlo.dll!SwNavigationChild::CreateImpl(vcl::Window * pParent, SfxBindings * pBindings, SfxChildWinInfo * pInfo) Line 69    C++
     sfxlo.dll!SfxChildWindow::CreateContext(unsigned short nContextId, SfxBindings & rBindings) Line 457    C++
     sfxlo.dll!SfxWorkWindow::CreateChildWin_Impl(SfxChildWin_Impl * pCW, bool bSetFocus) Line 1436    C++
     sfxlo.dll!SfxWorkWindow::UpdateChildWindows_Impl() Line 1328    C++
     sfxlo.dll!SfxWorkWindow::UpdateObjectBars_Impl() Line 1283    C++
     sfxlo.dll!SfxFrameWorkWin_Impl::UpdateObjectBars_Impl() Line 1114    C++
     sfxlo.dll!SfxDispatcher::Update_Impl(bool bForce) Line 1329    C++
     sfxlo.dll!SfxBaseController::ConnectSfxFrame_Impl(const SfxBaseController::ConnectSfxFrame i_eConnect) Line 1326    C++
     sfxlo.dll!SfxBaseController::attachFrame(const com::sun::star::uno::Reference<com::sun::star::frame::XFrame> & xFrame) Line 584    C++
     sfxlo.dll!`anonymous namespace'::SfxFrameLoader_Impl::impl_createDocumentView(const com::sun::star::uno::Reference<com::sun::star::frame::XModel2> & i_rModel, const com::sun::star::uno::Reference<com::sun::star::frame::XFrame> & i_rFrame, const comphelper::NamedValueCollection & i_rViewFactoryArgs, const rtl::OUString & i_rViewName) Line 604    C++
     sfxlo.dll!`anonymous namespace'::SfxFrameLoader_Impl::load(const com::sun::star::uno::Sequence<com::sun::star::beans::PropertyValue> & rArgs, const com::sun::star::uno::Reference<com::sun::star::frame::XFrame> & _rTargetFrame) Line 721    C++
     fwklo.dll!framework::LoadEnv::impl_loadContent() Line 1122    C++
     fwklo.dll!framework::LoadEnv::startLoading() Line 379    C++
     fwklo.dll!framework::LoadDispatcher::impl_dispatch(const com::sun::star::util::URL & rURL, const com::sun::star::uno::Sequence<com::sun::star::beans::PropertyValue> & lArguments, const com::sun::star::uno::Reference<com::sun::star::frame::XDispatchResultListener> & xListener) Line 116    C++
     fwklo.dll!framework::LoadDispatcher::dispatch(const com::sun::star::util::URL & aURL, const com::sun::star::uno::Sequence<com::sun::star::beans::PropertyValue> & lArguments) Line 55    C++
     sfxlo.dll!RecentDocsView::ExecuteHdl_Impl(RecentDocsView * __formal, void * p) Line 335    C++
     sfxlo.dll!RecentDocsView::LinkStubExecuteHdl_Impl(void * instance, void * data) Line 327    C++
     vcllo.dll!Link<void * __ptr64,void>::Call(void * data) Line 84    C++
     vcllo.dll!ImplHandleUserEvent(ImplSVEvent * pSVEvent) Line 1949    C++
     vcllo.dll!ImplWindowFrameProc(vcl::Window * _pWindow, SalEvent nEvent, const void * pEvent) Line 2497    C++
     vcllo.dll!SalFrame::CallCallback(SalEvent nEvent, const void * pEvent) Line 282    C++
     vcllo.dll!ImplHandleUserEvent(HWND__ * hWnd, __int64 lParam) Line 4110    C++
     vcllo.dll!SalFrameWndProc(HWND__ * hWnd, unsigned int nMsg, unsigned __int64 wParam, __int64 lParam, int & rDef) Line 5774    C++
     vcllo.dll!SalFrameWndProcW(HWND__ * hWnd, unsigned int nMsg, unsigned __int64 wParam, __int64 lParam) Line 5907    C++
     [External Code]
     vcllo.dll!ImplSalDispatchMessage(tagMSG * pMsg) Line 572    C++
     vcllo.dll!ImplSalYield(bool bWait, bool bHandleAllCurrentEvents) Line 591    C++
     vcllo.dll!WinSalInstance::DoYield(bool bWait, bool bHandleAllCurrentEvents, unsigned __int64 nReleased) Line 654    C++
     vcllo.dll!ImplYield(bool i_bWait, bool i_bAllEvents, const unsigned __int64 nReleased) Line 510    C++
     vcllo.dll!Application::Yield() Line 556    C++
     vcllo.dll!Application::Execute() Line 472    C++
     sofficeapp.dll!desktop::Desktop::DoExecute() Line 1327    C++
     sofficeapp.dll!desktop::Desktop::Main() Line 1646    C++
     vcllo.dll!ImplSVMain() Line 177    C++
     vcllo.dll!SVMain() Line 216    C++
     sofficeapp.dll!soffice_main() Line 165    C++
     soffice.bin!sal_main() Line 48    C
     soffice.bin!main(int argc, char * * argv) Line 47    C
     soffice.bin!WinMain(void * _hinst, void * _dummy, char * _cmdline, int _nshow) Line 47    C
     [External Code]

Change-Id: I4f699952d144a9c276ce53604dc66418abc4baee
Reviewed-on: https://gerrit.libreoffice.org/27977
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
(cherry picked from commit b0f4291e5c0fe6d61d621e692cbc8bea4cb4c401)
Reviewed-on: https://gerrit.libreoffice.org/27985
Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>

fftester: guard against empty table manager stack

Change-Id: Ibb4cadb58aa17eacfc9741f7f13c780be347cd8b
(cherry picked from commit 145e2dfc4501cb3fa01a1771cb6829b5945df4a6)
Reviewed-on: https://gerrit.libreoffice.org/27937
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: David Tardon <dtardon@redhat.com>

tdf#101008 fix XOR rendering

Change-Id: Ied73966633e5ffd56faccea7ec1408bd83642b58
Reviewed-on: https://gerrit.libreoffice.org/27862
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
Reviewed-on: https://gerrit.libreoffice.org/27922
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

tdf#100915 draw antialiased line just for polygon outline

To get the anti-aliased polygon we draw a anti-aliased line around
every trapezoid. This works fine until we draw a transparent
polygon where the lines become visible because of blending. A much
better and faster way is to just draw the polygon outline with
anti-aliased lines. This is done with this commit.

Same fix as aeb0c407a620ea8c28903f61d9d53e6d9ae7c53a in master,
but the code differs in 5.2 from master so much that it is generally
a separate implementation.

Change-Id: I95f98cc930caa7138a59048af68d4015046334d4
Reviewed-on: https://gerrit.libreoffice.org/27923
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

GetInputHdl may return nullptr

As can be seen in the other places we need to check for nullptr.

See
http://crashreport.libreoffice.org/stats/crash_details/553b9ef5-c85f-4cda-b432-472501795634

Change-Id: Ic5cab914f403fcae7f4e1e6f0e66c6295f7f56dd
Reviewed-on: https://gerrit.libreoffice.org/27949
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
(cherry picked from commit 8b86768a9a72ce6ef326615cb7c9ad5c76172921)
Reviewed-on: https://gerrit.libreoffice.org/27979
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

don't try to memcpy from and to a nullptr

See e.g.
http://crashreport.libreoffice.org/stats/crash_details/2218a489-b64c-4193-a7d4-cd01c6a607cb

Change-Id: I6c123d3e9e5c6dbcb7756a686503904cbfd944a4
Reviewed-on: https://gerrit.libreoffice.org/27948
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
(cherry picked from commit 855df248c8a380e253a35feefec21ebc1dee1aca)
Reviewed-on: https://gerrit.libreoffice.org/27978
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

mapping ERROR.TYPE to ORG.OPENOFFICE.ERRORTYPE was always wrong

Change-Id: I02b941a1a884a984b39632cad36579130cd71d6c
(cherry picked from commit 3d70765218986abba8b6d7c8e3cadd83a62ee035)
Reviewed-on: https://gerrit.libreoffice.org/27852
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

tdf#75256 Sifr icon update

Change-Id: I6d08448056f42b80c9ba04a3b5d99a5679952c5e
Reviewed-on: https://gerrit.libreoffice.org/27846
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: jan iversen <jani@documentfoundation.org>
Tested-by: jan iversen <jani@documentfoundation.org>
Reviewed-by: Yousuf Philips <philipz85@hotmail.com>

Sifr: Add thesaurus duplicate to links.txt

Change-Id: I0cfecbef999bb251c1fd8829aed0a4f0268e176c
Reviewed-on: https://gerrit.libreoffice.org/27870
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Yousuf Philips <philipz85@hotmail.com>
(cherry picked from commit c63aaa624aa6e949e127fb8ef00c4f3744c72b76)
Reviewed-on: https://gerrit.libreoffice.org/27876
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

tdf#78506 RTF import: fix handling of invalid \levelnumbers

In case ';' is written in \u form in \levelnumbers, then Word ignores the whole
\levelnumbers contents, do the same.

(cherry picked from commit 428a1da60b88415e7db21353a42bed85b8b76ed9)

Change-Id: I93ce5810af2b5ed703e804199c0b236d2c4c36b5
Reviewed-on: https://gerrit.libreoffice.org/27873
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

tdf#101147 New approach to fix minute/month detection

Other formats may be differently interpreted
HH YY/MM => Month expected
HH DD/MM => Month expected

Change-Id: Iabb17911bf2907c02b6b898052b6933692f8074a
Reviewed-on: https://gerrit.libreoffice.org/27886
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>

break the Excel rule for YMD and DMY, tdf#101147 follow-up

Clearly in {HH YYYY-MM-DD} the MM should not be minute. Also not in
{HH DD.MM.YY}. Don't follow every bullshit. Period.

It is debatable how to treat MDY, {HH:MM DD/YY} should be different from
{HH MM/DD/YY}, Excel ironically takes both as minute, even in an en-US locale.

(cherry picked from commit 88134dcba680418496f7a1c70a47fde0159ce390)

 Conflicts:
svl/qa/unit/svl.cxx

Change-Id: I13d39a36294e3c40cc0e9bf72026804b299bb264
Reviewed-on: https://gerrit.libreoffice.org/27813
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Eike Rathke <erack@redhat.com>

Resolves: tdf#101269 cannot click/edit text inside a frame

regression from...

commit b0c0a074a2318a231f3f635784da494f1b713c53
Date:   Mon Apr 18 11:03:43 2016 +0200

    clang-tidy clang-analyzer-deadcode.DeadStores

    suspect that some of these may reveal latent bugs....

and restore method then removed by

commit 47f62540bd2c2f107313bb0c6f141cd4460b6379
Date:   Thu May 19 10:31:47 2016 +0200

    loplugin:unusedmethods in sw

Change-Id: I14826eecadbc1a74e408f8907eea8746debe4db4
(cherry picked from commit 24c009de4d6d27233027e9bcd86ea181d230be8f)

pPos can never be null

Change-Id: Ib7b57d628fbd80a1d2214917cd439e09d4803d91
(cherry picked from commit 2702796806cae2b28ae78e78d29c3559bf97607e)

SwLinePortion::GetCursorOfst is const and its ret is ignored here

so nOfst is unnecessary

Change-Id: I27f73057b55ebdf30eb69a19ee1b0647ba25583e
(cherry picked from commit 35f4248206e844cabd83757ec57ddde1c2ef66b4)

only called if !bDraw

Change-Id: I7310b5f8cbbc4176e76d92ce2bcd49168470843c
(cherry picked from commit cc7a55450658696009fe01e94632e80bb3785d5b)

the sole caller is ok with rPoint being modified, so change to non-const arg

Change-Id: I40a24502e2e8ad312e13e0c47ee1fde3118d2251
(cherry picked from commit 2b3112b30a8dea8a009c11ddb209e12697f72e32)
Reviewed-on: https://gerrit.libreoffice.org/27865
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

add more nullptr checks, related tdf#100820

See
http://crashreport.libreoffice.org/stats/crash_details/8569250c-b04a-43b0-b4b8-4b3a94daffe1

Change-Id: I9070be3db57e77befe70c0a32ef6aa54c6c1cf85
Reviewed-on: https://gerrit.libreoffice.org/27842
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
(cherry picked from commit 626a1aa960ed36f5c3370e2aab1e72d41df875cf)
Reviewed-on: https://gerrit.libreoffice.org/27859
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Resolves: tdf#97193 restore pre 5-1 values for INS_INSROWS and INS_INSCOLS

i.e. INS_INSCOLS_BEFORE and INS_INSROWS_BEFORE have the same
meaning as the old INS_INSROWS and INS_INSCOWS, so reorder the enum
so they are at the positions of their corresponding originals.

This makes macros recorded before 5-1 work like they did prior to 5-1-0.
This unfortunately will have the same effect to macros recorded during
the 5-1 period that the original change did.

regression since...

commit f97beeef352fdd9fd4f157a7a4c4f6af8341cbe6
Date:   Wed May 20 16:54:27 2015 +0200

    Calc: Insert row/col before/after [1/2]

    This part renames all INSROW / INSCOL to
    INSROWS_BEFORE / INSCOLS_BEFORE which is
    the current (default) behaviour.

Change-Id: Ide90b8cfebe4af1a3718d93c34657663c20ccc62
(cherry picked from commit 7ec4a495291c7427bba5256bf2c54f1da999ea52)
Reviewed-on: https://gerrit.libreoffice.org/27834
Reviewed-by: jan iversen <jani@documentfoundation.org>
Tested-by: jan iversen <jani@documentfoundation.org>
Reviewed-by: Eike Rathke <erack@redhat.com>

Resolves: tdf#96873 include template image to make it available in help

alternatively move it into the helpimg dir and change the usage
in helpcontent2, but do it this way to keep it together with the others

Change-Id: I84c6b10aed948a5fea835dc86bade5a4390c716b
(cherry picked from commit 61062608a27ceaef5424203d849f88c7fb77b8d0)
Reviewed-on: https://gerrit.libreoffice.org/27826
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#94940 don't load font history if option is disabled

Change-Id: I70348b4e4bb85850331044e901c03aabb0d22ef8
(cherry picked from commit 171586cc4df4a9470251e50ce51338c2d53acf9b)
Reviewed-on: https://gerrit.libreoffice.org/27657
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#100123 prefer notation priority in detecting range/address

i.e. try and parse whether something is a range or an address in order of
document address convention, Calc A1, Excel A1 and Excel R1C1, rather than
check if something is a range in any of those conventions before checking if it
might be an address in any of those conventions.

Reviewed-on: https://gerrit.libreoffice.org/27821
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
 Conflicts:
sc/source/ui/view/tabvwsh3.cxx

Change-Id: Ibb744c3eda78a80f33bdbfa5f5ddf0aa5b6361af
Reviewed-on: https://gerrit.libreoffice.org/27824
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>

Use "dot" for "source"ing shell files

Quote from stackoverflow: "source" is non-standard and specifically
not supported in ash, dash or pdksh... it's a Bashism.

Change-Id: I4b47385c1e44c5e56bd84f23ff3645712015dcf4
Reviewed-on: https://gerrit.libreoffice.org/27746
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
Tested-by: Jan-Marek Glogowski <glogow@fbihome.de>
(cherry picked from commit 3e241399476ca3c2e3775f485356197bb93863a9)
Reviewed-on: https://gerrit.libreoffice.org/27751
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

Resolves: tdf#101196 line wrap very long checkbox content

Change-Id: Ie451b88240f755a0d64dd3db14d7981b8ed2a2a2
(cherry picked from commit 1c33b3f5f693d432f74c795a87f87bd487b09d47)
Reviewed-on: https://gerrit.libreoffice.org/27811
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

fftester: null DestinationText

Change-Id: Icdd71733c4e9f4b36e6e957e4dea772087890faa
(cherry picked from commit 0dd22165346f4296718656d17ae98ebbfe563719)
Reviewed-on: https://gerrit.libreoffice.org/27666
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

masses of MessBoxes not being disposed promptly

since...

commit ba81e5c6bd420b41a84ade6ccd774011a8089f7f
Date:   Thu May 28 21:35:43 2015 +0100

    tdf#91702 - fix stack-based MessBox allocation.

There is no special ScopedVclPtr<X>::Create or
ScopedVclPtrInstance<X>::Create just
VclPtr<X>::Create and a raw VclPtr<X>::Create()->foo
doesn't call dispose on the owned X

(cherry picked from commit a860df25dd7bf62ecb6b3d3ed38803b981f56d52)

Change-Id: Ifacc8d5e742820701307c3c37b9b86487667d84f
Reviewed-on: https://gerrit.libreoffice.org/27537
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Fix the WebDAV lock owner name shown on file locked dialog UI

Probably due to 8d411a4a1ef6844c00bc714f8b144d3729e4f4e8.

See comment in the added code for details.

Change-Id: I3e8a856a715385ed2ab1d9ecda31dd25e05e6c59
Reviewed-on: https://gerrit.libreoffice.org/27716
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Giuseppe Castagno <giuseppe.castagno@acca-esse.eu>
(cherry picked from commit 97353e5c43815b45a6f365ccaea474899683d38e)
Reviewed-on: https://gerrit.libreoffice.org/27728
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#57978 don't show duplicate index keys in dropdown list

Change-Id: I25e25157063ea4424f770b397d879cb7bfd46bcb
(cherry picked from commit c755336853f20c7b2dd14e9c56eeb441d831be2c)
Reviewed-on: https://gerrit.libreoffice.org/27576
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

fftester: guard against no drawing object property set

Change-Id: I51736459f9f098d9d793bff3b9a1a403962e099d
(cherry picked from commit db686815b41c52598f0952613ff8c6be34b2f0e0)
Reviewed-on: https://gerrit.libreoffice.org/27800
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

don't catch SEH exceptions in sal thread code

...just for libreoffice-5-2, which does not have
62c047ffb397802c09df9070492e70725928cadf "switch to EHs on windows" and
needs this for crash reporting in spawned salhelper::Threads.

Change-Id: I1e8f15d8f2fb5a741538d16de34c47f66dfaba0d
Reviewed-on: https://gerrit.libreoffice.org/26968
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

tdf#95251 MM just allow print as single document

Remove kept block from revert of
  138d29aa09417eba4d15ade4c9f4dab2620b6326

After this revert MM just supports pinting via a single merged
document. Printing as inidividual files should have still worked,
but as a workaround we remove this functionality, until the print
backend can handle multiple-file print jobs correctly.

Reviewed-on: https://gerrit.libreoffice.org/27721
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
(cherry picked from commit b6a698b093f78dc90836d502db3d0276f335fbfe)

Conflicts:
sw/source/uibase/dbui/dbmgr.cxx

Change-Id: I70d0852cdcc369b0fc9f344086830f15d8dea451
Reviewed-on: https://gerrit.libreoffice.org/27722
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#96677 effects page didn't link labels to what they labelled

(cherry picked from commit b3b7669e7fb74b04d925f21a7f0b048434eeffa3)

Change-Id: I29503324f221c42208267f4f0d36922e89119fdd
Reviewed-on: https://gerrit.libreoffice.org/27677
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#84121 wrong password passed to xpdfimport

regression from...

commit eb276d11a652601f743d69ada7a38016bd97d73e
Author: Fridrich Štrba <fridrich.strba@bluewin.ch>
Date:   Mon Jul 8 14:01:17 2013 +0200

    fdo#38878: Upgrading internal xpdf -> poppler

before this the hashed _OOO_pdfi_Credentials_ was used
in the removed OOoImportSecurityhandler::authorize

but after this the hashed password continued to be
passed through stdin, while the unhashed password
was the required data.

Change-Id: Iadc8a372abcfab514f724dd95f05776e5153d007
(cherry picked from commit 2a66c02553bbcfbb9e646c39ec291fb2eefbf6bb)
Reviewed-on: https://gerrit.libreoffice.org/27673
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#101246 errors are not strings

Change-Id: I5a63dcc0452f0254e66e0ad743ae0d0926dfc14a
(cherry picked from commit 8e495518c36359e565d6c172688e7120c477137f)
Reviewed-on: https://gerrit.libreoffice.org/27775
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

limit SelectionFillDOOM to 24117248 cells, tdf#60021 tdf#60056 related

Change-Id: Idabf5e51b7f423d7d58094ad1caef166728c3bed
(cherry picked from commit 5cf5975cef114870268bee792e44570ddfdaafe8)
Reviewed-on: https://gerrit.libreoffice.org/27776
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Related: tdf#92516 make menu accelerators always work even if...

keyboard not used yet.

The gtk2/3 menus appear to work this way. (And when not in gtk2/3
it shouldn't disable the accelerators for other platforms anyway)

(cherry picked from commit f6d0d92a5e249413df99a7f3677a28efe6d45426)

Change-Id: Ib7a99bd9039cd07120b3b77380f810b5b028fd57
Reviewed-on: https://gerrit.libreoffice.org/27794
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

fftester: another table manager stack checks

and a m_xTextFactory check too

Change-Id: I9352410c42048b4dd7d6dbc3514351ab8f16790b
(cherry picked from commit 8a6b2fb5b94de43316ab3ea95ff07cf5f46b6134)
Reviewed-on: https://gerrit.libreoffice.org/27759
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

add missing nullptr check

See
http://crashreport.libreoffice.org/stats/crash_details/173c2c82-7241-44b4-8394-2b5635f7f938

Change-Id: I2a03c6d6a77e3c872b8507c72b47a05db20fd564
Reviewed-on: https://gerrit.libreoffice.org/27791
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
(cherry picked from commit dadd4abdadb970dd4912caaa56b7250c3a30cc9e)
Reviewed-on: https://gerrit.libreoffice.org/27799
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Resolves: tdf#101241 Revert "fix paint table border in DOUBLE_THIN style"

because of the never-ending invalidate and paint of tdf#101241

there was no bug or route to reproduce the original problem so can't
try to solve it a different way.

This reverts commit 9a95520948de1d92c55252aa5f90606b01d6ac9e.

(cherry picked from commit 772191616a0cceebf91a9674ce7417d5cad0050f)

Change-Id: Ib3c0421247a20f047d396b13ec17dbfe27cb780c
Reviewed-on: https://gerrit.libreoffice.org/27789
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

tdf#101147 Improve (again) minute/month detection

For format like HH:MM:SS MM/DD
second MM is detected as minute: see tdf#95339

New rules:
- first M following each H is minute
- first M following first S is minute

Detection Month/minute now fully compatible with Excel even with unwanted detection:
SS:MM:HH DD/MM/YY
second MM is minute even if user would expect month

Reviewed-on: https://gerrit.libreoffice.org/27560
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
(cherry picked from commit aa6bca0dfa0a58aae91eb5fa3564add642cd4238)

 Conflicts:
svl/qa/unit/svl.cxx
svl/source/numbers/zforscan.cxx

Change-Id: Ia789fbc2dbd6d2dcbe9b9c34d3f288674966dd10
Reviewed-on: https://gerrit.libreoffice.org/27768
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>

Padding adjustment to Custom Animations Panel

Change-Id: Ib09d138d2f8e82d08ce1f2082a25690ce01ac7cc
Reviewed-on: https://gerrit.libreoffice.org/27752
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Yousuf Philips <philipz85@hotmail.com>
(cherry picked from commit 4244236d1bfc6426b7d81f0a9d13e473c091b345)
Reviewed-on: https://gerrit.libreoffice.org/27763
Tested-by: Yousuf Philips <philipz85@hotmail.com>

Add border padding to Slide Background Panel

Change-Id: I6b0ae2e65cba6768f566dfffb3e54181987d5757
Reviewed-on: https://gerrit.libreoffice.org/27749
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Yousuf Philips <philipz85@hotmail.com>
(cherry picked from commit 82111a5f8206ce05fc762db14f4b066ce23d6f07)
Reviewed-on: https://gerrit.libreoffice.org/27762
Tested-by: Yousuf Philips <philipz85@hotmail.com>

add missing nullptr check

See
http://crashreport.libreoffice.org/stats/crash_details/07407baf-17dc-4c39-8020-4975797d89cc

Change-Id: Ia10cf6555869a402bd16fbb5a272b1ccb03b12fb
Reviewed-on: https://gerrit.libreoffice.org/27737
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

VclPtr: password prompts never dispose

Change-Id: I917e5457be6de152b39e91ad41b6ed196dfe3e45
(cherry picked from commit 4b73aff7d2e4af7d3a1359df52133cd2c59058c5)
Reviewed-on: https://gerrit.libreoffice.org/27675
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

fftester: another empty table manager stack

Change-Id: If3148cb6e16cff4aad28c4f86467c66ed04bcd05
(cherry picked from commit d00fc303bca7765762a602531b7d3b40ce8f1740)
Reviewed-on: https://gerrit.libreoffice.org/27703
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

fftester: context stack check

(cherry picked from commit 2acc9d06291b3f271c2a88bb4995d52d2d8d2152)

Change-Id: I4a135a9f9ac2f16f9dab096f5c234ff1d6e5e853
Reviewed-on: https://gerrit.libreoffice.org/27730
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

Resolves: tdf#60056 disallow Fill and Random when entire sheet is selected

This is a combination of 3 commits.

move range detection to ScViewData::SelectionFillDOOM()

(cherry picked from commit 4b90c9dc83d96908d3732d95e75c2f71d0f5d988)

Resolves: tdf#60056 disallow Fill when entire sheet is selected

(cherry picked from commit c729ee7622b1d54b2dc82b1807c68899efeab6d7)

disallow Random Number when entire sheet is selected, tdf#60056 related

(cherry picked from commit 77327759c9053da493430ee01b4d7fe98c174574)

866f84d72c27be962dbad8f8e4dcb345aa336bff
e9e229c3d3e5a5a7e3f3e90ebd181471927bd452

Change-Id: Ic5205428136fc323affed8d956d6ee0434d43988
Reviewed-on: https://gerrit.libreoffice.org/27690
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

missing nullptr check

See
http://crashreport.libreoffice.org/stats/signature/ScPreviewObj::getSelectedSheets%28%29

Change-Id: I210426b8ac23fdf6b1195afe7dbc388802335e71
Reviewed-on: https://gerrit.libreoffice.org/27711
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
(cherry picked from commit 71835fd8b5b04651e5be38b3b26a8dc80f5c3355)
Reviewed-on: https://gerrit.libreoffice.org/27732
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

tdf#97831 follow up : remove unused arguments from stack

Intention:
Make e.g. [1+SWITCH(1;1;1;6)] work. Without this fix, the result is
7 but should be 2.

Reviewed-on: https://gerrit.libreoffice.org/27664
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
(cherry picked from commit 3ee2cae3a36e5d052e02c3f7b226dc3b79bd2ebc)

 Conflicts:
sc/qa/unit/data/functions/fods/Functions_Excel_2016.fods

Change-Id: I2a72e458e5458c02c3ce2f4214b86ba13f7b1d11
Reviewed-on: https://gerrit.libreoffice.org/27697
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>

update emoji autocorrect files from po-files

Change-Id: I13b36b9280245db81bf68391765fa2d5c1261155

fftester: throw on empty stack access

rather than spend the rest of my life protecting each one

Change-Id: I181df33b052a0303f072ce0252d98562231569e2
(cherry picked from commit 2710211eb2333cafdb894742a8fa73fb02dc513b)

fftester: empty states stack

Change-Id: I05dfffced9a8677650a46b43f65a29e9b21c5524
(cherry picked from commit 5de2d02806669812d43e7f23db58ab7a16373ce6)

Related: tdf#75757 remove inheritance from std::deque

Change-Id: Ia50ea146052c2014ea16474186e2d15ce93581c1
(cherry picked from commit 7a887df4db129ac5222fd4068173b5a06d107a59)
Reviewed-on: https://gerrit.libreoffice.org/27639
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

Updated core
Project: translations  b2f028f1d2f7779c25d2de039cdc42f8590f6ddf

update translations for 5.2.0 rc4

and force-fix errors using pocheck

Change-Id: If338de1b28c0e26c2a9069a35de02bb21fe80d06

do not force the Paste command to enabled for Calc, tdf#60021 related

... so it actually gets disabled in menu or removed from context menu if we say
so. Used with disallowing Paste for Ctrl+A selection or protected cells.

Change-Id: I7d5f8bdc7b7adee65d08b93ddee015aef953d95d
(cherry picked from commit fe46041a6b2fd29715389cc5eb2dfbaf65f7f1f9)
Reviewed-on: https://gerrit.libreoffice.org/27648
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

Resolves: tdf#60021 disallow Paste when entire sheet is selected

... which exhausts memory unless you have 100GB or more of free RAM.

Change-Id: Ie6f02c48457f80acad33d2286194765f8343f2fb
(cherry picked from commit 984b0d1599ff1672cb0d28019bd652d58d6bdefa)
Reviewed-on: https://gerrit.libreoffice.org/27647
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

tdf#99434 docx enable docprotection only when enforced

Change-Id: I9454a34d7b386acffc50e74d5ef6eed1966d572a
Reviewed-on: https://gerrit.libreoffice.org/27456
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Justin Luth <justin_luth@sil.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
(cherry picked from commit d40ed9d0cd3c4a66d4ebf65d3a530f0fc5a8da17)
Reviewed-on: https://gerrit.libreoffice.org/27625

fftester: empty table manager stack

Change-Id: Ia7f7ace8130d5dfe290207e0cd3d2e6a43b8ab46
(cherry picked from commit df9414084b46c1712dc7151f50023438c62301e2)
Reviewed-on: https://gerrit.libreoffice.org/27631
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#101132 a11y crash in hyperlink

regression from...

commit aa6c54f337a616c9ca6a39cfc8708657018d8613
Date:   Mon Jan 18 09:01:15 2016 +0200

    loplugin:unusedfields

Change-Id: Ic2b7fe5f2d49a447a9936c4fb04bafcc7349ecf9
(cherry picked from commit 4fc3c8a3df485f6dccdcb2c51c6266fbd0dace3e)
Reviewed-on: https://gerrit.libreoffice.org/27609
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Resolves: tdf#63662 'Manual Column Break' always visible...

even when non-printing chars is off.

I think this began as an intermediate step for both page and
column breaks to the header/footer overlay we now have for
page breaks. Leaving this just for column breaks. It may
have been set to always on with page-breaks in mind.

Change-Id: I1617683e2bb8c1f9229e9c2313e7817e184fa5ea
(cherry picked from commit 9e63e60d6f00ee690fafb9f21f2cafb08a6ad92e)
Reviewed-on: https://gerrit.libreoffice.org/27579
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Updated core
Project: dictionaries  6417dbd05e13c2e1165751feb2bc7866328dec2d

cut size of README_en_GB.txt

400 KB -> 4 KB
list of changes is now factored out (official change)

Change-Id: Ib1a1f2701015c73bc580766e0b15a974f87ce873
Reviewed-on: https://gerrit.libreoffice.org/27621
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
(cherry picked from commit 3380b51b42480693d8b3ac97f08c6568a4f73201)
Reviewed-on: https://gerrit.libreoffice.org/27627

Drop OpenCL details from the 'Detailed Calculation Settings' dialog

There is really no need for end-users to have access to these
settings. We don't want end-users to shoot themselves in the foot. If
power users really think they need to modify these settings, they can
poke the LibreOffice registry.

tdf#101063: Get rid of the whole "Test OpenCL" functionality while at it

Its intended use case has never been particularly clear. (I wrote it,
so I know.) We now test OpenCL usability silently at each from-scratch
start.

(cherry picked from commit a92da660cd4aae511785da4f3a276309c64b47f7)
(cherry picked from commit d78012872a15cdd0347cbabfc13e03ce5511bf90)
(cherry picked from commit 70258bb2d637319d8ff1d19bf304f813c5b2e948)
(cherry picked from commit e4781d09418fe1dc7219f6ae11f516da622329aa)

Change-Id: Idb800ea7e35d09fd1d72da83e93c2c2b6a025115
Reviewed-on: https://gerrit.libreoffice.org/27617
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

WaE: C4100: 'pHelper' : unreferenced formal parameter

(loplugin:staticcall)

(cherry picked from commit 3849758adc0592327568ff465ccf255786632b4b)

Change-Id: I800eef0517f063ff7e08a95de9da268fb0e9d621
Reviewed-on: https://gerrit.libreoffice.org/27618
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Tor Lillqvist <tml@collabora.com>
Tested-by: Tor Lillqvist <tml@collabora.com>

handle notes on empty cells, display search results, tdf#65334 follow-up

This is a combination of 6 commits.

include empty cells if notes are searched, tdf#65334 follow-up

(cherry picked from commit f027c77c520adbdf8cec59e0484fc87b33cf203b)

use GetCellArea() when searching for notes, tdf#65334 follow-up

... so empty cells with notes are included as last "data" position if they are
below or right of the last "real" data, which GetLastDataPos() doesn't.

(cherry picked from commit 09cc958dee93ad0ad2ab0d8cc9cc4c09e46c3653)

do not search in empty string if there is no note, tdf#65334 follow-up

(cherry picked from commit a02f7aa735c52f5d20df0e2a94cc06879cb3dfac)

use GetCellArea() also in SearchAll() and ReplaceAll(), tdf#65334 follow-up

(cherry picked from commit 156590e7c0bf3c81f50b3d9862961e50010ea08a)

use HasCellNotes() when looping and searching for notes, tdf#65334 follow-up

... so we actually find all if on empty cells.

(cherry picked from commit d93284f65cf14fdf5e2520b474eb7383ca1d99e4)

display the SearchAll() and ReplaceAll() results for notes, tdf#65334 related

(cherry picked from commit 4f719263ae8dc44eabfba4654f9dbed92a9c5360)

9bb7464033736e7e7fa24e635ef1a3d39626002b
6c347b5bccd4ffd6a43a80e9d73a0b5fcf82926b
50bfeee96e57010455bbb9e8be4312a39d392411
e504ce7df2a17877eddf21cf309990a2a8b692ba
b9ff40b26526efdf242db2ef1804e54611f16b0e

Change-Id: Ib67b9745ffff730b209df2b82dc7e46fc4774900
Reviewed-on: https://gerrit.libreoffice.org/27603
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Resolves: tdf#65334 cannot search in cell notes

regression since...

commit c06dbbe7594c2a0b5a5b19f8e183d9c421e6e094
Author: Markus Mohrhard <markus.mohrhard@googlemail.com>
Date:   Thu Feb 23 23:36:49 2012 +0100

    remove mpNote from ScBaseCell

and requires restoring some code subsequently removed in

commit a1a1131c16227327d17bcf5ad2a5209ddc0f01a1
Date:   Sun Mar 4 15:27:27 2012 +0800

    Remove unused code

Change-Id: If931c3162cacaeb126e730e3e79c1c47be4ce434
(cherry picked from commit 7c84a219e3cc8d623826343f547ba8adacca6d56)
Reviewed-on: https://gerrit.libreoffice.org/27577
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins <ci@libreoffice.org>

Resolves: tdf#100046 the dreaded 1945-04-02 Double DST, also tdf#79663

This is a combination of 5 commits.

two more testDateInput() dates

Found in comments of i18npool/source/calendar/calendar_gregorian.cxx
Calendar_gregorian::setValue()

(cherry picked from commit 17fb4d43e86bd9622fb1b2c4378ab1e63a0025f0)

display data on test failure

(cherry picked from commit 4097b95d62db14f738f69e3bb4d1162739f3a06a)

add test data from tdf#79663, tdf#100046 related

... before changing anything.

(cherry picked from commit 22dceea0ca3cd68463e390395aa8de479b680240)

Resolves: tdf#100046 the dreaded 1945-04-02 Double DST, also tdf#79663

A coward not having changed this earlier, partially taken from a patch that was
submitted for tdf#79663 but broke API, thanks to Isamu Mogi again.

This removes all the clutter around timezone and DST correction that was
necessary for early ICU versions, which apparently gets things right now and
the clutter can still be confused by corner cases.

(cherry picked from commit 8e91b0177cbf6554acf7ccfbcc573f647e7fcf4d)

 Conflicts:
i18npool/inc/calendar_gregorian.hxx
i18npool/source/calendar/calendar_gregorian.cxx

activate the remaining 1945-04-02 test cases, tdf#100046

(cherry picked from commit 157b56a049a9a5f868d1f3b9856a631a40043a14)

432ba81929dd0badd48d2bd7b4344ceed4df2b43
edf89713de1eb02e501723f679bf059b5a43415e
9a90f933e8db8c6e0db145520ebf71cc27621abc
4d2f3d5602b162924c14348c0307a74ddeb93079

Change-Id: I4ab0aae15b7d97b22bc3a16d039542b5b49d3f14
Reviewed-on: https://gerrit.libreoffice.org/27559
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

update credits

Change-Id: I9679f7b0693ecc380d756b18025b09225afd414c
(cherry picked from commit ce3577056dd44d35b38894b7ec573951208209eb)

tdf#101138 opengl: blacklist intel drivers for Win 7

Change-Id: I1b3db15b2fbdd948dcc9bacf7891f8429f066150
Reviewed-on: https://gerrit.libreoffice.org/27549
Reviewed-by: Jan Holesovsky <kendy@collabora.com>
Tested-by: Jenkins <ci@libreoffice.org>

No need for own implementation of erfc() in OpenCL

The own code was copied from the C++ one we used to have in
sal/rtl/math.cxx but which was removed in
a62bc6a65abb47adb0e4caff7e38823c15b302fc. However, it did not work
correctly on some machines at least, like my AMD A10-7800 running
Windows 10. I was unable to figure out why not. This lead to OpenCL
being disabled by the Desktop::CheckOpenCLCompute() code we now run
early on startup.

Anyway, as OpenCL has erfc(), just use that.

Change-Id: I7ba6104fc4975cd570358760fa97a19390a54cce
(cherry picked from commit 4afa88f289de1150850b52d36f2345fd9a9fbc1e)
Reviewed-on: https://gerrit.libreoffice.org/27540
Reviewed-by: Tor Lillqvist <tml@collabora.com>
Tested-by: Tor Lillqvist <tml@collabora.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Resolves: tdf#101076 drag-create creates an unfilled shape...

after using the slide properties->background->all slides

This SetAttributes with nothing selected will end up setting the
default object properties. Here I think we are setting the bg color to
the master, and then unsetting the bg color of the slides to none
so the master color shows through.

So the additional SetAttributes call unsets the bg of the default
object style

Change-Id: I5cfd74eef5637b87d2a8778d4b8c8a0cbc0960db
Reviewed-on: https://gerrit.libreoffice.org/27487
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
(cherry picked from commit 60ab08c6c5c4b5730c1cb54eca5558be214e8925)
Reviewed-on: https://gerrit.libreoffice.org/27505
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>

Related: tdf#101057 inherit MalformedVersionException from std::exception

so we don't get std::terminate on MacOSX if its thrown up to
JavaVirtualMachine::getJavaVM

Change-Id: I1c6b79bf16cbaa66b8f79d495239e301d3ca58de
(cherry picked from commit a3b3c5fcd7efd3e33a0f0078bd6fd23689c45c41)
Reviewed-on: https://gerrit.libreoffice.org/27508
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>