Change-Id: Ia8097b6d92c9bc4e3841d96bcb8cd553969f54a0

EscherPropertyContainer may allocate a 1-byte buffer if there are no
vertices / segments. (happens in sw_filters_test)

 in EscherPropertyContainer::CreateCustomShapeProperties(MSO_SPT, com::sun::star::uno::Reference<com::sun::star::drawing::XShape> const&) filter/source/msfilter/escherex.cxx:3853:25

Change-Id: I56dedbd57d38cd017183060e924cb1340ea58ebf

Change-Id: I5e92ca9dbb0c9f33c9511cc4b12db1433e40fa68

Change-Id: I62ff8a172a9c4f361ef5fcb6d60bd9fd2e959c2d

hopefully..

C:/cygwin/home/tinderbox/master/svl/source/items/srchitem.cxx(111) : error C2872: 'Locale' : ambiguous symbol
could be 'C:\cygwin\home\tinderbox\master\workdir\UnoApiHeadersTarget\udkapi\normal\com/sun/star/lang/Locale.hdl(17) :  com::sun::star::lang::Locale'
or       'C:\cygwin\home\tinderbox\master\workdir\UnpackedTarball\icu\source\common\unicode/unistr.h(79) : icu_54::Locale'

Change-Id: I9d9650cd058fb4328fcaa2ae8f5adbb903001817

Those attributes contain a lot of interesting properties, among them
the heading level which fixes this bug.

Change-Id: I634ef404123cb1b2831563b9b186db4b7e9e7d8b

So move it there. While at it, make it use SAL_WARN() instead of
printf.

Also, add a few more SAL_WARN() and SAL_INFO().

Change-Id: Ib058fb20d07757331ca364a8d7649abc59e9494b

Change-Id: I9c31d91dab2bc47148c0439947c331e411159891

(and no need to hold css::uno::Sequence objects by pointer either).  Regression
from e537d227 "hard to find accidental leaks
with all the deliberate ones" caused JunitTest_svx_unoapi to sometimes crash.

Change-Id: I3579c4bbac58d7e341c8acb52dd3f0f06f7b2245

Change-Id: I58fd91354fe41673216674f5ae9c11c17cb3f29d

Limit duplication for SfxRequest FN_NUMBERING_OUTLINE_DLG in sw shells

Change-Id: Ibbf58480232f3b393d13548508b7cdbc98690847
Reviewed-on: https://gerrit.libreoffice.org/14313Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>

Limit duplication for SfxRequest FN_FORMAT_FOOTNOTE_DLG in sw shells

Change-Id: I6f1ce98acf44a3811e4666a6bd9952232429f55a
Reviewed-on: https://gerrit.libreoffice.org/14312Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>

Change-Id: I704e7d0ca6422f1388cd111c6998ba6ceb48d2bf
Reviewed-on: https://gerrit.libreoffice.org/14311Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

Change-Id: I5a594b3d6ef84b022ce4a92a865beba735d47113
Reviewed-on: https://gerrit.libreoffice.org/14322Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

... as the mdds storage segment may change when a formula cell is
inserted at a position such that two segments are merged into a new one.

Change-Id: I449a89005418aff7ea12099ea931b786959dbc3b

Change-Id: I955c8ac4dbe002d23531df7eb10fb4444d6b5157
Reviewed-on: https://gerrit.libreoffice.org/14292Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>

::dbtools::ensureRowSetConnection throws on failure, and it can fail
of course if the database doesn't exist

its not generally useful to throw through a vcl event handler as that just
leads back to the dispatch loop.

Change-Id: I8e8f4cff06de5684f163ed1b658a8794f54a7df2

Change-Id: I540d192a361938cd4dca811e7170d762841662ee

The code used to draw a black circle and populate it with thumbnails
of the documents inside the folder, but they are always empty because
we are not generating any thumbnails.

Change-Id: Id0b63f7e901525b5a962d12441cda84dfeb59d52

Lossless PNG optimisation for all Icon Themes.
Icons were run through ImageOptim 1.5.4 (https://imageoptim.com/). Total saving ~839KB.

Breeze: Saved 13.9KB out of 29.1KB. 47.5% per file on average (up to 54.5%).
Crystal: Saved 44KB out of 1.2MB. 3.7% overall (up to 85.8% per file).
Galaxy: Saved 10.7KB out of 2MB. 0.5% per file on average (up to 56%).
Hicontrast: A few bytes.
Human: Saved 288KB out of 1.8MB. 15.8% overall (up to 81.1% per file).
Industrial: Saved 22.3KB out of 1.1MB. 1.9% overall (up to 88.7% per file).
Oxygen: Saved 177.8KB out of 2MB. 9% overall (up to 95.3% per file).
Sifr: Saved 41.2KB out of 315.8KB. 13% overall (up to 63.4% per file).
Tango: Saved 130.8KB out of 1MB. 13.1% overall (up to 86.4% per file).
Tango_testing: Saved 110.3KB out of 546.7KB. 20.2% overall (up to 86.4% per file).

Change-Id: Id43faba59fa738cd2cde60df61c16d8cf2e06daf
Signed-off-by: Andrew <dent.ace@gmail.com>

* Add get() method to tools::SvRef

Change-Id: I3825852176c9c37f3ee21f4fd57328b1c7451951
Reviewed-on: https://gerrit.libreoffice.org/14268Tested-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

XclXmlUtils::ToPsz is the one I needed

Change-Id: If4aee9a2a9c943535d7e07b08f61b419ac2c8a22

the Parent window isn't a SwScrollNaviPopup anymore

Change-Id: I5c539d125aac28bfee0beb842ff2df001e19c945

Change-Id: I2ada1394c4cbee7c27b02f241ee4a254597a5dd8

use the same guards as the SwFldEditDlg does

Change-Id: I7d135cf969d9925cec57a90ef0fba8fe40f664eb

Change-Id: Ib01e444b24ad9288970066a47f3fcf3dbee1fc49

The code falls back to the traditional interpreter when necessary.

This reverts commit ad582ce7.

Change-Id: I4cbafba2c469c0814dcc5c5210db5ce495e6b641

Put using it in #if 0 for now, though.

This reverts commit 2021275f.

Change-Id: Ia6541df12ee97747badbbedd758873688190b00c

For iOS, the pBuffer argument to LibreOfficeKitDocument::paintTile()
is not a pixel buffer, but a CGContextRef. Thus we can keep the
existing CGContext-oriented vcl code for iOS as is.

I commented out the <touch/touch.h> -using code in sw, and it probably
would be OK to just kill <touch/touch.h> and remove use and
implementations of its API. I don't think they are used by our current
Android or iOS code.

This concludes the commits of the work I did at the post-FOSDEM
hackfest and on the flight home. TiledLibreOffice now works as before,
but uses LibreOfficeKit.

Change-Id: Ibbde6e11d660561840eab6335f7fc0db6758173e

Change-Id: Ic5c6d6501d477544c6e4a4d6bb2510b6a5a92956

Change-Id: I5971aa4db91adc89f8aeeb7384a1893dab329411

Change-Id: I4104f39227cb42d5cc2013c9ae0f4bb15e0f0f00

Either build mozab or mork driver, and install it in Repository.mk too.

Change-Id: I456d6cb4ebe99625a64a960d1133cb06d836e8c7

provide a callback when a menu item gets deleted

Change-Id: I5b5f1a181fb10f53f6b1fe7b5637d385e1517530

Change-Id: I2e121c36021508bbdbc9111087f5cc6e1aae6524

This reverts commit c0f08040,
logic error, true fix will follow.

Change-Id: I85a5f35fb8b44734e84c686ecdffc6096226b91d

If e.g. a document is slow to redraw, scrolling it by holding down a key such
as PageDown can make LO lag behind and continue processing the queued up
input events even (long) after the key has been released. Since with autorepeat
keyboard events the normal user expectations are to hold it until something
happens and the exact number of the events doesn't matter, simply discard
excessive autorepeat keyboard events if LO can't keep up with them.

Change-Id: I45acdc9aa5033647fb80760991437dddfcb6591c

There is one event per wheel step, so wheeling more can create a number
of wheel events, and LO processed those one by one. If processing one took
long (e.g. the repaint was complicated), the scrolling visibly lagged.

This commit works only for X11 VCL backend (and by extension, KDE3/4 backends).

Change-Id: I5eff7446da16167ec75925e75243314c68bc81a4

Without this, they can activate after any call to the event processing,
so they may activate in cases such as when updating progressbar while
loading a document, or on repeated user input (so things like showing
spellchecking get updated when the app is busy redrawing). This change
makes idle timers activate only when there's nothing more for the event
loop to process. It's a bit of a question if this doesn't break something
that happens to expect idle timers to be not-really-idle timers, but oh well.

No change for non-X11 platforms, as there's I don't know how to check
the event queues.

Change-Id: I074a88f2f5eeb4b456a11916a0ec2ad6f54dfbab