Change-Id: I91feab63b77055b8f73d0fb481a76586b249160c
Reviewed-on: https://gerrit.libreoffice.org/24595Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: jan iversen <jani@documentfoundation.org>

some places are marked with "dodgy"- need to check those to see
what is going on, because they are leaving dangling pointers behind
in the Menu class

Change-Id: I41d5c7c0fec2f70ce9e3ffdc48cd03d26c0a869b
Reviewed-on: https://gerrit.libreoffice.org/26516Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>

Change-Id: I5026590506b4ddf3db95fb14fe6b0890255a5a25

Project: help  c23a2a94aae6b2f1290145323460848eff867305

Replace ASCII approximation of trademark symbol with real character

Change-Id: I70c466baaaa8e2fbcba8c4c36fd4df6f14cda22b

Validation is done with Microsoft Office Binary File Format Validator if it
is enabled. However binary files are failing validation, so validation is
disbled in corresponding tests.

Change-Id: I9f6ff5c3dd9e3e3c0bb37ce895e60e38a0b43e9a
Reviewed-on: https://gerrit.libreoffice.org/26510Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>

Change-Id: I0fc0b7e2abbadee2f13d477c5ddd2290f37487a0
Reviewed-on: https://gerrit.libreoffice.org/26563Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Takeshi Abe <tabe@fixedpoint.jp>

Change-Id: I1ff2bab766eadac1dd5d483453af0e760634083d
Reviewed-on: https://gerrit.libreoffice.org/26421Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>

id:000178,src:000000,op:havoc,rep:16.bmp
id:000383,src:000047,op:flip2,pos:18,+cov.bmp
id:000403,src:000047,op:arith32,pos:22,val:-33.bmp
id:000317,src:000025,op:havoc,rep:8.bmp
id:000119,src:000000,op:havoc,rep:16.bmp
id:000167,src:000000,op:havoc,rep:8.bmp

revealed since

commit 81e3ca4f
Author: Tor Lillqvist <tml@collabora.com>
Date:   Tue Jun 21 10:34:21 2016 +0300

    Use real assert() instead of DBG_ASSERT()

sanitize invalid palette entry indexes at the outer perimeter on initial load
to try and avoid having to do it in all sort of places in the interior.

we load the palette before getting here, so we should always know the palette size here

Change-Id: I8856787d28310898c3806647fc181a97cfa88514

Change-Id: I98c383106c8d5874bc8eb3daf05ca3f172d9f41a

Change-Id: Ic8887571e376483e84d7c2eb270d3daaa37c63f3

fpicker module TCHAR elimination and A/W clarity

NB fpicker was already being compiled as UNICODE

Use SendMessageW() as appropriate for clarity

Use SAL_N_ELEMENTS for buffer capacity (number of WCHAR) when retrieving class name

Change-Id: Ie6d0b1df1b9a2fb18f4cdfc51e5d4c3c55c9ca17
Reviewed-on: https://gerrit.libreoffice.org/26526Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

Change-Id: I7dc68ef48a69f4878a20e6cb1dd3203dfc3097fd

The wrapper has references to vcl objects and the destructor might have
the last reference to a vcl::Window causing a delete. Avoid potential
problems (I have seen one deadlock in the chart wizard code) by taking
the SolarMutex.

Change-Id: I8afcf79842b2f8e8f822b975629d5c8e18a86163

Change-Id: Ic4c4ba74701303df1ff38c5ed6ed7d521b369053

Change-Id: Ib1fc73c317186ede3f41029aa40b0c6a075c085d

These sleeps are only necessary when you want to run them non-headless
and see what is actually happening.

Change-Id: I20bd80c680c8569bc3c885f261e03db494afee06

Change-Id: Ifff8fe9cb0f005c8a4e892d756ae6b0b95db3544

Change-Id: I7a5cf03daac1b0426672a9d6611aa4cab4aaf91c

Change-Id: I70222d701191f28f9dffb3bfdfbb7510d64d0251

Change-Id: I6ad7898f92c74f37ed2936337d6570904860d535

Change-Id: Ic6e0d99a567b7080928accb61ed036b1c25ec7ff

Change-Id: Ic0589be9b3769279b201dfd314534a087c7f4309

Change-Id: I41796a2a557143030cbd016399c2c1bf5cde37bd

Change-Id: Iaa1a49d7e38c8d87bb3cfd749af94a0b92feee0d

Change-Id: Ibcdf5c1a43ed943c22745545257d5d324ed139ec

Change-Id: I4cdffe62cdabaf110eaaeb0d31a45c44833d354a

Change-Id: Ia9c9accf3c5509393c3e09692b6a7eae756907f9

Change-Id: Ia4b0439af933653bcff5aed9127b2503c622ce0e

Change-Id: I09b724cb80d75b17fd3ad5f3736245f9fca1d17a

revealed since

commit 81e3ca4f
Author: Tor Lillqvist <tml@collabora.com>
Date:   Tue Jun 21 10:34:21 2016 +0300

    Use real assert() instead of DBG_ASSERT()

sanitize invalid palette entry indexes at the outer perimeter on initial load
to try and avoid having to do it in all sort of places in the interior.

png spec says that the palette has to appear before the first IDAT so we
should always know the palette size here

Change-Id: I6e04223adce1c88d037f9cf34862e6f54e381bb0

Change-Id: Ie4b92f3763d2d695709a4919faceb520f874897f

The test fails on a tinderbox. Adding a temporary printf shows that
the number there is 71.

Change-Id: Ic8fbaab9e3d3434e20ba6aef5cc1acd66b5f4283

shell module largely already compiled UNICODE: A/W clarity

Verified by installing release build and trying property sheets etc.

Change-Id: I4965cdadbce635d86da1996c3c1bbd37a981ffde
Reviewed-on: https://gerrit.libreoffice.org/26556Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>

Condition multiply inherits so the Builder::disposeBuilder call was missed

and prenotify occurs during ::dispose when there is no m_pConditionPlayground
anymore, so m_pConditionPlayground == 0 where pPlaygroundCandidate is 0
passes, so a random widget gets cast to somethings its not, hence the bad
cast exception

Change-Id: I6d081dd55df1a68ab893294858e1ab0054d5e049

dtardon says: i forgot about it. it didn't help me anyway

This reverts commit 44819f98.

Change-Id: I017c0107f421590bc61c4b1b7f3df458dd0b18f9

Change-Id: I57f2d016eb6d35d2bc9be090444be1190d68d11d

sal/osl/unx/conditn.cxx:90: pthread_cond_destroy failed: Device or resource busy
sal/osl/unx/conditn.cxx:92: pthread_mutex_destroy failed: Device or resource busy

"Desktop disposed before terminating it"

seen under valgrind with CppunitTest_sccomp_lpsolver

Change-Id: I643cf114b902c38a6a54487fd78c55d84ed78cfc

Change-Id: I7761ac05fbc9be39a6982ba66054c75a331816be

revealed since...

commit 81e3ca4f
Author: Tor Lillqvist <tml@collabora.com>
Date:   Tue Jun 21 10:34:21 2016 +0300

surely here we should ask the palette how many entries are
in it, not assume from the original bitcount that all entries
exist.

presumably the palette can only have <= 1 << BitCount entries in it

Change-Id: Ieb1b98f2f13f702a6a6a20d8cf3d8e9a695141b2