Existing documents unwittingly may have used that as decimal
separator in such locales (though it never was as this is always
the en-US locale).
"Regression" from
commit 9336286a7e
CommitDate: Fri Nov 30 22:15:22 2018 +0100
[API CHANGE] Resolves: tdf#42518 new KParseTokens::GROUP_SEPARATOR_IN_NUMBER
Change-Id: I0ffc1b8ec7c1820fccd8277036c64093dddf82fe
Reviewed-on: https://gerrit.libreoffice.org/80023
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
when validation dialog is run modally, but changed itself to unmodal
during execution.
only NotifyModalHierarchy on modal<->unmodal transition and not on
modality depth
Change-Id: I85d4f67e5b0426aafbf80b1ae25fc3e3f19717af
Reviewed-on: https://gerrit.libreoffice.org/80039
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
instead of just a point
Change-Id: Ibadecd64f3296264790373528427a7a528646c73
Reviewed-on: https://gerrit.libreoffice.org/80038
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
Unfortunately the backspace key-events we emit trigger uno accelerator
handling, which happens another PostMessage further out, so cheat by
doing it synchronously, and relying on the PostMessage inside to get the
ordering right.
Change-Id: Ibee80af7674fd5107cb1c9ba323071ac024c45ae
Reviewed-on: https://gerrit.libreoffice.org/79883
Tested-by: Jenkins
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
When we disable LOKit at the end of a unit-test, we do so
before any of the smart-pointers go out of scope. These
instances are then destroyed when LOKit is no longer active,
in spite of the fact that they were created while it was.
This leaves some dialog/window routing notifiers dangling,
which cause assertion failures and access-after-free (when
they are invoked on destruction inadvertently). The latter
issue was found with 'make CppunitTest_sd_tiledrendering
VALGRIND=memcheck'.
The fix is to move these flags to the setup and teardown
stages of cpp-unit.
Change-Id: Ic4481305ca1a5cca14d70a4ebbc820a3fdeea9cc
Reviewed-on: https://gerrit.libreoffice.org/80028
Tested-by: Jenkins
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
The problem is that the image in the bugdoc's footer is anchored in one
cell, but its position has a value that shifts it to the next column and
next row. The next column is the problem for the horizontal position.
So build on top of the previous vertical position fix, and make sure
that CalcRelPosX() doesn't limit the position inside the current cell
for an in-table, follow-text-flow, wrap-though image.
Once that's in place, make sure that we don't try to grow the cell due
to follow-text-flow, wrap-though objects in CalcPosition(), since
in a wrap-through vs follow-text-flow situation, the wrap-though should
have priority (should not affect size of cells).
Finally, now that cells don't grow in this case, the previously added
special-casing of footers in SwFlyFreeFrame::CheckClip() is no longer
necessary. Not growing the cells means we don't try to re-position the
object.
Change-Id: Ic55e4b5188704fa70314f91fe9a01987b6a56d7b
Reviewed-on: https://gerrit.libreoffice.org/80032
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
0e68bac852 "set up python unit test
infrastructure" had introduced unotest/source/python/org/libreoffice/unotest.py
with the comment and commented-out code. It claims it has extraced "boostraping
code from convwatch.py to unotest.py", but I can't find a trace in the git
history of a prior incarnation of that comment and commented-out code. It is
unclear what problem exactly the commenting-out of the check was supposed to
address. So put the check in place and see what, if anything, breaks.
(The commented-out return appears to be truly unnecessary, and has been removed
completely.)
Change-Id: Ie2108d70fbe076cd0832bd23bfb29fc60eedffef
Reviewed-on: https://gerrit.libreoffice.org/80029
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
(f5bbb2db98), which had erroneously been dropped
by 1ae450504c "drop gtk2 support"
Change-Id: Ib7e2fc5bfa394dd6c49d5c9710e1c415fdfe2186
Reviewed-on: https://gerrit.libreoffice.org/80026
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
No need to inherit, just create a typedef
Change-Id: Ifeabb90e640e10e590f793716c022f91850ae8b9
Reviewed-on: https://gerrit.libreoffice.org/79991
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
If a shape is vertically flipped, then the text is rotated by 180deg
around the shape center. The rotation is done by SdrTextObj. There
text rectangle and shape rectangle are the same, so the rotated text
rectangle has the correct position in regard to flipping. But the
text rectangle for SdrObjCustomShape is set by the shape author in
the TextFrames attribute and might have an asymmetric position. The
patch translates the text rectangle so, that is will be at the
correct position after the 180deg rotation.
Change-Id: I42e552394cc4b0103530eab48750dd621c52cc5e
Reviewed-on: https://gerrit.libreoffice.org/79984
Tested-by: Jenkins
Reviewed-by: Regina Henschel <rb.henschel@t-online.de>
...which fixes losing the background color of the page. That was
broken by LO63 commit cc899c6967
since an obsolete RES_BACKGROUND was not necessarily created.
Change-Id: I944a1b3f3df1468c283f93a49ffacfbd223fb392
Reviewed-on: https://gerrit.libreoffice.org/79912
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
Reviewed-by: Michael Stahl <michael.stahl@cib.de>
No idea why this only started to show now in my clang-cl build, but this is
similar to what 23a8d5ffbb "Upgrade external/boost
to Boost 1.69.0" did at ucb/Library_ucpcmis1.mk:
workdir/UnpackedTarball/libepubgen/src/lib/EPUBGenerator.cpp includes
boost/uuid/uuid_generators.hpp, includes boost/uuid/random_generator.hpp,
includes boost/uuid/detail/random_provider.hpp, includes
boost/uuid/detail/random_provider_include_platform.hpp, includes
boost/uuid/detail/random_provider_bcrypt.ipp, includes boost/winapi/bcrypt.hpp,
includes system bcrypt.h.
Change-Id: I1dcfb9a13515d3e6f0a249c4a42ae5ebbf11fcf4
Reviewed-on: https://gerrit.libreoffice.org/80004
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Also return the proper kind and creator values set in
WKS4Parser::checkHeader() to callers of
WPSDocument::isFileFormatSupported()
Change-Id: I3adfde0e18669b80956a9b577080fe8388dd3b1f
Reviewed-on: https://gerrit.libreoffice.org/79981
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
look for places where we can replace sequential additions to
OUString/OString with one concatentation (i.e. +) expression, which is
more efficient
Change-Id: I64d91328bf64828d8328b1cad9e90953c0a75663
Reviewed-on: https://gerrit.libreoffice.org/79406
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
regression from
commit 1cd32bcf1b
Date: Mon Dec 10 11:28:59 2018 +0200
loplugin:useuniqueptr in vcl
where I failed to note that sort needs a different kind of comparator to
qsort.
Also fix another similar issue I introduced in that commit
Change-Id: I5a93bd0567cd5dd4344b9cf2c362ebff60fa0027
Reviewed-on: https://gerrit.libreoffice.org/80007
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
LO can actually create clipboard content, which is not backed by
an XTransferable, for example when copying / selecting the text
of the file name in the QFileDialog. So the ownership check in
Qt5Clipboard::handleChanged is wrong and we just have to prevent
freeing the content in handleChanged while changing the handle
from Qt5Clipboard::setContents.
This patch simply sets a boolean member, while the QClipboard is
changed by LO, to handle this case.
Change-Id: Icc41c32c1f9807e7adff7a9ae16a6c6cacc83f1b
Reviewed-on: https://gerrit.libreoffice.org/79992
Reviewed-by: Michael Weghorn <m.weghorn@posteo.de>
Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
Tested-by: Jenkins
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: I986ba9f661854d1b5fafc337ec109b0ee558a86a
Reviewed-on: https://gerrit.libreoffice.org/79904
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
- in soffice.ini (sofficerc) the entry "CrashDumpEnable" default is "true"
- when false then the Dump.ini and the dump-file are not written
- when the switch --disable-crashdump is set, then the
switch "CrashDumpEnable" set to "false"
- when the entry "CrashDumpEnable" is missing, in this case is the
default true, too
- the checkbox under Options-General "Send crash reports to ..."
is deactive and shows off (only view, not change the config)
- when set the environment variable "CRASH_DUMP_ENABLE" to any char
then the switch "CrashDumpEnable=false" are overrules with true
and the Dump.ini and dump-file are write
Change-Id: I34e7795640eb95d111e18b0ad46ec67b2c126b19
Reviewed-on: https://gerrit.libreoffice.org/79273
Tested-by: Jenkins
Reviewed-by: Juergen Funk (CIB) <juergen.funk_ml@cib.de>
...a namespace enclosing 'std'" (clang-cl). (Upstream <https://github.com/
unicode-org/icu/commit/5a34bfb1516a6719b5f470063c6be2f47446f0b2> "ICU-20209 Fix
build failures on Windows with std::atomic not in enclo…" covers more things, so
just include here what is absolutely necessary for our needs.
Change-Id: I10e61b24a5d73b372bfd719d97fc9678029dc205
Reviewed-on: https://gerrit.libreoffice.org/79953
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
regression from
commit 78b9dac2ee
remove internal use of 16-bit packed formats
we are trying to create a 32-bit image in
drawinglayer::primitive2d::createDefaultCross_3x3
Change-Id: Iaf19326b1e2c11c5d535dd8f317086bebbaaebec
Reviewed-on: https://gerrit.libreoffice.org/79948
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
- remove double code
- using of all the methode of the CrashReporter-Class
- all methode only active when crash-dump enable
except the addKeyValue
With this change the handling for the patch
tdf#127711 A runtime-switch for the MiniCrashDump
would be simpler
Change-Id: I339b3b8e06f7fc2cd3c0d34ece112a6fd352913a
Reviewed-on: https://gerrit.libreoffice.org/79272
Tested-by: Jenkins
Reviewed-by: Juergen Funk (CIB) <juergen.funk_ml@cib.de>
There were two problems here:
1) CalcHeightWithFlys() considered all follow-text-flow objects when
determining the cell size, while wrap-through objects should never
influence the layout of text (i.e. when they conflict, the second should
have priority).
2) Once the cell had correct height, the oscillaction described in the
SwFlyFreeFrame::CheckClip() comment started. Such a position update was
already disabled for headers, but footers have exactly the same problem.
In the case of the bugdoc, we jumped between 14618 and 14744 twips, till
finally layout loop control kicked in.
[ FollowTextFlow is meant to be same behavior as Word's layoutInCell
shape property, so this is expected to improve rendering of existing
documents. It's not likely that any user would opt in for FollowTextFlow
to have the old close-but-not-exactly-matching behavior. ]
Change-Id: I6b3b672fc82c6c67dbbdd35c349613fe4cda610d
Reviewed-on: https://gerrit.libreoffice.org/79980
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
...MSO 'washout' to our watermark colormode
After d151ed35ad,
Contrast and Brightness was reset only in writer,
so do it earlier for Impress and Writer
Change-Id: Icaf2b565880e31d86e09e2f56bac25eba52c8467
Reviewed-on: https://gerrit.libreoffice.org/79905
Tested-by: Jenkins
Reviewed-by: Xisco Faulí <xiscofauli@libreoffice.org>
The C++ standard leaves it unspecified how ATOMIC_FLAG_INIT is defined, and at
least MS Visual Studio 2017 version 15.9 defines it as {0}, asking for list-
initialization, but std::atomic_flag is no aggregate (it has a default ctor) and
has no suitable ctor for that initializer list. Other standard library
implementations solve that by e.g. adding a std::atomic_flag ctor taking a bool
parameter, and defining ATOMIC_FLAG_INIT as std::atomic_flag(false); but MSVC
apparently relies on some non-standard behavior here that allows to initialize
std::atomic_flag from {0}. But that is apparently not supported by clang-cl,
causing failures like
> [build CXX] workdir/UnpackedTarball/boost/libs/locale/src/shared/date_time.cpp
> In file included from workdir/UnpackedTarball/boost/libs/locale/src/shared/date_time.cpp:11:
> In file included from workdir/UnpackedTarball/boost\boost/thread/locks.hpp:10:
> In file included from workdir/UnpackedTarball/boost\boost/thread/lock_algorithms.hpp:11:
> In file included from workdir/UnpackedTarball/boost\boost/thread/lock_types.hpp:18:
> In file included from workdir/UnpackedTarball/boost\boost/thread/thread_time.hpp:9:
> In file included from workdir/UnpackedTarball/boost\boost/date_time/time_clock.hpp:17:
> In file included from external/boost/include\boost/shared_ptr.hpp:27:
> In file included from workdir/UnpackedTarball/boost\boost/shared_ptr.hpp:17:
> In file included from workdir/UnpackedTarball/boost\boost/smart_ptr/shared_ptr.hpp:36:
> workdir/UnpackedTarball/boost\boost/smart_ptr/detail/spinlock_pool.hpp(77,5): error: no matching constructor for initialization of 'std::atomic_flag'
> BOOST_DETAIL_SPINLOCK_INIT, BOOST_DETAIL_SPINLOCK_INIT, BOOST_DETAIL_SPINLOCK_INIT, BOOST_DETAIL_SPINLOCK_INIT, BOOST_DETAIL_SPINLOCK_INIT,
> ^~~~~~~~~~~~~~~~~~~~~~~~~~
> workdir/UnpackedTarball/boost\boost/smart_ptr/detail/spinlock_std_atomic.hpp(81,38): note: expanded from macro 'BOOST_DETAIL_SPINLOCK_INIT'
> #define BOOST_DETAIL_SPINLOCK_INIT { ATOMIC_FLAG_INIT }
> ^~~~~~~~~~~~~~~~
> C:/PROGRA~2/MIB055~1/2017/COMMUN~1/VC/Tools/MSVC/1416~1.270/Include\atomic(158,26): note: expanded from macro 'ATOMIC_FLAG_INIT'
> #define ATOMIC_FLAG_INIT {0}
> ^~~
> C:/PROGRA~2/MIB055~1/2017/COMMUN~1/VC/Tools/MSVC/1416~1.270/Include\atomic(169,2): note: candidate constructor not viable: no known conversion from 'int' to 'const std::atomic_flag' for 1st argument
> atomic_flag(const atomic_flag&) = delete;
> ^
> C:/PROGRA~2/MIB055~1/2017/COMMUN~1/VC/Tools/MSVC/1416~1.270/Include\atomic(168,2): note: candidate constructor not viable: requires 0 arguments, but 1 was provided
> atomic_flag() noexcept = default;
> ^
The internals of external/boost appear to be the only place where we currently
use ATOMIC_FLAG_INIT, and we can work around that easily by picking a different
backend in boost/smart_ptr/detail/spinlock.hpp for the problematic combination
of Clang and _MSC_VER 1916 (which may need to be extended to further MSVC
versions).
Change-Id: Ie6aa62556efbedd3897ad7803b546cb8088ebe10
Reviewed-on: https://gerrit.libreoffice.org/79967
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>