uno::Reference is only allowed to used with classes that have a
::static_type member.
So convert all those places to rtl::Reference.
Maybe we need some LIBO_INTERNAL_ONLY constructors on rtl::Reference and
uno::Reference to make this a little smoother?
Change-Id: Icdcb35d71ca40a87b1dc474096776412adbfc7e3
Reviewed-on: https://gerrit.libreoffice.org/25516
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Most of these syntax changes are found by pylint. Also:
- Use argparse instead of optparse, which is deprecated
- Fix a bug where we tried to multiply by float, which does not work
Change-Id: I7de5a29bd431755e6c28a8bc80b804c775a0c2cb
Reviewed-on: https://gerrit.libreoffice.org/25669
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: jan iversen <jani@documentfoundation.org>
If there is no integer part (with format such as ?/?):
- sStr should be skiped
- SFrac should be completed with the right number of ?,
j==0 must also be treated in ImpNumberFill
Change-Id: I57448eb3b0c68e10779d7fa565379e2604f7f63b
Reviewed-on: https://gerrit.libreoffice.org/25612
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
Yet another reminiscence of String to OUString conversion..
where the terminating NULL-character was obtained and
SvNumberformat::InsertBlanks() effectively did nothing.
Could be triggered already by entering an '_' underscore character as
number format code, which is a place holder for blanks of the same width
as the following character, which there isn't then yet.
Change-Id: I0534e1417d4bd35e9e7ed4bd0170b9ea3b5fb575
The new call to SwTable::SetTableStyleName() was not recorded in
SwUndoTableAutoFormat and hence persisted even after Undo.
(regression from 73f4a06c0b)
Change-Id: Ia7f769dafa62f02ff8e4b0596b48266190c7a69b
it's the same as HelpId in default implementation and extended with
'=?config=foo' string otherwise, depending on different configurations
the dialog/tabpage can have (e.g. different widgets hidden/shown)
Change-Id: I01312a5a88ef7ba784bca315b336420cdb63c8dc
Reviewed-on: https://gerrit.libreoffice.org/25455
Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
Tested-by: Katarina Behrens <Katarina.Behrens@cib.de>
... so ScMarkData doesn't need to be passed. Which it wasn't anyway
because the destination sheet span is part of CopyFromClipContext.
Change-Id: I0addd370d96ab78b22ed957081a1212fc70419a1
we're already using libxslt internals, so using XSLT_STATE_STOPPED isn't an
additional exposure
This probably isn't all that useful to the original reporter in terms of
importing useful data, but it does turn a hopeless situation into something
that can be cancelled.
Change-Id: I08e9a1dcd9ee78e1804faec500bbcca36a546988
* The data argument to the Any ctor call needs to point at sal_Int32, not int.
* All calls to int2enum guarantee that rType is an enum type, so assert that.
Change-Id: I0ccd498420638fee80aeeccc93d9c1e3309f83b0
One of the tinderboxes fails with mysterious warning:
include/tools/gen.hxx:418:36: error: '*((void*)& aViewportExt +16)' may
be used uninitialized in this function [-Werror=maybe-uninitialized]
nBottom = rSize.Height() ? nTop+rSize.Height()-1 : RECT_EMPTY;
Change-Id: I62bf99d1a623ba4bd9d7db5fe049366d0f20e26f
Upstream maintainers, with few exceptions, generally don't care about
warnings in boost headers, hence we re-base our warning fix patches on
every upgrade, which is a pointless exercise in frustration.
Most of the patches are for GCC/Clang warnings, where we could use
-isystem to suppress the warnings (with corresponding hacks in the build
system, because -isystem also disables dependency generation) - but
clang-cl does not support -isystem.
So generate a bunch of wrapper headers into external/boost/include, that
disable all known warnings and use #include_next to get the real boost
header. This allows us to get rid of most of the existing patches.
There is however a bug in GCC that preprocessor warnings like -Wundef
cannot be disabled with a #pragma, so those patches cannot be removed.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53431
Change-Id: I2992bf4a463015f1140489df867bd80757f84541
Reviewed-on: https://gerrit.libreoffice.org/25563
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
...now that sal_Unicode no longer clashes with sal_uInt16 on any platform (in
LIBO_INTERNAL_ONLY code), after e16fa715c4 "Handle
wchar_t as native C++11 type on windows"
Change-Id: Id423dd6235bf14823fa5611b804c0974edbe64b3
Add max-denominator-value attibute to <number:fraction> to comply with ODF 1.3
See https://issues.oasis-open.org/browse/OFFICE-3695
Specify loext or number according to ODF version
This value is only used to stored the number of digits in numerator.
No other values than 9, 99, 999,... are expected.
Add loext:max-numerator-digits attributes to <number:fraction>
to distinguish
??/?? and ##/## or any combination
Not saved for ODF strict
min-...-digits indicates the number of '?'
max-...-digits indicates the total number of '?' and '#'
This needs to be saved with both min-...-digits at least one
because previous versions cannot have min-...-digits of 0. See tdf#38097
Change-Id: Ibd785579aa75edc315e3bd94935a146d82bbeeef
Reviewed-on: https://gerrit.libreoffice.org/24903
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Eike Rathke <erack@redhat.com>
If the shader compilation or loading of an already compiled shader
fails, disable OpenGL in the registry and exit with the
EXITHELPER_CRASH_WITH_RESTART status. The wrapper process will thus
run soffice.bin once more, and this time OpenGL will not be used.
Change-Id: I3cc4a615f00a8a1adb584493861e4956c83cec32
setup_native/source/win32/customactions/shellextensions in MSI Installer compiled as UNICODE
Functions suffixed with A/W (ANSI/Wide) as needed for clarity
Change-Id: I711f781b47c0efdb1e3d9bf995f7658dbb0e38d1
Reviewed-on: https://gerrit.libreoffice.org/25556
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
on drop that it set on drag. It does some uno tunnel foo to drag the data it
needs back out of it in some grotesque fashion.
So we have to follow the same style of hackery as under MacOSX to detect
on drop that there is an active drag started by ourself and so use that
active drag's transferable as the source transferable for the drop, rather
that use the intermediate universal GtkDnDTransferable.
Change-Id: I3c3a94416db908603bde8f15dc5b1c9d726b8dbd
This restores the state before commit
e1b78d3600 (#95181#; call the setValue
method of the XStatusIndicator as often as possible to enable
reschedule, 2001-11-26), which doesn't seem to be necessary anymore,
perhaps due to the current scheduler that has priorities.
Rather than a plain revert, still allow the progressbar to jump back, as
that seems to be used relatively frequently. So just filter out the
calls that would increment the value, but only with a small difference,
compared to the shown value.
Change-Id: I7136b20f1c64e267b0b4a35bbe2564e5163d9468
Reviewed-on: https://gerrit.libreoffice.org/25654
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
This patch fixes the problem that the build for x64 fails in basic module
on 64bit Windows installed 32bit Excel Application.
New code checks the existance of ODBC driver for excel insted of the
existance of Excel application(at this time the bitness of ODBC driver for
excel would match that of building LibreOffice).
What we need is probably not Excel Application but ODBC drivers for proper
bitness.
Change-Id: I62285eb2351f2022754fc34cb2d54db1bd9e8142
Reviewed-on: https://gerrit.libreoffice.org/25301
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
...that was there ever since at least 84a3db80b4
"initial import". None of the "make check" code appears to trigger this code
path.
Change-Id: Ic85b2ef4cf0a0ad06b467558e823cfa4ee31a08c
