new statusbar element property mandatory=true/false to determine
if this element can be hidden if total statusbar width is not
enough to fit all elements.
marked some calc and draw statusbar elements as not mandatory.
Change-Id: I20e26d3c4bd865e94ea48632a1e97d55f3fa712f
Reviewed-on: https://gerrit.libreoffice.org/56443
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
The case when the thread is re-scheduled exactly after checking for
m_bExecute but before taking the solar mutex.
Change-Id: I494c123a6225da60407391771622208c055ae49b
Reviewed-on: https://gerrit.libreoffice.org/56585
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins
GCC trunk towards GCC 9 emits -Wdeprecated-copy because SalVisual has implicit
copy functions but a user-declared dtor. The implicitly-defined copy functions
are actually used (so cannot be deleted), but in fragile interaction with the
semantics of the user-provided dtor in the SalColormap(sal_uInt16) ctor.
Changing SalVisual into a move-only class (moving the "this instance must delete
its visual member" information) doesn't work well, as SalDisplay::ScreenData is
used as a supply of SalVisual instances (but which will never be of the "this
instance must delete its visual member" variety) that are then copied into
SalColormap::m_aVisual.
As only SalColormap creates SalVisual instances of the "this instance must
delete its visual member" variety, what actually works is to track that
information in SalColormap instead of directly in SalVisual, and make
SalColormap a move-only class.
Change-Id: Ib968a38c40b660ce91981b3c7b281f4add6ece6b
Reviewed-on: https://gerrit.libreoffice.org/56579
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
When subsets are non-contiguous, we need to include the parts between
subsets.
Change-Id: I28214dccc75e6a6af5c65397b2126049a65bf79f
Reviewed-on: https://gerrit.libreoffice.org/56571
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: Ib2301526d7aa6982af6c8c79ed7e9a4c34b7bbf7
Reviewed-on: https://gerrit.libreoffice.org/56491
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
There appears to be a lot of caching involved in the code invoked
by ScGetPivotData(). Testcase is e.g. tdf#89186-1 (with minimal
formula size disabled again).
Change-Id: If7e2a972898c358dc8f80cf238f8d5ccf541f16d
Reviewed-on: https://gerrit.libreoffice.org/56525
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
Lets assume that that was an oversight, and change all those user-provided
copy assignment ops into implicitly-defined ones (that will now also copy
Bundle::mnColor). Avoids -Wdeprecated-copy with GCC trunk towards GCC 9, as the
copy ctors were already left implicitly-defined (and so already copied
Bundle::mnColor).
Change-Id: I43e323b55d059c7afecaddab0d777b6df47182c8
Reviewed-on: https://gerrit.libreoffice.org/56539
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
The assertion fails when one reproduces tdf#69460 but removing the
assert() does not cause any problems.
Change-Id: Idaa5d41a0cf76073e406d7152fc3b4e9ebd3e876
Reviewed-on: https://gerrit.libreoffice.org/56538
Reviewed-by: Tor Lillqvist <tml@collabora.com>
Tested-by: Tor Lillqvist <tml@collabora.com>
...as that member is only used in m_aIdToStyle, and it was confusing how the
user-provided HashedStyle copy ctor initializes it to zero while the implicitly
defined copy assignment op copies it. (And existence of only one of those
copy functions also triggered -Wdeprecated-copy with GCC trunk towards GCC 9.)
Change-Id: I176ef95af8780fefe21e53f61fd66f5e2d9156ab
Reviewed-on: https://gerrit.libreoffice.org/56528
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Which is used in e.g. the Calc cell border dialog. The approach is
similar to commit 60790935cc (tdf#96657 -
vcl opengl: implement invert: Track Frame., 2016-01-09) but that one was
for rectangles, this one is for polygons.
Change-Id: Ib1feebab2d14f4450fee0afe96afcea906965fdb
Reviewed-on: https://gerrit.libreoffice.org/56534
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins
add to NB grouped (full and compact) in Edit group
cause Format is only Format selected text but not Format page.
Edit group is most generic group that fit's for page selection
Change-Id: I49af8ff5b560655d43c55818da391a768d26ef04
Reviewed-on: https://gerrit.libreoffice.org/56430
Tested-by: Jenkins
Reviewed-by: andreas_kainz <kainz.a@gmail.com>
If the fon supports Graphite features we can get the feature
descriptions from the font, but this descriptions are mostly
limited (available in a limited number of languages), so it is
better to only use this descriptions only if the feature does not
have an OpenType description available.
Change-Id: Ic0b81cb8cc22e1e73f30e6a960167cfa982fbfc7
Reviewed-on: https://gerrit.libreoffice.org/56318
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
pasting content from another cell
In tiled rendering case (headless) each window has its own clipboard,
otherwise exists a unique clipboard (UNO service).
Change-Id: I7d5d0b085faeaffa3fc0a80914fbe9349f4aa402
Reviewed-on: https://gerrit.libreoffice.org/56508
Tested-by: Jenkins
Reviewed-by: Eike Rathke <erack@redhat.com>
When embedding fonts to ODF documents we can get into a situation
where the same font is embedded multiple times and this wastes
space. To prevent this we calculate the checksum of the font first
and track the fonts that are embedded. When the font has the same
checksum, the path to already embedded font is returned.
Change-Id: I087c42d9a63b1697b9b4ea985dbce2fda760a996
Reviewed-on: https://gerrit.libreoffice.org/56507
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
and spelling out using the new NatNum12 modifier
to support common "in", "on", "from", "to", etc. date
suffix variants, "dash-free" spell out format of years
of the new Hungarian orthographical reform, and months
with articles.
For example, "in 2018" is "2018-ban", "in 2019" is "2019-ben"
in Hungarian because of "vowel harmony", the word stem
dependent fluctuation of the suffix variants, a frequent
linguistic feature of other agglutinative languages, too,
including Estonian, Finnish and Turkish.
Note: some of the new date formats will work correctly only
with the upcoming update of the external libnumbertext.
Note 2: add also alternative (real) abbreviated month names,
because the default abbreviated month names are Roman
numbers.
Change-Id: Ibb33ff6a627b8e27fd02388653e3b33ebd446a10
Reviewed-on: https://gerrit.libreoffice.org/55637
Tested-by: Jenkins
Reviewed-by: László Németh <nemeth@numbertext.org>
If that (implicitly defined) ctor had ever been used, it would have failed to
increment m_nRef, so could have caused m_pInstance to be deleted too early. So
better mark it as deleted. (LO itself apparently didn't call it. If 3rd party
code would want to call it after all, it could be changed from deleted to
properly user-provided, incrementing m_nRef, in the future.)
The implicitly defined copy assignment op happens to already do the right thing,
and /is/ used, e.g., during
> m_aNode = _rhs.m_aNode;
in DriversConfig::operator= (connectivity/source/commontools/DriversConfig.cxx).
Mark it as defaulted to avoid -Wdeprecated-copy with GCC trunk towards GCC 9.
Change-Id: I527123d9de837c311d30a270feb09dc173d9f411
Reviewed-on: https://gerrit.libreoffice.org/56512
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
