while self compiled one still segfaults frequently, it is a tad better
than the included one (and compiling froms source instead of using
precompiled binaries is a net win in any case)
Out of 50 attempts with the mergedlo.dll, the new compiled dump_syms
succeeded 11 times, the shipped one succeeded 7 times - so could still
be regular variance in that small sample size, but at least not
significantly worse either.
Change-Id: I86c213fe6aece58f1391d4c2bf9906b85eee9c57
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117056
Tested-by: Jenkins
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
Another place where SfxItemState was used as placeholder
Change-Id: I54e549db50b7485024a305eb39fda848cbed6b78
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117036
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Armin Le Grand <Armin.Le.Grand@me.com>
Tested-by: Jenkins
The mechanism around NUMRULE_STATE does only do something
when the SfxItemState is actually SfxItemState::SET, so this
can be reduced to use a valid Item shared_ptr when this is
the case
Change-Id: I973648abff3696057f2ca21608ed64570742dddb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117043
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Armin Le Grand <Armin.Le.Grand@me.com>
Tested-by: Jenkins
Another place where SfxItemState::UNKNOWN was used as placeholder
Change-Id: Ie2757e8356bfea7010b0442d72ef7865cbb18d5d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117034
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Armin Le Grand <Armin.Le.Grand@me.com>
Tested-by: Jenkins
This extends the solution from
commit fa8db25af6
Date: Thu Jun 3 20:47:10 2021 +0100
gtk4 crash on exit
to cover the qt5/kf5 cases as well, after seeing
equivalent crashes with kf5 VCL plugin
(e.g. when starting Writer and no profile exists) after
commit 556243467a
Date: Wed Jun 2 13:06:54 2021 +0200
fix shutdown leak in CppunitTest_sw_filters_test
Change-Id: I4ef25bac0891c059588dbc8e6abe50db14230dea
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117042
Tested-by: Jenkins
Reviewed-by: Michael Weghorn <m.weghorn@posteo.de>
and place initial contents into a section by default
Change-Id: I1e7bb20bf5efb481e4f7ed0f38c8b46c06f7a7d5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117050
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
and
cid#1474351 Untrusted loop bound
cid#1474118 Untrusted loop bound
this time without second guessing the original intent
Change-Id: Iaa6b636a08ed29feaf709fbcbac7deac761a0fc7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117045
Tested-by: Caolán McNamara <caolanm@redhat.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
* Update helpcontent2 from branch 'master'
to ab81b6bfa02cd89c244bc55148486324863844f8
- sf_form(control) Python support
Change-Id: Iad14d1c5c7d353e50aba84d0cfa461e56f69f4ef
Reviewed-on: https://gerrit.libreoffice.org/c/help/+/115969
Tested-by: Jenkins
Reviewed-by: Rafael Lima <rafael.palma.lima@gmail.com>
of 'Related: cid#1474351 Untrusted loop bound'
and add in documentation for the records
This reverts commit 2134d83b8d.
This reverts commit 7c2e605843.
This reverts commit 3748fe3e4f.
Change-Id: I5b407eb4388a8d67fc863cc6070fbbf8b15321d9
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117044
Tested-by: Caolán McNamara <caolanm@redhat.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
The anchor type is determined during insertion, it's stored as direct
formatting after that.
We used to have a user-profile default for the anchor type. This commit
allows customizing the Graphics or OLE styles: if they specify a custom
anchor type, then that is user instead of the user profile setting.
This allows creating templates where the default depends on the used
template and not on the user profile.
The UI for this was added in commit
5951da5175 (sw image anchor type: add
style UI for this, 2021-06-10).
Change-Id: Id05342a5f38dc6267cdbe68b248dc50b87854ce2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117040
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
which is a problem since...
commit 33d2cf7008
Date: Sat Jun 5 19:30:46 2021 +0200
WMF Change the sign of read fields in BITBLT and DIB records
Change-Id: Iea1261a52866d691435f0612f672636009c00355
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117039
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
With previous implementation, the RestoreDC index argument was skipped,
and always the last entry was taken.
With this commit the support for reading SaveDC by specific index
was added. The SaveDC/RestoreDC index support was added for
both EMF and WMF, according to [MS-WMF] and [MS-EMF] documentation.
Change-Id: I9b8a1a41462ae01de25ac3c85e453bcd80e05537
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117033
Tested-by: Jenkins
Reviewed-by: Bartosz Kosiorek <gang65@poczta.onet.pl>
but it's not a spreadsheet
since...
commit dad7fe7227
Date: Thu Jan 2 16:16:55 2014 +0100
Fix math categories to make SUSE rpm check happy
Change-Id: I0397ea12cd7f29f4b0c1f0e0ad76a3429b55c455
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/116884
Tested-by: Jenkins
Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
ParaLineSpacingControl::Initialize() is another place where
SfxItemState::UNKNOWN was just used as a placeholder for
a non-set SfxItemState (aka could not be queried)
Change-Id: I95ad01579e5aa4c86ace619e2201481742297c2b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117016
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Armin Le Grand <Armin.Le.Grand@me.com>
FontworkBar::getState does not need to check for SfxItemState::UNKNOWN
at all, actions solely depend on FontWork object being selected.
This also greatly simplifies that method. Also, the optimization
by passing in a variable to checkForSelectedFontWork and remember
if already computed can be removed - also in other places where
it had to be given, but was not re-used at all
Change-Id: I35b1f36195feb1d645619665d2dd65a84b75b118
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117014
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Armin Le Grand <Armin.Le.Grand@me.com>
In this case SfxItemState::UNKNOWN was used as boolean
state to decide to keep extracted value or replace. First
method even throws an exception when SfxItemState::UNKNOWN
should happen.
I tried to be very careful in changing these methods, so
another look/check from reviewers will be welcome. I think
that code was unnecessarily complicated, too, butz hope to
not have touched semantics (behaviour) at all
Change-Id: Id213fedcafcfb655749caf2659f9a15b240ec27d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/116999
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Armin Le Grand <Armin.Le.Grand@me.com>
The two
> aSet.Put( aAnch );
had been like that ever since at least 84a3db80b4
"initial import", but only now aa9cb8e147
"-Werror,-Wunused-but-set-variable (Clang 13 trunk)" removed the lines of code
between them (and which should not have changed neither aSet nor aAnch as a
side effect, which could have explained the putative duplication), making the
duplication stick out. Lets assume this was an error ever since, and that just
one call is enough.
Change-Id: Ie1ad54fb31212849f242eea77bc9e436b6fd92f7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/116980
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
This was added in commit de0866b507
(sw-invert-border-spacing.diff:, 2010-09-14), it's off by default and is
on for Word documents.
Change-Id: Idf1999dbe9e641259e09eaafbebd347199cd5db4
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117015
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Two tests where added, IntersectRectClipRegion and
IntersectRegionClipRegion, along with their .svm files.
The tests create mutliple IntersectClipRegions and test
their attributes.
Change-Id: I33b7b13b4f3eef1de07cc063ef22fea034d65cfc
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/116967
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
by normalizing copy of PaM instead of reference
Change-Id: If2c8664de18eb9473151c5961ad7654932ca37e2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/116875
Tested-by: Jenkins
Reviewed-by: Jim Raykowski <raykowj@gmail.com>
With previous implementation, during reading of rectangles
the optimizations were made after reading every single
rectangle. This was causing performance issues, with many
rectangles (eg. 2500 rectangles).
With this commit, the optimization is made after reading all
rectangles. It is improving performance of FILLRGN, PAINTRGN and
EXTSELECTCLIPRGN records.
Change-Id: I1b8b844efddd08e9bf6f6726c3fdf213a629883f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/116996
Tested-by: Jenkins
Reviewed-by: Bartosz Kosiorek <gang65@poczta.onet.pl>
...ever since the code's introduction in
4e386a897c "INTEGRATION: CWS macosxspellchecker"
Change-Id: I932b7ff4cc8aabf2e70379ed9ebe361881246c46
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/116997
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
