Create a rewriting plugin for finding methods that simply return
object fields, and should therefore be declared in the header,
so that the compiler can reduce method calls into a simple
fixed-offset load instruction.
Change-Id: I7a620fc54250b79681918dc31ed9a8f21118c037
Signed-off-by: Stephan Bergmann <sbergman@redhat.com>
This required some changes to the framework:
* Init-/DeInitVCL is no longer done per individual test in BootstrapFixture, but
once per CppunitTest invocation in a new vclbootstrapprotector (similarly to
the exisiting unobootstrapprotector). CppunitTests that need VCL now need to
declare gb_CppunitTest_use_vcl.
* For things to work properly, the UNO component context needs to be disposed
from within DeInitVCL (cf. Desktop's Application::DeInit called from
DeInitVCL). The easiest solution was to introduce an
Application::setDeInitHook (where the hook is called from DeInitVCL)
specifically for vclbootstrapprotector to call.
* PythonTests don't (yet) call DeInitVCL; they still hook into
BootstrapFixture's original test_init functionality (to call InitVCL), and do
not make use of the vclbootstrapprotector.
Change-Id: I4f3a3c75db30b58c1cd49d81c51db14902ed68b2
Introduce BitmapFilter as a general bitmap filtering class, and
make scale "super" algorithem independent as BitmapScaleSuper
which uses BitmapFilter as superclass.
This is an ongoing work to make some bitmap algorithms structured
and more independent from the big bitmap class This will make them
easier to work with, test and optimize.
Change-Id: I37d29709b2af95cab2f6da21129302f5be79318b
...which does not make sense. On Linux and Mac OS X, they potentially end up
exported from multiple libs (weakly, though), while on Windows the potentially
even end up not emitted at all, which could cause link errors.
Change-Id: I092c9ba39e686c17b6e91581cdd4753f1c4d582f
In grouped list text area does not cover the whole
shape but just a part of it at the top.
To get the same visual effect modify text distance
attribute.
Change-Id: I32f30d0afbc1975f940c4562ec65f46596e97060
because those styles are "pseudo-styles" and a new one cannot be
added. The possibility is supposed to be disabled, and it is
disabled in the floating stylelist. The old code assumes there
can only be one of these stylelists and when a stylelist
queries if the "new" should be disabled the callback asks
the stylelist what family is selected, but only asks the floating one.
So, floating closed, sidebar open, the new is not disabled.
Implement the ancient TODO now that we have to. Instead of asking
the stylelist what family is selected, query the frame for what
is the current SID_STYLE_FAMILY as set by whatever is the active
stylelist.
What's disturbing is the SID_STYLE_FAMILY values are not SfxStyleFamily, but
indexes that have to be mapped to SfxStyleFamily. I bet there are a pile of
bugs around that, especially with little islands of different conversion
codesites
Change-Id: I85c8032d7c26ae6eea245685748f89b2a860e767
COMPHELPER_DLLPUBLIC of an inline function is wrong. Nevermind that I just
followed the example of the preceding functions.
And it is too bikeshedding-inducing anyway to use recursion.
This wasn't used anywahere yet anyway in committed code, just locally in a
tree of mine. Will add a local copy there.
This reverts commit ee0d7d4c52.
To prevent rogue DocBasicItem objects (which are global objects) from
attempting to stop listening after the main app is dead.
Change-Id: I68a667137ca4e0eff18278f4d2d5dd67466eaf34
This was keeping form controls (e.g. CheckBox and RadioButton) to
transition from flat style to 3d style, while the 3d-to-flat
transition was allowed, making flat some unescapable pit of doom.
Whatever the goal of that pinning, it will have to be achieved in a
different way.
Change-Id: Icc4dfb9673a7dfe4a7d8726fbfd23f75499f8b41
==3768== 512 bytes in 1 blocks are possibly lost in loss record 25,532 of 28,010
==3768== at 0x4A06965: operator new(unsigned long) (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==3768== by 0xB52FA41: __gnu_cxx::new_allocator<bool>::allocate(unsigned long, void const*) (new_allocator.h:104)
==3768== by 0xB52EC00: std::__cxx1998::_Deque_base<bool, std::allocator<bool> >::_M_allocate_node() (stl_deque.h:533)
==3768== by 0xB52DE4E: std::__cxx1998::_Deque_base<bool, std::allocator<bool> >::_M_create_nodes(bool**, bool**) (stl_deque.h:627)
==3768== by 0xB52C624: std::__cxx1998::_Deque_base<bool, std::allocator<bool> >::_M_initialize_map(unsigned long) (stl_deque.h:601)
==3768== by 0xB52ACA0: std::__cxx1998::_Deque_base<bool, std::allocator<bool> >::_Deque_base(std::allocator<bool> const&, unsigned long) (stl_deque.h:461)
==3768== by 0xB5296D3: std::__cxx1998::deque<bool, std::allocator<bool> >::deque(std::allocator<bool> const&) (stl_deque.h:791)
==3768== by 0xB52840E: std::__debug::deque<bool, std::allocator<bool> >::deque(std::allocator<bool> const&) (deque:73)
==3768== by 0xB52786B: framework::UndoManagerHelper_Impl::UndoManagerHelper_Impl(framework::IUndoManagerImplementation&) (undomanagerhelper.cxx:221)
==3768== by 0xB525882: framework::UndoManagerHelper::UndoManagerHelper(framework::IUndoManagerImplementation&) (undomanagerhelper.cxx:879)
==3768== by 0x6872683: sfx2::DocumentUndoManager_Impl::DocumentUndoManager_Impl(sfx2::DocumentUndoManager&) (docundomanager.cxx:88)
==3768== by 0x6870FB6: sfx2::DocumentUndoManager::DocumentUndoManager(SfxBaseModel&) (docundomanager.cxx:227)
==3768== by 0x687A272: SfxBaseModel::getUndoManager() (sfxbasemodel.cxx:1677)
==3768== by 0x260E562A: (anonymous namespace)::getUndoManager(com::sun:⭐:uno::Reference<com::sun:⭐:frame::XFrame> const&) (PagePropertyPanel.cxx:74)
==3768== by 0x260E6C60: sw::sidebar::PagePropertyPanel::PagePropertyPanel(Window*, com::sun:⭐:uno::Reference<com::sun:⭐:frame::XFrame> const&, SfxBindings*) (PagePropertyPanel.cxx:189)
==3768== by 0x260E593C: sw::sidebar::PagePropertyPanel::Create(Window*, com::sun:⭐:uno::Reference<com::sun:⭐:frame::XFrame> const&, SfxBindings*) (PagePropertyPanel.cxx:101)
The SfxBaseModel creates the DocumentUndoManager on demand at
sfx2/source/doc/sfxbasemodel.cxx:1685
m_pDocumentUndoManager is a rtl::Reference but debugging into the
acquire/release I see that the
DocumentUndoManager::acquire/DocumentUndoManager::release forward to those of
SfxModelSubComponent which forward them to its rModel without doing anything
else, so the implementations of
DocumentUndoManager::acquire/DocumentUndoManager::release don't actually do
anything directly to the DocumentUndoManager itself so there is nothing that
will really release it.
Either the rModel needs to explicitly destroy it somehow, given that the
acquire/release delegate responsibility to it, or the whole thing is addled.
I rather feel it's addled, so implement as a normal WeakImplHelper1, but
cowardly in addition also keep acquire/release on the rModel.
Change-Id: Ib52544a9276fd8d9d489ad6b6afda12498cc39fa
regression from 0c17ccc493
I think it might actually make more sense for SetName to always Reindex
and have another SetNameNoReindex for selective use where its known
safe to get away with it
Change-Id: I6bf7e1f45130f741954ab525b838f3ac5ff61a89
