Idea originally from caolan.
Found using the following command:
find . -name *.cxx | xargs /opt/local/bin/grep -zlP '(?m)if\s*\(\s*\w+\s*\)\s*delete\s+\w+\;'
Change-Id: I3338f4e22193a6dfd6219c8c75835224a3392763
regression from 13ef16423e
commit 13ef16423e
Author: Armin Le Grand <alg@apache.org>
Date: Wed Nov 21 13:23:01 2012 +0000
For backward compatibility take mirrorings in setTransformation into account
Also found an error in SdrObjCustomShape::TRGetBaseGeometry when MirrorY was used
(cherry picked from commit 4116c33b12d3787c406f0348f89efcb1cf409507)
Change-Id: I7bfb5dea32b8ab8498e3d92975c49b830c81e6fb
Sadly cannot forward declare "struct {...} TimeValue;".
rtl/(u)?string.hxx still include sal/log.hxx but removing osl/diagnose.h
was painful enough for now...
Change-Id: Id41e17f3870c4f24c53ce7b11f2c40a3d14d1f05
Turns out that for normal TextFrames and rectangular drawinglayer
shapes, the engine is written in ODF implicitly. Use the same trick to
describe if the shape content should be rendered by SW or editeng.
This reverts 9d310ecfce (sw textboxes:
implement ODF import/export, 2014-06-02).
Change-Id: Id8afa6368bedcd293e578244abb58e18949bec8d
I introduce a template method into the PTR_CAST machinery
to maintain constness.
There is now a FIXME in sd/../docshell.cxx because I needed
to use a dynamic_cast there to work around the games it appears
to be playing with OLE in-place activation.
Signed-off-by: Stephan Bergmann <sbergman@redhat.com>,
dropping the GCC-extension, unnecessary use of typeof from tools/rtti.hxx
Change-Id: Iba5ace1aa27e02b34fcc91af1e658c43371afd03
Actually make it experimental mode was unecessary on master.
This reverts commit 653bac184f.
Conflicts:
sd/qa/unit/import-tests.cxx
xmloff/source/draw/ximpshap.cxx
Change-Id: Ie8747745a9a89ce4d6500a35d15a98e28bda840a
The previous solution was good because older LO versions
import the fallback image without changing their code, but
it came out it does not fit to ODF standard so export
fallback image as the second object after the glTF model.
Change-Id: Ib9b2044b1f36b32d980cb79f6dac8dcf94d6209b
There is no an accepted mime type for glTF models yet, so
follow collada's mime type: model/vnd.collada+xml.
Change-Id: I1e23a95dbb0caea781dade819f8e08235cb5c1dc
We already had the concept of having a XML_TEXT_TYPE_SHAPE type (for
editeng-based text) and a XML_TEXT_TYPE_TEXTBOX (for Writer-based text),
but in case of draw:custom-shape, the previous was always assumed,
implicitly.
Check if the shape is marked as having a textbox, and if so, use the
XML_TEXT_TYPE_TEXTBOX handler, which makes e.g. the table import work.
Change-Id: I45f28b36df1836eeba89e9eef44c7abd87fcbd56
A simplified version of the semantic match that finds this problem is
follows: (http://coccinelle.lip6.fr/)
// <smpl>
@r1@
statement S;
position p,p1;
@@
S@p1;@p
@script:python r2@
p << r1.p;
p1 << r1.p1;
@@
if p[0].line != p1[0].line_end:
cocci.include_match(False)
@@
position r1.p;
@@
-;@p
// </smpl>
Change-Id: Ib9708d37fbb4c6060f88d5dae3814a2d37b2091e
Reviewed-on: https://gerrit.libreoffice.org/9493
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>
Old versions of OOo and AOO need a fix for Z paths. But AOO 4.0 did not
fix it neither.
Change-Id: I29af75035c7c059b69fcdc3a1ce27b617a30dbd1
Reviewed-on: https://gerrit.libreoffice.org/9012
Tested-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Default mime-type for all media objects:
"application/vnd.sun.star.media"
The problem of missing mime-type detection
still exists. For now only glTF model has
a concrete type.
Change-Id: I4dca26c1c47a564579bbed926bffa3aa5eda6c04
In case the underlying UNO object supports that, which is the case for
Writer. Export was already working before.
Change-Id: I4676c8349ebe1959da004d6e1a024a342da45049
and draw:glue-point it is necessary to move the GluePoints from the last
draw:image where they were automatically imported to the surviving one if these
are different
(cherry picked from commit c011af1087411a9bacd29cd479c807e698b2e92c)
Conflicts:
xmloff/inc/xmloff/xmlictxt.hxx
xmloff/source/core/xmlmultiimagehelper.cxx
xmloff/source/draw/ximpshap.cxx
xmloff/source/draw/ximpshap.hxx
Change-Id: I8f6c875767e9cbfee74838742401356df002b051
examining svx/source/unodraw/unoprov.cxx "Geometry" is a
indeed a drawing::PolyPolygonBezierCoords for the BezierShapes
and a drawing::PointSequenceSequence for the PolyShapes
regression since e44335abe0 because after
223f6b631c we started getting
drawing::PointSequenceSequences in maPath which is the wrong type for the
argument to property "PolyPolygonBezier" for those shapes. Which led me to
incorrectly assume that the all PolyPolygonBezier properties were named
"PolyPolygonBezier" which isn't the case.
so reverting the non maPath hunks of e44335abe0
Change-Id: I013a66778d11a472fc4567e53a9e17e73e2b91ce
See f0a9ca24fd "rhbz#887420 Implement 'block
untrusted referer links' feature" for details. This adds some further /*TODO?*/
comments, and one known problem (marked /*TODO!*/) is that movies/sounds are not
blocked during a slideshow presentation.
Change-Id: Ib2d0c7e4f7b02c4bdec0d8a90cee5e7e1bee8325
For now, this checks for a trusted referer (if the BlockUntrustedRefererLinks
configuration prop is set) in utl::MediaDescriptor::impl_openStreamWithURL and
SvxBrushItem::GetGraphicObject. Checking in additional places will probably be
necessary to block /all/ unwanted communication. Also, some places marked
/*TODO?*/ currently pass in an empty referer (which is always considered
trusted) and will probably need to be adapted.
Ideally, Referer URIs would never be empty (and consistently use something like
<private:user> for cases where access is explicitly initiated by the user and
should never be blocked), but that's a very daunting task, so start small by
identifying the places that potentially need blocking and adding appropriate
Referer URIs there. Also, Referer information should always be computed as
freshly as possible from the context in which an access attempt is made, but,
again, always carrying the information from the context all the way to the
relevant functions is a very daunting task, so for now store the information
upon object instantiation in some cases (SvxBrushItem, SdrGrafObj, ...).
The Referer URI (css.document.MediaDescriptor property; SID_REFERER) was already
used to track macro execution, and there is one place in
SfxApplication::OpenDocExec_Impl where opening of hyperlinks (explicitly clicked
by the user) is done that needs the current document's URI as Referer to check
execution of macro URIs but needs an empty (or <private:user>, see above)
Referer to not block non-macro URIs. Special code has been added there to
handle that.
Change-Id: Iafbdc07a9fe925d9ee580d4f5778448f18f2ebd9
Convert code like:
0 == aStr.compareToAscii("XXX")
to
aStr.equalsAscii("XXX")
which is both clearer and faster.
Change-Id: I2e906d7d38494db38eb292702fadb781b1251e07
