Color tags like <a:schemeClr> can have children tags that modify the
specified color. These modifications were applied on import time in
the Color object, but they were not preserved.
We added a member to Color object to preserve the unaltered list of
transformations. The method getTransformations() returns that member,
and the methods getColorTransformationName and
getColorTransformationToken were added to transform the tokens into
strings that can be added to an InteropGrabBag and viceversa.
The transformations are added to the Shape InteropGrabBag in the
method Shape::createAndInsert, and they are written back on export
time at DrawingML::WriteStyleProperties.
The data files for some /sd/qa/ unit tests were updated to reflect
the new properties inside the Shape InteropGrabBag.
Change-Id: Ieb164268c3b79f2d9b7ed3a4954b5de3b7a5811c
The InteropGrabBag property adds a lot of noise to this tests and it
doesn't actually have any purpose, because we are building their own
tests for InteropGrabBag-supported document attributes.
We have removed this property by default but we left a flag to enable
it when it's convenient for debugging.
Change-Id: I97cc7ba171783222a57ab5a96e5755070b949505
Issue:
The <wp:align> is missing after roundtrip
XML Difference:
Original:
<wp:positionH relativeFrom="page">
<wp:align>
center
</wp:align>
</wp:positionH>
Roundtrip:
<wp:positionH relativeFrom="page">
<wp:posOffset>
0
</wp:posOffset>
</wp:positionH>
Conflicts:
sw/qa/extras/ooxmlexport/ooxmlexport.cxx
Reviewed on:
https://gerrit.libreoffice.org/7571
Change-Id: I3c8ef2f0ee3dd84a23fab197ab95f152b850067e
Application::InitSettings should only assert when
pSVData->msAppData.mpSettings is already set. There is an errant not
symbol.
Change-Id: Id18848774f3f3a1b37969f057eaf8685231ad291
Reviewed-on: https://gerrit.libreoffice.org/7711
Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com>
Tested-by: Norbert Thiebaud <nthiebaud@gmail.com>
Marking the cell dirty alone appears to be sufficient to trigger resetting
of number format. SetDirty() would mark it dirty *and* put the cell into
formula tree, which would prevent proper value propagation as seen in the
bug report.
Change-Id: Ie68f996112938fe286a9bd50c38404f9df6f4ca1
Using OSL_LOG_PREFIX in the exceptions is not a good idea; it blows the size
of the string literals tremendously - full build path for every line that uses
that, including the line.
Let's not remove it for good, but hide it for non-debug builds
(OSL_DEBUG_LEVEL == 0) so that anybody who would like to use this during
debugging could still do (but I doubt it has any value for anybody).
Change-Id: Icc8db95ae0862671a206e681f92c60cdf51ffc32
When the source and destination bitmap do not have the same size,
ImplDrawAlpha() does not use direct paint, but instead it gets the image from
the screen, blends it with the provided bitmap, and again draws it.
Unfortunately, the blending uses the most trivial (and ugly) way of scaling;
so to produce much better results, let's scale to the destination size before
even calling ImplDrawAlpha().
The sideeffect is that we use the direct paint in most cases now; so hopefully
it pays off to do the (a bit more expensive) scale first.
Change-Id: I3b6b275710220910709ae4345ad6be3d6e4bf79c
This was generated previously in writerfilter, but keeping around an XML
file of 8000+ lines just to generate these constants is an overkill.
Additionally, these constants are hardwired in all available WW8 files,
so they won't change. By copying it here, it's possible to incrementally
remove what's still remaining from writerfilter's not used doctok.
Change-Id: I1d30da1f635946b15a37de68080fd9eebc2ad88e
A LineProperties object can contain some preset dash defined by the
shape style and some custom dash defined by the shape properties. In
that case, the custom dash must be used.
The existing code only used the custom dash if no preset dash was
present.
Change-Id: I6254d95184312360e8782dba7329d5ec2f5441fa
