Age | Commit message (Collapse) | Author |
|
Change-Id: I42119f656ca528286fb25d2d36c0af54b7d04a6b
|
|
Change-Id: I8cf274902bb5fda9fa70ab2af9e399db82d85d1d
|
|
fdo75692-2.xlsx and fdo#75692-3.xlsx crash.
Change-Id: I56353e7da1850a49e18d3a570641843600d34b1c
|
|
Change-Id: I692fa22132cd3a722b58de22e3dbb759ff888e5d
|
|
importExtDrawings() must be called as soon as possible,
before parser starts to parse the next shape.
Call it when graphicFrame tag is closed. This tag include
the reference to the SmartArt.
Plus fix up import tests.
Change-Id: I9e8d54c2b1afeb78a1122390dc4982d580c152ae
|
|
Change-Id: I54a51189e1c595841b8b02f3b4436da4a29f1dac
|
|
SmartArt import ignores some fragments during import if
drawing fragment exists, which seems to be not complete.
In this case font style is blank (white) in data (and drawing)
fragment and the real value is defined in the ignored color fragment.
So first make color fragment parsing work, then apply font
color of "node0" style on nodes of the SmartArt.
Actually, it's a workaround, because "node0" style label
is hardcoded, for a proper solution layout fragment should
be parsed too to get the right style label, but
it interferes with the drawing fragment by now.
Change-Id: I7db89176a07eee928563d42d3896fbd02190dfa8
|
|
Text list styles were copied, without proper
copy constructor and operator. It lad to mix
up list styles and so text font.
Change-Id: Iee7a6c0c1f74322fd7b80e41a262849f948e463a
|
|
Description:
In RT file the dash length (d) is going out of range,
as after RT the dashing scheme changes to custom dash
which was causing the corruption. Changed code at export,
which will divide the DashLen, DotLen and Distance by
base line width.
Reviewed on:
https://gerrit.libreoffice.org/9559
Change-Id: I0e644b5a2b692a9e717026a14d1f0058199f53b1
|
|
DOCX shape import normally works by oox creating the shape, then
writerfilter handling the shape text. For drawingML shapes, having shape
text, this a bit more complicated, as there are shape properties after
the shape text as well.
ShapeContextHandler::endFastElement() assumed that shape text is only
possible on css.text.TextFrame shapes: also handle shapes having a
TextBox as well.
sw/qa/extras/ooxmlimport/data/mce-nested.docx is a reproducer for this
problem (group shape missing), when TextBoxes are enabled by default in
oox.
Change-Id: I7a412b31965cf363da0b0c7fcc732741f2037542
|
|
Change-Id: Id3811dbe0cf2510ef6a851804b3886c14eca01b6
|
|
Change-Id: I9e34e14d1266310458bb491259e4bf9880e8a19f
|
|
Change-Id: I4bef14453d076f11066a695bc4a948cea5cfd40b
|
|
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
|
|
Change-Id: I171df68539fc41046b706157c04ab1e8cc1e60ca
|
|
Change-Id: Ifdbf7d69ecee226e551bc9c65de6c3a53acef53b
|
|
Change-Id: I5316693452427ed76a7738b090de023b110caa40
|
|
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>
|
|
bt shows:
Program received signal SIGSEGV, Segmentation fault.
0x00002aaadba213fb in oox::vml::InputStream::updateBuffer (this=0x8d7fd80) at /home/julien/compile-libreoffice/libreoffice/oox/source/vml/vmlinputstream.cxx:339
339 while( (mnBufferPos >= maBuffer.getLength()) && !mxTextStrm->isEOF() )
(gdb) bt
0x00002aaadba213fb in oox::vml::InputStream::updateBuffer (this=0x8d7fd80) at /home/julien/compile-libreoffice/libreoffice/oox/source/vml/vmlinputstream.cxx:339
0x00002aaadba21048 in oox::vml::InputStream::available (this=0x8d7fd80) at /home/julien/compile-libreoffice/libreoffice/oox/source/vml/vmlinputstream.cxx:326
0x00002aaacf5a0249 in sax_fastparser::FastSaxParserImpl::parseStream (this=0x89aea30, maStructSource=...)
at /home/julien/compile-libreoffice/libreoffice/sax/source/fastparser/fastparser.cxx:810
Indeed, mxTextStrm is invalid, so let's test its validity in InputStream constructor
Change-Id: Ifed79603e33b64d11eb07656df17824b7f98058f
|
|
Change-Id: If87cdfb2c605254f6d69baa4ca5aec09091caa68
|
|
Change-Id: I5002f3e935edcc9f09603a5b535e2b339ebed402
|
|
Change-Id: Ie4a0334ddb393726d982e9f4e51a45e391a1b1f0
|
|
Change-Id: Ia49f4e99e6663ea95dc85d4dd09e161413a2f419
|
|
Change-Id: Ia62459945cd45f493754a1412b74242d3994f7f0
|
|
Change-Id: Ibabd73d06135a3ee500ce9d52fef42caa3ad7f35
|
|
Change-Id: I4a74eb76e5fcda915e5d12257fedf3ee84b62baa
|
|
Change-Id: I6fc18afd6189060de6943b003dea933713e19773
|
|
Now that we have default values for Exception constructor params,
remove lots of boilerplate code.
Change-Id: I620bd641eecfed38e6123873b3b94aaf47922e74
|
|
Change-Id: I05aa9a5ad5d97a1ead3d30298f45e7aeb62d0652
|
|
When two pictures apply different effects to the same picture, it is
only saved once in the original document. Added a cache to DrawingML
to know if the picture has already been exported, and added a test
for it.
Change-Id: Ia25f3d8f2f46d61f18aefc22fdfdbcdc72f2d916
|
|
When Word applies an artistic effect, it creates two embedded files;
one contains the bitmap with the effect and the other one contains the
original bitmap to be able to undo the effect.
This patch reads the original bitmap, stores it in the shape grab bag
and saves it back to the docx file. Added unit tests too.
TODO: right now, if two effects point to the same original bitmap it
is stored twice, we should improve this.
Change-Id: Ia72034a257739abe4ffafa0f42b2a912e4bf9436
|
|
Bitmaps can define artistic effects like in the following example:
<a:blip r:embed="rId5">
<a:extLst>
<a:ext uri="{BEBA8EAE-BF5A-486C-A8C5-ECC9F3942E4B}">
<a14:imgProps
xmlns:a14="http://schemas.microsoft.com/office/drawing/2010/main">
<a14:imgLayer r:embed="rId6">
<a14:imgEffect>
<a14:artisticMarker trans="14000" size="80" />
</a14:imgEffect>
</a14:imgLayer>
</a14:imgProps>
</a:ext>
</a:extLst>
</a:blip>
LO core doesn't support them, but I'm preserving them using the shape
grab bag. Bitmaps must not be transformed to a SwXTextGraphicObject
so the grab bag of the XShape is not discarded.
Added several Context and Properties objects on the import side to
traverse and save the relevant tags, and added the corresponding code
on the export side to extract the grab bag and output the effect back.
Also added a unit test for a selection of artistic effects.
TODO: Word saves the original bitmap as an embedded wdp file so the
effect can be undone. We must preserve it too and add the reference to
the a14:imgLayer tag.
Change-Id: I61d427f83e4c8f353eb073da0114cd73ba50ba4b
|
|
Added the artistic effects defined in the MS-ODRAWXML extension. Ref:
http://msdn.microsoft.com/en-us/library/dd905216(v=office.12).aspx
Change-Id: I24c5e5bbe569f160d300a7a7293ca047cb7a1b88.
|
|
Groupshape text is provided by editeng, redlining is available in Writer
text only. But still, instead of ignoring both deletions and insertions
in group shape text, we could at least show the latest version
correctly, i.e. ignore deletions but import insertions.
Change-Id: I12df2ca1a8d27f9496a8036e521f8820d7075053
|
|
Change-Id: Ia1614759558e73686fce137eb52fb623eae4a4e5
|
|
Change-Id: Iba7bb88139a97e919739363cd85f3e822062b694
|
|
This time we did preserve this info, and we should not.
Change-Id: If2ddd7a266cd7da5642dc234411ce34e94bd3384
|
|
Problem Description: Docx file containing a chart (line chart / scatter chart)which has used a builtin
marker'x' gets corrupted when we save it in LO.The reason was that while exporting LO was writing the
marker information 'x' as 'X' which MS Word doesn't recognize.‒<c:marker><c:symbol val="X" />. Also
the size of the marker was coming 1 less than the actual value. Ex: if size is 7 then it was being
written as 6.
Solution: During export I have made changes so that now LO writes 'x' in the tag information
‒<c:marker> <c:symbol val="x" />. Now the size of the marker is also being correctly exported.
Change-Id: I26b747f9576625bf3acb941322ae418a0bbc6b64
Reviewed-on: https://gerrit.libreoffice.org/9273
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
While export, locked Canvas is missing and drawing is
exported inside textbox. However a locked Canvas has to be exported
inside a text-box and drawing should fall under locked Canvas
for the RT file to work in MS Office 2010.
Reviewed on:
https://gerrit.libreoffice.org/9377
Change-Id: Iea2c411302b552db18527001cc4c6d0290085afe
|
|
Change-Id: Iaf33c536c86a372aed2a719b6f87a8e03a5481ea
|
|
The original file contains two text boxes one contains image and the other contains chart.
Image in text box is not getting imported in LO which leads to no contents inside the a:graphicData tag which is causing the corruption.
Root cause is found in ShapeContextHandler::getShape.
mxChartShapeContext.is() returns true even when mnStartToken is not set as NMSP_dmlChart which is causing the issue.
I have added one more condition to handle this.
Change-Id: I6c567d7618b34c1a24f6809801e4460af6894c67
Reviewed-on: https://gerrit.libreoffice.org/9363
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Change-Id: I42787db31769e8bde984c5f4f0aa90335e889b1c
Reviewed-on: https://gerrit.libreoffice.org/9356
Reviewed-by: Thomas Arnhold <thomas@arnhold.org>
Tested-by: LibreOffice gerrit bot <gerrit@libreoffice.org>
|
|
It seemed a bit pointless to waste CPU cycles on PNG-compressing a
bitmap image only to later then uncompress it anyway. vcl's PNG
writing code showed up as 13% on the time profile of TiledLibreOffice
when displaying a document full of SmartArts.
Miklos suggested I try using SVM (which I guess means "StarView
Metafile") instead. When using SVM, no rendering of diagrams to
bitmaps during loading is done, but the diagram stays stored in a
resolution-independent (vector-ish) form. Which means it will be
rendered nicely and crisply regardless of the zoom level.
At least, that is my understanding, and experimentation (on OS X and Linux)
seems to confirm.
ce8c0ff07559ddcc729bffd7a68f4c6f281882e3
Change-Id: Ice8c0ff07559ddcc729bffd7a68f4c6f281882e3
|
|
If a picture contains some 2D (glow, shadow...) or 3D effect
(rotation, extrusion...), we prevent the importer from transforming
it into a XTextContent so the XShape grab bag is not removed and
the effects are preserved using the existing mechanisms. Added a unit
test for this issue, and modified some existing unit tests to match
the new behaviour.
Change-Id: I3b87069ea208604383a592d34d0a4ceb6b0f9fc7
|
|
Transformed the preservation process of shape effects to be able to
store more than one effect. For that we:
* Created the Effect struct and added a vector member to the
EffectProperties struct.
* Changed the shadow effect to use the new Effect struct,
EffectShadowProperties struct is preserved because the direction
field still has some use but we should remove it.
* Changed the structure of the grab bag to store more than one effect.
* Modified an existing unit test to check shapes with several effects.
Change-Id: I0dd908fa1d9578827c02ef6272fc9e2b914391be
|
|
This was actually a bug. It didn't matter much because a document with
an a:sp3d tag but without a:scene3d section would be invalid, but the
code was logically wrong.
Change-Id: Ifa838e425849642c2a1bf6fca6b6a8dc8ed3b465
|
|
Change-Id: I3e51a62710bb46c8255fd228d41d9300c90a1fb5
Reviewed-on: https://gerrit.libreoffice.org/9360
Reviewed-by: Thomas Arnhold <thomas@arnhold.org>
Tested-by: Thomas Arnhold <thomas@arnhold.org>
|
|
Change-Id: I00a4ebdcfc278274a567403bc1ccb5332ada836f
|
|
Implement a clang plugin to find them, and clean up existing code
to pass them by reference.
Change-Id: If642d87407c73346d9c0164b9fc77c5c3c4354b8
Reviewed-on: https://gerrit.libreoffice.org/9351
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>
|
|
If it wasn't for the desperate need to minimize code size for LO-based
iOS apps, I definitely wouldn't use -Oz in gb_COMPILERDEFAULTOPTFLAGS
(in solenv/gbuild/platform/IOS_ARM_GCC.mk).
Change-Id: Icb231b5f30b32afe1c0f3b43e5744bf0e41e96df
|