Age | Commit message (Collapse) | Author |
|
The export -> import part of this file took 72s on my machine, now it
takes 4s.
Change-Id: I693a922027a7285b5e967bbb2bcf8289728eae82
|
|
Re-enable the unit test with the fixed xPath expression.
This reverts commit 0e5f9958a57a68a3707db5b1eb9f82e0326327c1.
Conflicts:
sw/qa/extras/ooxmlexport/ooxmlexport.cxx
Change-Id: Ibc848b2ae89b3fb7f009cd984d0ecc5fc165673a
Reviewed-on: https://gerrit.libreoffice.org/8679
Tested-by: LibreOffice gerrit bot <gerrit@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
- m_oldTableReference values were not being reset.
- A call for DocxAttributeOutput::switchHeaderFooter from
DocxExport::WriteHeadersFooters would end up using the m_oldTableReference
value saved by DocxAttributeOutput::EndParagraph, which is incorrect.
- Fixed this issue and added an UT for the same.
Change-Id: I5ac51c3449e78bd26389e93e0d733d429dcc6874
Reviewed-on: https://gerrit.libreoffice.org/8659
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Change-Id: I637cec62eb6cf951cb3d7c9e83c896494d6e1701
|
|
MS Office has an internal limitation of 4091 styles in styles.xml
and cannot load a .docx with more. However the documentation seems to allow
that. So if there are more styles, don't export those.
MS Office follows the same principle and repairs the corrupt document by
removing extra styles.
Change-Id: I20c8775ee9d697a6613be96eb01283844c1d78f5
Reviewed-on: https://gerrit.libreoffice.org/8653
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Change-Id: I932440ae95a6e7599168e88e9a17686b67763247
Reviewed-on: https://gerrit.libreoffice.org/8667
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Change-Id: I45d81df485ff62417f94c01234c3ddc41b6f3c77
Reviewed-on: https://gerrit.libreoffice.org/8665
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Reviewed on:
https://gerrit.libreoffice.org/8664
Change-Id: I719fe3266b80fd6e4da73729c74f3d606708b691
|
|
Change-Id: Ib9ca82b1b5f19dff439cd3119781a5f025b45231
Reviewed-on: https://gerrit.libreoffice.org/8663
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Change xPath as suggested by Luke Deller in
https://gerrit.libreoffice.org/#/c/8572/.
Change-Id: Ibcf2bc6cd2b35df501aeeaf143a3928603cf6038
|
|
SDT blocks can contain document properties, like the author. This is
done with the dataBinding tag inside sdtPr.
Added a resource definition to the ooxml model file, so the parser can
process this tag. The attributes are stored in the existing grab bag
used for other sdt properties and written back on export.
The code to write the SDT block in the exporter has been modified to
be able to extract the new tag from the grab bag and write it to the
document.
Added a unit test.
Change-Id: I18f35a61784eb0a03ab9ca2cc398ca0df289956c
|
|
Change-Id: Iefd8de05107c6e9ad1dec04711ee3fd8e4c05b17
|
|
The exporter writes the sdt blocks enclosing an entire paragraph, but
that's not the proper behaviour in all cases. The documents that mix
checkboxes and text in the same paragraph export the text inside the
sdt box and that's incorrect.
We have added code to be able to write sdt blocks that enclose
paragraphs or text runs, depending on the type of block. We have
applied it to checkboxes now.
There are two sets of properties in DocxAttributeOutput that store
sdt properties for the text run and for the paragraph. We have
modified the method WriteParagraphSdt to be generic enough to support
both cases. To write the sdt block enclosing the text run we used
parser marks.
We have renamed the property id PROP_PARA_SDTPR to the more generic
name PROP_SDTPR.
Checkbox unit test was modified to match the new structure of the
generated document.
Change-Id: I81ffe0062e1a5f80fc4638f7ee454a9bc18e63ee
|
|
Checkbox controls in docx are implemented with a sdt block containing
the property w:14:checkbox, which has several child elements.
To implement the preservation of the field, we had to add several
tags to the ooxml model file. Notice they belong to namespace w14,
that's why we had to add a redundant definitions of CT_STRING,
ST_STRING and their actions. Then, we just used the existing
mechanisms on the importer and the exporter.
Finally, added a unit test.
TODO: the checkbox is preserved on export but it cannot be modified.
We should import it like we do for combo and date boxes.
Change-Id: I40f29b48ec24719c48482b9c8cff10176d1b188d
|
|
This patch strictly reuses the code introduced at commit
55211e612d2cbed03dd81c039d07ea4e936c2804.
Change-Id: Ie9151c81c2cb63d0f4a320293b6e5d27f20d4043
|
|
Change-Id: Id24cfcd34355ed39f80baa9e7b90ab7705e852a2
|
|
Change-Id: Ifc003e3e986fbe79f8aff5a9edb05a7f22e0f24b
|
|
Change-Id: Ic5bd2ecf827be68510877589b64305b5704d7488
|
|
Change-Id: I8a94b7ab0fb7ca30da1665ae4745aafbaeed687b
|
|
Change-Id: If57e5e4459bc8e07d980c0f5d68151e26b9b8bf2
|
|
Change-Id: I4d0e9057007aebfe9ed37899b36f8db6f51d95d2
|
|
Change-Id: I3b66764c6c783d2a076d6b201be2d4473891bd83
|
|
Change-Id: I605a8c7f0dc1519a70db1fc6d6dd9acff32e3bd5
|
|
Change-Id: Ie7ab3fa80cb7231cee1f741f53af541633fa427e
|
|
Change-Id: I3e2ad2d3dd1713336b1a01cb7a2b1ff28bb833ea
|
|
Change-Id: I7d180e0386d50f0a943b68494a9458016a6a51a4
|
|
Change-Id: Icd1922a47ea3b59bdec0ef21481df5f4a969a4e0
|
|
Change-Id: I83b2335ffe215c36d5d69bec7489f8f025d6b930
|
|
Change-Id: I466c795b53a0e9b9f282ff225134d2f6e2edf52c
|
|
Change-Id: I454f350863a540c2e686550e3e7ff16b3a33ba19
|
|
of a set of paragraph which have more than one different List Style applied
create a new List Style and put the paragraphs into a new list.
(cherry picked from commit 0087ca89e3905009ed947c651f3dc70f3d61ea93)
Conflicts:
sw/inc/doc.hxx
sw/inc/docsh.hxx
sw/inc/editsh.hxx
sw/source/core/doc/docnum.cxx
sw/source/core/docnode/ndcopy.cxx
sw/source/core/edit/autofmt.cxx
sw/source/core/edit/ednumber.cxx
sw/source/core/uibase/app/docst.cxx
sw/source/core/uibase/docvw/edtwin.cxx
sw/source/core/uibase/inc/textsh.hxx
sw/source/core/uibase/shells/listsh.cxx
sw/source/core/uibase/shells/textsh1.cxx
sw/source/core/uibase/shells/txtnum.cxx
sw/source/core/uibase/uiview/formatclipboard.cxx
sw/source/core/uibase/uiview/view2.cxx
sw/source/core/uibase/wrtsh/wrtsh1.cxx
sw/source/core/undo/unnum.cxx
sw/source/core/unocore/unocrsrhelper.cxx
sw/source/filter/ww1/fltshell.cxx
sw/source/ui/misc/num.cxx
Change-Id: Iadb5b386cada296d90d96aa75574024baac7ae3e
|
|
There are two issues that are handled in this fix.
1)File created in MSO2K10 on RT gets corrupted.The root cause is found in CustomShapeProperties::pushToPropSet
2)File created in MSO2K7 on RT gets corrupted.There is an issue in shape import <a:gd> values for any shape (ex. circular arrow, hexagon etc).LO cannot import right <a:gd> values for any shape which is created in MSO-2007.Due to missing values of <a:gd> tag, after roundtrip the file gets corrupted.To avoid corruption a check introduced:-
if(aAdjustments.size() == nLength)
after http://opengrok.libreoffice.org/xref/core/oox/source/export/drawingml.cxx#1784
that will verify the number of <a:gd> tags associated with the shape during import and the number of <a:gd> tags during export.If there is a mismatch <a:gd> willnot be written.Changes made in DrawingML::WritePresetShape.
I have written 2 test cases for the same.
Conflicts:
sw/qa/extras/ooxmlexport/ooxmlexport.cxx
Reviewed on:
https://gerrit.libreoffice.org/8657
Change-Id: Ibb1e2dc1c098b7399c06d7b4f59fac4e80887062
|
|
Issue :
DOCX containing nested Hyperlinks, have mismatch xml tags.
End tag for outer Hyperlinks does not get closed.
Implementation :
1] Added code at Export side to maintain Count of
Hyperlinks started and ended.
Added check endParagraph() to to close Hyperlink tag
if not yet closed but started.
2] Written Export Unit Test case.
NOTE : Corruption was occurring on "Without debug-util" build and Crash was occurring on "WITH debug-util build".
Reviewed on:
https://gerrit.libreoffice.org/8654
Change-Id: I9a0ab48578f3d5f4a13c615f4e42a69e790d3ced
|
|
Change-Id: I26d27bfaafb83ee3ae74ce64bf72146b157ed7db
|
|
Reviewed on:
https://gerrit.libreoffice.org/8623
Change-Id: I042eeb0829c55d6a0fd5d85c1712c7b1cddc8c96
|
|
The IA2-integration added some special cases for calculating textbounds for
attributeruns when automatic spellchecking is active. This does not seem to
effect linux but messes up accessibility on mac. This patch removes the special
cases avoids the problem on Mac. When daily builds gets available for Windows
I will do some testing there to see if I can find any problems.
Change-Id: I203da4fe1c6e5d4915d19e1704073e58f8aa88c9
Reviewed-on: https://gerrit.libreoffice.org/8591
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Regression from commit 75934fc178ed90b56d77eb1efebed5bdac8427b6 (VML
import: handle shape with text inside groupshape as TextShape,
2014-01-17), the problem was that TextShape is also a kind of rectangle,
so that shouldn't be used for all shape with text in groupshapes, just
for v:rect elements.
This fix gives the required text wrapping for the referenced bugdoc
without turning the rest of the shapes into rectangles.
Change-Id: I6e343c65a869e335b641728adbc7984ee85377b5
|
|
Change-Id: I55fdcdfb25a2d6456eff044eb4af8fa00905d6e8
|
|
Change-Id: I73acb3a150b01114d32274a6842f6db9654b3e63
|
|
Change-Id: Ief810c49b821e2e60ee2c8b14187dd8d7d81c461
|
|
And mirror the right and bottom borders while I'm at it. I guess it was
originally meant to.
Change-Id: I408b077c1524d19bbadd6f0b284ce204064eb735
|
|
Change-Id: Icb11a2e2f802cbf2af2362315f3acbc66f15334d
|
|
Change-Id: I9467ae2ba5155175d5be8218248edada136a5ca1
|
|
So that e.g. parseDump() in case of an incorrect xpath does not just
plain give an equality assertion failure, which on its own is confusing.
Change-Id: Ib9fc2b4f6720b5c472773afbe8dc18cd8961465c
|
|
Change-Id: I4d36e4b86c77a7356a8c221cbfc5735e925392ba
|
|
Change-Id: Ic416289d93f9e2a5bc0b3a1016d6222cfc724912
|
|
Change-Id: I3773f5432ff4e2a005e0c67d9c65fdb00de6f16b
|
|
The closed borders connect much better this way. Now all double lines look
at least half-way decent.
Also, we shouldn't be doing any sort of pixel correction in the common code
that gets used both raster and vector drawing modes. In case we need to do
some pixel correction, we should do it in the pixel processor code.
Change-Id: I78e245c10752d0f4790a9ac664a14ce06571434e
|
|
Change-Id: I9d27359fe2c343593455eebf69bca6efdb7c5079
|
|
1) Offset top and left border lines only for double lines.
2) Don't apply zoom level scaling to border widths. That's used for line
pattern spacing only.
Change-Id: I12bdd1492531cc6651e22955c3dca4bdda0f1e4c
|