Age | Commit message (Collapse) | Author |
|
... b60ce8465c8f01242354abccebe00742d164af60
Change-Id: Iba1a059b3aeb5b2266398d80e6995f98fd580f14
|
|
Boost is clearly following the C++ tradition of surprising omissions.
Change-Id: I205ef17f87b176da938ebfa3e1a0748e94605daf
|
|
... if the parameter is currently in the same list.
The "boost::intrusive::circular_list_algorithms::transfer" has a
precondition that the 2 parameters must not be in the same list.
This causes an infinite loop in SwFindParaText::Find(), which is hiding
the infinite loop that i'm trying to debug...
While at it, remove some unnecessary complexity.
Change-Id: Ib41f52c6d5c44ecc358c6170ee1e6e98729e1302
|
|
Change-Id: I122f5f23e510ae42664d2b69504f4722aa05597e
|
|
Partially revert commit d30a8ec448bcd08c6a52a37d6ae41a4b71c235da
(Correct number format mapping for CJK numbered lists in rtf/doc/docx
filters., 2014-11-04).
Conflicts:
sw/qa/extras/ooxmlexport/ooxmlexport.cxx
Reviewed on:
https://gerrit.libreoffice.org/14211
Change-Id: I0876ecd982cfd8d97887495b23af0e0370821101
|
|
* pib structure in rtf shape is processed always, if exists, not only
for shape #75 (picture frame): it is possible for other shapes, at
least for #1 (rectangle)
* picture inside shape gets width and height from shape: this picture
looks like pib object and should fit shape frame
Conflicts:
sw/qa/extras/rtfimport/rtfimport.cxx
Reviewed on:
https://gerrit.libreoffice.org/14396
Change-Id: I5cf2e692940c4ddd8ab8b291ef65c0cc44c45f5d
|
|
Change-Id: Id56bc94e95e59f9bfe1dca02ff1f5e16844de45a
|
|
Change-Id: Ia9ac12c0d03d960aced5d7066306a6f68c95996b
|
|
The problem was that the bugdoc had a table, and inside the table there
was a long paragraph that flows to the next page, but only the paragraph
mark of it does so. We first split the frame to have space for the
paragraph mark, but later decided that all the content would fit the
first frame, and this way the last hard line break and the paragraph
mark was painted on each other.
This is normally not a problem without tables, because
SwTxtFrm::FormatAdjust() just calls SplitFrm(), sets its nNew flag to
non-zero make sure that later SwTxtFrm::_AdjustFollow() doesn't try to
join it, and we're ready. However, when the paragraph is inside a table,
then the paragraph was formatted multiple times, and next time when we
already had a follow nNew was not set, so even if there was a correct
split first, the new frame was joined later.
Fix the problem by explicitly setting nNew for the "in a table and
paragraph ends with a hard line break" case, that way we don't blindly
join the frame, only in case there is enough space for the follow in the
master.
Change-Id: Iede654740dcb0d8aa768d742ee330208291a383a
|
|
Change-Id: I29ca828a2049a72e2238986892c172e7c135b44b
|
|
Change-Id: I934ef3cbe67bba7d7439e464f71b9fd4a1da4e50
|
|
they should keep their control models
Change-Id: Iaa27cd8a9bc3a1165db665e0ad172da622e00c6b
|
|
Accept gl linkage in hope of future offscreen mesa support.
Avoid linking with GLX though.
Change-Id: I4e666f60e74fe34075a8da9eeba95807cf8fe38c
Signed-off-by: Riccardo Magliocchetti <riccardo.magliocchetti@gmail.com>
Reviewed-on: https://gerrit.libreoffice.org/13452
Reviewed-by: Michael Stahl <mstahl@redhat.com>
Tested-by: Michael Stahl <mstahl@redhat.com>
|
|
Deprecated comphelper/sequenceasvector.hxx.
Rewritten code using it. Using instead the functions containerToSequence and
sequenceToContainer, found in include/comphelper/sequence.hxx.
One class that inherits from it (in framework/inc/stdtypes.h), and
the code using that has been left.
Signed-off-by: Michael Stahl <mstahl@redhat.com>
Conflicts:
writerfilter/source/dmapper/DomainMapper.cxx
writerfilter/source/dmapper/TblStylePrHandler.hxx
writerfilter/source/dmapper/WrapPolygonHandler.hxx
Change-Id: Ice61c94ffb052e389281aebb7cdf185134221061
|
|
Change-Id: I273af8b570adfcb7bfb784495bc31d2f4f1ee00b
Reviewed-on: https://gerrit.libreoffice.org/14333
Tested-by: Michael Stahl <mstahl@redhat.com>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
|
|
and boost:make_shared->std::make_shared
Change-Id: Ic1e187c52c856a7b27817967b2caa8920f23a98d
|
|
Change-Id: Iaa880528cf3c899ce66e4349c6d989dfbe5cbeb6
Reviewed-on: https://gerrit.libreoffice.org/14346
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Change-Id: Ia112cf626126149ea9cf09c5d6ff812d5d5ffec5
|
|
- style:border-line-width's first value should be positive (follow-up
fix could be to handle such existing documents, but this is a test
document, so let's just modify it)
- style:contextual-spacing is written as loext:contextual-spacing today
Change-Id: I4463ad83d45bb85f39257cc5bfda96ae1ef2c2a1
|
|
Change-Id: Iedf8eacd37b8ed8e307a10e8ade32f53c7417c4a
Reviewed-on: https://gerrit.libreoffice.org/14306
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
1. Revert koreanDigital2 back to NUMBER_HANGUL_KO (import).
2. Map NUMBER_LOWER_ZH in Simplified Chinese document to
chineseCountingThousand (export).
Change-Id: I62ec255626c250802288bf7fbec7371076e727d3
Reviewed-on: https://gerrit.libreoffice.org/14155
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Using the DPI from Application::GetDefaultDevice() is a much better
idea, especially that now oox::GraphicHelper::GraphicHelper() and
oox::drawingml::DrawingML::WriteSrcRect() are in sync.
Should fix the testCropPixel() failure in CppunitTest_sw_ooxmlexport
that appears on HiDPI systems.
Also, fix all the rounding problems that now became visible when the DPI
is the same for both import and export.
Change-Id: Iceb34a8a5a1eaa8ce0824491521ad6b4d2f6949c
Reviewed-on: https://gerrit.libreoffice.org/14280
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
It was a problem since the initial commit
062eaeffe7cb986255063bb9b0a5f3fb3fc8e34c (sw: Improved document
comparison based on RSIDs., 2011-12-22) that this new feature -- which
is annoying for some use-cases -- could not be disabled, let's allow
that.
Change-Id: I33fa77382919586fb00198246f737caa68dcbd85
Reviewed-on: https://gerrit.libreoffice.org/14277
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
|
|
Change-Id: Iab7b4941e9b41249f2a30fa6139313239a98c501
Reviewed-on: https://gerrit.libreoffice.org/14276
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
In theory this is to be in sync with the ODF import. In practice the old
UNO property seems not to have a proper fallback to populate the doc
model with the fillattributes, so without this even if the import result
is visible, it would be lost on ODF export.
Additionally, this detected a bug in SwUnoCursorHelper::makeRedline(),
where paragraph format redline tried to use the map of a text portion
instead of a paragraph.
Change-Id: I026e38e1990ed2a460624a8d967a16ae3fb6c512
|
|
Regression from 7d9bb549d498d6beed2c4050c402d09643febdfa (Related:
i#124638 Second step of DrawingLayer FillAttributes..., 2014-06-02), the
problem was that exporters still expect an SvxBrushItem for the para
background color, while doc model was changed to have an XFillStyleItem
/ XFillColorItem pair instead.
Change-Id: Ib94fda103ec35a6f087307aafdd890183d9d935f
|
|
The problem is that SwUndoDelete will move the fully selected nodes to
the UndoNodes but it leaves bookmarks with their SwIndex pointing to the
deleted nodes. The SwNodeIndex are corrected by SwNodes::_MoveNodes()
so they point to a different node than the SwIndex.
This only happens if only one position of the bookmark is inside the
deletion range; if both are, the bookmark will be deleted by
SwUndoSaveCntnt::DelCntntIndex().
Also joining the 2 start/end nodes of the selection will accidentally
correct the bookmarks but only if it happens to delete the end node.
(and apparently there is also a DeleteRange method that doesn't join)
Change-Id: I91ec362bb833328f8d681fd9458cb915c4efb267
|
|
...by adding some further SAL_DLLPUBLIC_RTTI type annotations (cf.
b4f6b26b5a1a78fecfa95ec2eb7ac8b80495d8aa "SAL_DLLPUBLIC_RTTI for proper RTTI
visibility for LLVM") and by making sure relevant function types do not use
incomplete types in their parameter and return types (which would make the RTTI
hidden).
Change-Id: Id7aadcbc0704b9759968ae36266fc9ce11a2e340
|
|
If there is a SAXParseException, OOXMLDocumentImpl::resolve() should not
ignore it, because if it occurs in a substream some end tag handlers may
not have been run and the DomainMapper may be in an inconsistent state,
so continuing to parse the outer document is probably not a good idea.
Also add some exception mangling so sfx2 can present a useful error
dialog.
Change-Id: I169ba6db25f2ae264af08a64edf76a6bf6757f85
|
|
Change-Id: Ia23865a38eff10304a1c1748b2b4f17af7b4d3ae
|
|
LibreOffice recognizes MS Office Table Of Contents, but LO files
saved as .doc didn't show the Cross-Refence heading - it was blank.
Now bookmarks using the LO naming convention are also imported
as TOC bookmarks from .doc files.
Testcase included.
Change-Id: Icae5c0de31856ccc75ede9a97c0ea046dd97325f
Reviewed-on: https://gerrit.libreoffice.org/14018
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
note the two TODOs about suspicious casts
Change-Id: I324fa05b30a5c8aa1d9e9a6d488a1e295226e788
|
|
Change-Id: I67f472d72efe123b533d4d94be0084986c0e8349
|
|
Change-Id: Id3aec16d17b55366c2a35810b21b0a1c73439741
|
|
Preset shape text rotation values are read in a few places and stored in
moRotation. With these changes, moRotation is always read unchanged, but
multiplied by -1 when it's used.
Change-Id: I633d665c21daa69e15fa828a43300f10d2bf2054
Reviewed-on: https://gerrit.libreoffice.org/13561
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Change-Id: If83f4900fcecde07f75d88c7f0b7c9088974b093
|
|
Change-Id: If96b46bac7f8212cfb83dc8b4528986137f1eaea
|
|
We can't do anything sensible with these CustomXML elements but now we
have to handle them because.
(regression from 9dbf817fe5c5253fba0831aefa17575ae0ba3af1)
Change-Id: If4247890ff9961a77434587802670d28608a7922
|
|
since all of it's methods and fields are static.
Change-Id: Ia2b652969489c5a21e01a35cda65c6e84346aac0
|
|
Change-Id: I4d0e1de89d0bbdbea23bc5a46bf75ae0ce4e2796
|
|
Change-Id: Idae98f65bd87d08c62b27c5c604e4d98fcacbe1e
Reviewed-on: https://gerrit.libreoffice.org/13660
Reviewed-by: David Ostrovsky <david@ostrovsky.org>
Tested-by: David Ostrovsky <david@ostrovsky.org>
|
|
Change-Id: I57f90a84b49c4a3bd49c66e23d059939684a6aef
Reviewed-on: https://gerrit.libreoffice.org/13765
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
and a handful of other currently uncompiled spots,
after my commit 7f8f277b94704a289fbbd1b836e4e5d66311580d
"fdo#84938: convert STREAM_ #defines to 'enum class'"
Change-Id: I550f6fb850e1d71a6f08767eeb222a18071b89d5
|
|
Change-Id: Ibbf14c7e9a5c1883c1311d4c86f948f74f8e473e
|
|
use C++11s range-based for() loops instead
Conflicts:
sd/source/filter/eppt/pptx-epptooxml.cxx
Change-Id: I0868eb345932c05b7e40b087035da252b99bf0b9
Reviewed-on: https://gerrit.libreoffice.org/13714
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Reviewed on:
https://gerrit.libreoffice.org/13541
Change-Id: I40ba14503ad9f5458d8326741a4fe7a13653b05a
|
|
Reviewed on:
https://gerrit.libreoffice.org/13542
Change-Id: I8d6289b814a063ba8037d07a7600a44fa77f860d
|
|
... only in case the whole table is already selected
Change-Id: If7626954460e16945af6b21402a84e90c71ae138
|
|
This makes CppunitTest_sw_ww8export pass on Windows again.
Change-Id: I16fed4eabbe7b9ccdcc0c71361b85b0e13f2245a
|
|
And add a minimal reproducer that shows how the old way was broken.
Change-Id: Ic2dadf9905d603b0fd0573651b235ecd5dd70e71
|