Age | Commit message (Collapse) | Author |
|
all SfxFilter instances should now be hold inside of a std::shared_ptr.
This fixes a number of huge memory leaks in the test framework and
removes one huge source of memory issue in sfx2. SfxMedium contains a
pointer to the SfxFilter but does not own. Therefore it is required that
any SfxFilter belonging to a SfxMedium lives longer. However this seems
to work mostly by hoping that all SfxFilter instances are stored in a
global array. As we have seen with the tests this is not true (there are
also some cases inside of sd that seem to not follow that pattern as
well).
Change-Id: I12fd04a504cc4efc0a94967abd91c6fe2c6a8ce8
Reviewed-on: https://gerrit.libreoffice.org/23140
Reviewed-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
Tested-by: Markus Mohrhard <markus.mohrhard@googlemail.com>
|
|
With this, it can't happen that "paste" returns an error, but "paste as
rtf" succeeds (from a classification point of view).
Change-Id: Ia6807e5a6c065557cdd735a574ea858d29a97160
|
|
So that input filters can build their own doc properties and can also
use CheckPaste() without having the full original document around.
Change-Id: I0b8a63702f73ce04e8728a360b56ab9d1ca67af1
|
|
(as some tests derive from the latter only for the Directories part, not for the
setUp/tearDown overrides: those tests will be cleaned up next)
Change-Id: Ib6b78eea868b8bc21d4cc6e8fd9e1d025deca05f
Reviewed-on: https://gerrit.libreoffice.org/23078
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Change-Id: I07dfd626f7d248e264fd1724f6f2bd1241be39d0
|
|
With this, user-defined document properties are imported from RTF.
Change-Id: I8dfb8e802bd26906827620550d6f5d88f047d364
|
|
Change-Id: Ic3ac02fe28bee0ea3189fb2c9d3548f5ca13fe5c
|
|
Change-Id: Ic2e08245dd022555ad6308283d406d81141a9124
|
|
Need to store the document twice to actually reproduce the bug.
Change-Id: I2332bce1e30546b3a9fe2ec020f2eea8e8d0da58
|
|
Regression from commit e2b260fc98e833d4e64426b90992094f2da0498c (sw: let
layout not mark embedded object as modified, 2014-06-03), an infinite
loop was caused by:
1) SwDoc::SetOLEObjModified() triggering the maOLEModifiedIdle Idle
2) which at the end called SwWrtShell::CalcAndSetScale()
3) which at the end called chart::ChartModel::setModified()
4) where chart code called back into SwDoc::SetOLEObjModified() via the
modification listener, and this happened again and again.
The original fix wanted to avoid marking the document as modified
without a user interaction, so fix the bug by only calling setModified()
if it prevents a not-modified -> modified transition. This keeps the
original bug fixed, but prevents the infinite loop, that is always a
modified -> modified transition.
Change-Id: I3b56a91afaacd3e0b7cb646a492fd15f1b5168ee
|
|
Change-Id: I57f4dc90bc2ad8488af8cd42515b195f4320a9d9
Reviewed-on: https://gerrit.libreoffice.org/22939
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
|
|
If the Basic library is newly created it may not have been stored to the
document storage yet.
(regression from 5246fa262450f686674850c53df666422f441c86)
Change-Id: I9c2fc1d7446795b9c1c2224671118b2a671dcad8
|
|
Change-Id: I37afde801a6468c79259fe51acbdea0614c2c8c5
|
|
The RTF spec is quite terse on how the form filed result should be
interpreted, but the binary equivalent documents properly that
checkboxes have 3 valid states: 0, 1 and 25, the later meaning
undefined. Use the default value in that case.
Change-Id: I672bf8d1f63d7880227b7fa7b5c81f91e1877b2a
|
|
Change-Id: Ie540aca97cddbd657faae5d15293a922ba5196d5
Reviewed-on: https://gerrit.libreoffice.org/22702
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
|
|
Change-Id: I6e22654d78ba8e043bab1009a295f078f9666000
|
|
Change-Id: Ia152f9cd97ead8e4289541ae6cd3e21a9635ffe0
|
|
And all those places that interface to SvxSearchItem.
Change-Id: I08915824a596cd0f247a89f44e4684cd22b98082
|
|
Change-Id: I5a2a465bec4fe38de683ce989264062a70a12d36
|
|
On one hand, don't handle a fieldmark for it in dmapper.
On the other hand, handle the field in the RTF tokenizer as it would be
{\pict ...hexdump... }, that will result in an inline picture, as
wanted.
Change-Id: I554fdf017920350144300fd86617bf74eed8995b
|
|
Change-Id: I34d7cdc7146c8e4d89b44a8a2824cd69ff9d7d6a
|
|
Change-Id: I5c3a4ab39768393090ca0ddc3cf3efe600184fcc
Reviewed-on: https://gerrit.libreoffice.org/22606
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com>
|
|
Change-Id: I899b57a485b83cd7c10ddc5192c92b4c87fe583a
Reviewed-on: https://gerrit.libreoffice.org/22582
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
|
|
stage 2 of replacing usage of various checks for the windows platform
with the compiler-defined '_WIN32' macro
In this stage we focus on replacing usage of the WIN macro
Change-Id: Ie8a4a63198a6de96bd158ecd707dadafb9c8ea84
Reviewed-on: https://gerrit.libreoffice.org/22393
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
|
|
Change-Id: I2a33fa5df5afed804adcbf904d5b47a507924034
|
|
Because the writerfilter::ooxml::OOXMLFastContextHandlerShape
constructor only sent the input stream (and not the full media
descriptor) to oox::shape::ShapeContextHandler, it's startFastElement()
tried to reconstruct a media descriptor from just the input stream. As a
result, the base URL of the document got lost at the time the
oox::drawingml::BlipContext constructor tried to call getAbsoluteUrl(),
to convert the relative URL to an absolute one.
Fix the problem by sending not only the input stream, but the full media
descriptor to the shape importer. As a bonus that makes code simpler,
too.
Change-Id: I1f06c04d0745aa7e2e06d4848cf454790d5073ca
|
|
Although LO was exporting whether an image was flipped
horizontally or vertically, it wasn't reading it in during
an import.
Change-Id: Ia85e3459d00f40d852bbcd3fcfe15e7b53d518ba
Reviewed-on: https://gerrit.libreoffice.org/22344
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Justin Luth <justin_luth@sil.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
|
|
Commit ca80f73 made it possible for one of dimensions to be zero.
This commit goes further, allowing opening real-life documents
with graphics having both width and height equal to 0.
Thanks to libreoffice@arbruijn.dds.nl for debugging and initial
patch!
Change-Id: I96a13b776adfd9fe46fc2c7691eb7904400c20a1
Reviewed-on: https://gerrit.libreoffice.org/21287
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Chris Sherlock <chris.sherlock79@gmail.com>
|
|
1. Export pStyle because writerfilter regard a list as an outline
only if associated pStyle is set to "Heading N".
2. Export English style name to ensure style names created by LO
with non-English UI is always recognized.
Change-Id: I88c590e4bae6cf30e78ebc1011023461eac4e883
Reviewed-on: https://gerrit.libreoffice.org/22086
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Mark Hung <marklh9@gmail.com>
|
|
1. Allow character entity ( &#nnnn; ) to exceed 0xffff in HTMLParser::ScanText()
2. Return a character as sal_uInt32 ( utf32 ) instead of sal_Unicode ( utf16 )
from SvParser::GetNextChar().
Conflicts:
sw/qa/extras/htmlexport/htmlexport.cxx
Change-Id: Ida455040970fae800f0f11471b27f53461fb78e4
Reviewed-on: https://gerrit.libreoffice.org/21152
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Mark Hung <marklh9@gmail.com>
|
|
After the import some of these are kept in RUNNING state. For Math
objects imported from MathType3 OLEs in particular, first a new
Math object is created and stored to the XStorage, only then is the
MathType3 stream imported. This means the Math object is modified and
contains data that must be stored.
The problem is then that SfxObjectShell::ImportFrom() simply calls
setModified(false), clearing the flag without storing the object.
For Flat ODF export we lose all the objects that are cached in sw's
SwOLELRUCache; for the bugdoc something more inexplicable happens for
ODT export where we lose "Object 214" (which is the first one in the
cache) but no other ones.
(The main difference is that for ODF there is an optimization to copy the
embedded object's storage without loading the object, but for Flat ODF
every object must be loaded and exported.)
(regression from 83777cd6e0f3f1a4458af896fd13344c696ecb1e)
Change-Id: Id1474fba9f4da2d5247c7ff4dc6819ddb9829fe8
|
|
... is reverted in case the test function throws an exception, so that
subsequent^Wfollowing tests don't inherit the changed configuration.
Change-Id: I748f9edf15a7f860607ae4cce891450db254c73e
|
|
... in the same paragraph. The intention was to avoid replaying the
supertext buffer and using its contents as a custom footnote mark at the
same time. However, it's enough to check if the buffer is empty to do
so, and that avoids the mis-import of the bugdoc as well.
Change-Id: I1e3b0c7f0c6d8eb8250d8b1d0d7d196039c40e79
|
|
A ridiculously fast way of doing this is:
for i in $(pcregrep -l -M -r --include='.*[hc]xx$' \
--exclude-dir=workdir --exclude-dir=instdir '^
{3,}' .)
do
perl -0777 -i -pe 's/^
{3,}/
/gm' $i
done
Change-Id: Iebb93eccbee9e4fc5c4380474ba595858a27ac2c
Reviewed-on: https://gerrit.libreoffice.org/22224
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Chris Sherlock <chris.sherlock79@gmail.com>
|
|
In the sequence
1 Title 1
1.1 Subtitle 1.1
Not numbered item
1.2 Subitle 1.2
2 Title 2
The numbering element for the Subitle 1.2 was missing the html start attribute
when it was followed by a numbering element of lower level.
So let's force it, when it comes after a not numbered item.
Adding a test case that validate that the Subtitle now contains the right attribute
Removed a reference to an openoffice bug that brings no value too
Change-Id: I928592a667f59128a4aa079d6b042ffb14244181
Reviewed-on: https://gerrit.libreoffice.org/22191
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
using an idea from dtardon:
<dtardon> noelgrandin, hi. could you try to run the unusedmethods clang
plugin with "make build-nocheck"? that would catch functions that are
only used in tests. e.g., i just removed the whole o3tl::range class,
which has not been used in many years, but htere was a test for it...
<noelgrandin> dtardon, interesting idea! Sure, I can do that.
Change-Id: I5653953a426a2186a1e43017212d87ffce520387
Reviewed-on: https://gerrit.libreoffice.org/22041
Reviewed-by: Noel Grandin <noelgrandin@gmail.com>
Tested-by: Noel Grandin <noelgrandin@gmail.com>
|
|
Change-Id: I4bb19d6103c4a6a902d86b62a857e3478493924c
|
|
Change-Id: Ice72f8d9971e15dd6ef365e64cd567b8581a92d3
Reviewed-on: https://gerrit.libreoffice.org/21797
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Ashod Nakashian <ashnakash@gmail.com>
|
|
Change-Id: If331fb2b0c03cd153387ddad7b33bbc28727ac11
Reviewed-on: https://gerrit.libreoffice.org/21995
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Michael Stahl <mstahl@redhat.com>
|
|
Change-Id: I8afd1db3859b99cf09d0d52e999bbd95cd88ed10
|
|
When we export the customshape as polypolygon, we need to
take into account, that the polypolygon of the shape already
has flipped coordinates, but coordinates has to be relative
to the rotated bounding rectangle.
Change-Id: I4f3b015e439d3d6041adc67eddd460f5bd4dfca1
Reviewed-on: https://gerrit.libreoffice.org/21905
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: Andras Timar <andras.timar@collabora.com>
|
|
Change-Id: Ie4554214e475da1bddba7ab0694ce23d1d93f06f
|
|
... attributes that happen if both CharHighlight and CharBackColor
properties are used, because the CharBackTransparent property wasn't
taken into account, and combining the CharBackColor and
CharBackTransparent properties happens *after*
XMLTextExportPropertySetMapper::ContextFilter() runs.
Also, it looks like a transparent highlight wouldn't export properly but
apparently DomainMapper::getColorFromId() won't create such.
(regression from f880962f5bf26bfaef06bd3f9e67e2d901a2e74c)
Change-Id: Ib628ef8bb377482f74fadb97c81afb95fbbf7184
|
|
Change-Id: I2bc99b4cbb86a368b6218b3c1a7dfc693858d1c9
|
|
... paragraph, it will be copied to every following paragraph.
(Before commit 2b78f2cd7b9e4bab0f3b3b9119238f36a1bbc7b2 the properties
of the end-of-section paragraph would be overwritten by the properties
inside the header)
Change-Id: Ibe095cf873a09ef9e12f898d09b5fd6f61c914c9
|
|
v2: Deleted CheckTable.java in JunitTest_sw_complex.mk
Change-Id: I737a887d6f48fb2cacbe893289a430dd8312b3d8
Reviewed-on: https://gerrit.libreoffice.org/21949
Reviewed-by: jan iversen <jani@documentfoundation.org>
Tested-by: jan iversen <jani@documentfoundation.org>
|
|
Change-Id: Icb94cdf3601d6d985ea147e3c80eadba2999ccee
Reviewed-on: https://gerrit.libreoffice.org/21877
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: jan iversen <jani@documentfoundation.org>
|
|
Change-Id: Ie091a38b34f14f6f641d591092bcbc38b0c548f6
Reviewed-on: https://gerrit.libreoffice.org/21910
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
|
|
Change-Id: I590d015aa50350cceef70313498e72474d242789
Reviewed-on: https://gerrit.libreoffice.org/21897
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins <ci@libreoffice.org>
|
|
Change-Id: Ib8cfaccaaf5bf75169c96e46d95ff118cc850f79
v2: Deleted whitespaces
Reviewed-on: https://gerrit.libreoffice.org/21886
Tested-by: Jenkins <ci@libreoffice.org>
Reviewed-by: jan iversen <jani@documentfoundation.org>
|