Age | Commit message (Collapse) | Author |
|
Regression from commit d620629521aee48dc820b6970ca446c329d0b09a (Use
constants for ST_LineNumberRestart values, 2014-10-01) where I forgot to
adapt the constructor's initializer list to also use the new constants.
(cherry picked from commit dc8a82488d4a8e5192d3cde43c10965fa89adf3d)
Change-Id: I7113d23cceb3a00f0d466b63c99a18bf3dfd0ee3
Reviewed-on: https://gerrit.libreoffice.org/64517
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: If18c80fc64e55d797953e24e40e5d5e62bd9c625
Reviewed-on: https://gerrit.libreoffice.org/63453
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
(cherry picked from commit 492ea7e08571e466e37f870b7642a79df55c2e92)
Reviewed-on: https://gerrit.libreoffice.org/63768
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Regression from commit af313fc149f80adb0f1680ca20e19745ccb7fede
(tdf#105143 DOCX import: enable DoNotCaptureDrawObjsOnPage layout compat
option, 2017-01-06), the group shape which covers most of the single
page in the document had a negative left position, while it should have
a small positive one (checking the drawingML markup).
This was a pre-existing problem, but now it's visible since we no longer
force objects to be on the page in Word compat mode.
Seeing the ODT import never positions group shapes (it's just a
container for child shapes), probably the DOCX import shouldn't do that
either. Start moving into this direction, first only for
absolute-positioned toplevel group shapes only, which already fixes the
bug.
(cherry picked from commit d220e476df38e86e094066a690985c624316d37f)
Change-Id: I152ba06a81a2bd09195a4c724da4b8878b0457bb
Reviewed-on: https://gerrit.libreoffice.org/63629
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Change-Id: I954c12d9e1c493be6ac8c7b15076077b5bff5b74
Reviewed-on: https://gerrit.libreoffice.org/62811
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
GrabBag the settings which LO doesn't implement
(which is all of them) so that the document
round-trips without losing the config settings.
Change-Id: I00de6c483af68073634430dd74fd445e981573ab
Reviewed-on: https://gerrit.libreoffice.org/62241
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Change-Id: I8889ce8a7d2309b54454cfe4c6421282e1c6e755
Reviewed-on: https://gerrit.libreoffice.org/63434
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Date SDT from DOCX is imported as date control since commit
3ec2d26dc2017ac4a27483febfc63328632f352d (bnc#779630 initial DOCX import
of w:sdt's w:date, 2013-04-30).
One detail I missed there is our date control is strict: it doesn't
allow free-form text. However, DOCX date SDT has an optional ISO date,
but the actual value can be free-form text. This means that importing
free-form text without an ISO date is lost on import.
Fix the data loss by restricting the creation of the date control: only
do this if we recognize the date format or in case we have an ISO date.
Otherwise just show the free-form text to avoid data loss.
Change-Id: I8125bdc749954a6a1c496de74b6682744adb7680
Reviewed-on: https://gerrit.libreoffice.org/63311
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Change-Id: Ifb3c9016019977aa2c51dbd0a0394a0f81150df5
Reviewed-on: https://gerrit.libreoffice.org/63270
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: I1ddf3fe0e5fad265ae14712a23469b684253079d
Reviewed-on: https://gerrit.libreoffice.org/63241
Tested-by: Jenkins
Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>
|
|
Change-Id: I0b3c407331bfa1fa0c5003250d327d4f26de3643
Reviewed-on: https://gerrit.libreoffice.org/63235
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: I7340a561e0df0c781fd834388deb4b9f83800f9b
Reviewed-on: https://gerrit.libreoffice.org/63221
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: Id16846c19c57ec437a64146d0fa771c1bfc05135
Reviewed-on: https://gerrit.libreoffice.org/63028
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Iea1227a9f13a0a618d9bb6b0bbedaa5ce8d1a4f5
Reviewed-on: https://gerrit.libreoffice.org/62732
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Do not move graphic nodes in SwXText::convertToTextFrame()
since it is not reasonable (they are anchored to something that
already moved) and hence it causes errors.
A static function checks wether the current node is a graphic.
A small change in writerfilter was done since a unit test failed:
The condition "m_nAnchorType != 0" was removed since it is not
reasonable at this place.
Change-Id: I8f27985f6f6c8329483370a7b8baaf9949513f1b
Reviewed-on: https://gerrit.libreoffice.org/60860
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
...just like sdt controls already turned off design mode.
I'm not sure why design mode wasn't entirely turned off
right from the start, but I don't want to do that now because
forms are becoming more prominent in Writer. If design
mode is ALWAYS turned off for .docx files, then the workflow
will be different for *new* forms, since writer defaults to on.
Change-Id: I91c771f1ebd5e640da8067d38cb06076d8c50d5a
Reviewed-on: https://gerrit.libreoffice.org/62491
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
This resulted from commit 009e3f4841378f13a2764a2c84e147e401b70525
which failed to replace some of numeric sprm ids with mnemonic names.
Change-Id: Ie00a89bee02da383df0cbe27b093f3a1d25e8b03
Reviewed-on: https://gerrit.libreoffice.org/62813
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: Id1a0749b78a7021be3564487fb974d7084705129
Reviewed-on: https://gerrit.libreoffice.org/62718
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
The document properly opens with all sections protected because
document protection (forms view) is enabled. However, when that
setting was toggled off, all sections wrongly became unprotected,
and remained unprotected when round-tripped
(including in Word - so loss of configuration).
Word does protection differently. It opens up in a forms only mode,
but upon enabling editing mode, the individual sections can
still be protected. Only when global enforcement is disabled
do all sections become editable.
So, if global enforcement is enabled, map the section protection to
LO native protection. On startup, the sections will all be protected
because of the global compatibility flag. If the flag is turned
off, then you enter a similar mode to Word's "Edit document" where
the sections are still protected. In LO, *each* section's protection
must be turned off individually to fully disable enforcement.
This patch keeps the same three-step process to fully edit the
entire document, but the meanings take on a different form.
"Compatability: Protect Form" changes from "enforcement" to
"edit document" in concept. BTW, that matches how export works,
where PROTECT_FORM is auto-enabled if any sections are protected.
Section protection in LO can be disabled through
Format - Sections - Write Protection.
Patch initially developed to support tdf#120499.
It depends on an earlier commit in order to round-trip.
Change-Id: I8a2399f79640115d689ae9093792eecef7dbaeec
Reviewed-on: https://gerrit.libreoffice.org/61918
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
As the floating objects anchored to footers aren't wrapped around properly
(they behave as if they are wrapped through unconditionally), which makes
imported tables to overlap the page body text making the document unusable,
let's just disable the conversion for the time being (until the actual bug
fixed properly).
Change-Id: I06c984ff7157b71fff2aa8122cc475a1199994c6
Reviewed-on: https://gerrit.libreoffice.org/62523
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: I444c43b9d549977039f25bec2b5bf666c3e15e0e
Reviewed-on: https://gerrit.libreoffice.org/62041
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
which seem to have snuck back in since the great rounds of removals.
Change-Id: I85f7f5f4801c0b48dae8b50f51f83595b286d6a1
Reviewed-on: https://gerrit.libreoffice.org/62229
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I1ae16467a8e58e8a50f59b7a140e9f8b68bde07e
Reviewed-on: https://gerrit.libreoffice.org/62254
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
looks for variables that can be declared const and static i.e. they can
be stored in the read-only linker segment and shared between different
processes
Change-Id: I8ddc6e5fa0f6b10d80c75d5952df8ddd311cf892
Reviewed-on: https://gerrit.libreoffice.org/61591
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
it's not enough to compare != end(), you also need to compare the key
against the iterator result
Change-Id: Ide5f151ba2297a35e5546f47fbc3c53cbe5ab533
Reviewed-on: https://gerrit.libreoffice.org/62014
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I1df70b7dff5ebb6048f7fc618789faa15ca5d422
Reviewed-on: https://gerrit.libreoffice.org/61967
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
...to avoid -Werror,-Wshadow from Clang trunk with
<https://reviews.llvm.org/D52400> "Improve -Wshadow warnings with enumerators",
warning about shadowing of BORDER_TOP/LEFT/BOTTOM/RIGHT in enum BorderPosition
in writerfilter/source/dmapper/PropertyMap.hxx.
Also, BorderPosition can apparently be made a private member of BorderHandler,
which gives extra confidence that any other uses of BORDER_TOP/LEFT/BOTTOM/RIGHT
across writerfilter/source/dmapper/ indeed meant to reference the other enum
BorderPosition.
And change BorderHandler::lcl_sprm to return early in default case, to avoid
having to come up with some new BorderPosition sentinel value to represent
"invalid pos".
Change-Id: I31e535326627cec25a17b99485bf7cf56f4c66d3
Reviewed-on: https://gerrit.libreoffice.org/61896
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Change-Id: I409eb644bdf519e097c87911b90f83c5ecd913ee
Reviewed-on: https://gerrit.libreoffice.org/61853
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
redundant get() call on smart pointer
Change-Id: Icb5a03bbc15e79a30d3d135a507d22914d15c2bd
Reviewed-on: https://gerrit.libreoffice.org/61837
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
tighten up the handling of binary operators
Change-Id: I262ec57bf7142fa094d240738150a94d83fd15ee
Reviewed-on: https://gerrit.libreoffice.org/61777
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
5d0c83fd4cf91083805f60f49e4fafd3d6ac73d4 follow-up
Change-Id: I3611172944072ac009c7c5191cbe05ec9600fd1d
Reviewed-on: https://gerrit.libreoffice.org/61847
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: I7abd715b6bb71d6e2e01939c4cf849d94eb6a103
Reviewed-on: https://gerrit.libreoffice.org/61843
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
error: accessing an element of the container does not require a call to
'data()'; did you mean to use 'operator[]'?
Change-Id: I022745aa84b80124f342ce6c596d51b4d904f012
Reviewed-on: https://gerrit.libreoffice.org/61820
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Iadab935efff35e1e8e5d70878414da1ed2f0fbf9
Reviewed-on: https://gerrit.libreoffice.org/61795
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
And fix
ScXMLCachedRowAttrAccess::Cache
which was never setting its mnTab field, and hence would never
be hit.
And fix oox::xls::CellBlockBuffer, which was never setting mnCurrRow.
Change-Id: I2c46aa050b9ebe3c2dc2e52579555f97945dd61c
Reviewed-on: https://gerrit.libreoffice.org/61772
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
message is "calling a base constructor other than the copy constructor"
these constructors are not in use, so just remove them
Change-Id: Ic8cede1eb153a6ae97fd1ca9c96ca8c1f43e3399
Reviewed-on: https://gerrit.libreoffice.org/61784
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Although this follows a very different code path, copy the ww8
import idea of "consuming" the bookmark of a BOOK_FIELD.
This is particularly important for textcontrols, especially
since LO keeps duplicating bookmarks as it adds another
bookmark for the field name at each save.
Existing unit tests that this matches are fdo53985.docx and
tdf111964.docx. I expected more, but apparently most fields
don't contain or represent any bookmarks.
This patch is for import only. A followup patch stops
the creation of duplicate bookmarks during export.
Change-Id: I1e11980e52dc523393fd6d621191228d676e9a17
Reviewed-on: https://gerrit.libreoffice.org/61615
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Repeating table headers consisted of more than 10 table rows
switch off table header repetition during DOCX table import
to fix non-visible table content and broken tables.
Repeating header lines are not visible in MSO, if there is no space for them.
OOXML (and ODF) standards don't specify this exception, and unfortunately,
it's easy to create tables with invisible repeating headers in MSO, resulting
OOXML files with non-standardized layout. To show the same or a similar layout
in LibreOffice (instead of a broken table with invisible content), we use a
reasonable 10-row limit to apply header repetition, as a workaround.
Later it's still possible to switch on header repetition or create a
better compatible repeating table header in Writer for (pretty unlikely) tables
with really repeating headers consisted of more than 10 table rows.
Note: This workaround could help to create standard and more portable OOXML
files in a mixed environment.
Change-Id: I17fbc0173ec1c4f188a46227b99dde5726530da3
Reviewed-on: https://gerrit.libreoffice.org/60689
Tested-by: Jenkins
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
Found with bin/find-unneeded-includes
Only removal proposals are dealt with here.
Change-Id: I04c5ba277d5b3398c07de6ae66713d977636088d
Reviewed-on: https://gerrit.libreoffice.org/61347
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
despite the documentation that states it must be a positive value,
the actual implmentation honors a negative number (as opposed to
treating it as a zero, or an absolute).
Change-Id: Ie38369ecdf5353b73572dd3533c2efe41adda290
Reviewed-on: https://gerrit.libreoffice.org/59401
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Change-Id: I9a763eacf8887230ba5d88ac69cb13d0cfb1a7ee
Reviewed-on: https://gerrit.libreoffice.org/61274
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
|
|
Change-Id: I7db0c27ff2213210ed4b46ebbadc1a2f74a18257
Reviewed-on: https://gerrit.libreoffice.org/61249
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
...and the column width must not be negative.
The previous column logic ensured that the total width was
larger than the reference by adding a fake buffer
and then subtracted the difference from the final column.
In the case of a zero-width final column, it could become negative.
The current logic ensures that the total width is less than
the reference width, and then adds the difference
(which should be a smaller difference now) to the final column.
Regression potential - early columns that need every single twip
of bonus space might not fit anymore. On the other hand,
ending columns might be fixed...
Change-Id: Ie75d455e8ed62dbec5a1b9c901417df8d842ace8
Reviewed-on: https://gerrit.libreoffice.org/59400
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
This mostly affects explicitly defaulted functions that had recently been user-
declared to silence new GCC trunk -Wdeprecated-copy. It is not entirely clear
to me what the best approach is overall, see my mail
<http://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20181001/245321.html>
"Re: r343285 - [cxx2a] P0641R2: (Some) type mismatches on defaulted functions
only", but lets just explicitly delete those functions for now.
Change-Id: If8c72f612f67a8feb8b03c2fb988c807e704ef03
Reviewed-on: https://gerrit.libreoffice.org/61259
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
Use range-based loop or replace with STL functions.
Change-Id: Ie1e169f9d2d44e188180164bd1f93af43740425d
Reviewed-on: https://gerrit.libreoffice.org/61141
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I5f9ef043d76c55f2c761fd08a2bc1dae66b675c8
Reviewed-on: https://gerrit.libreoffice.org/61073
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
If the tblBorder does not exist, or if first row's
left-border has a different sized border than the
overall table settings, use the cell's lineWidth
value to determine the the spacing needed.
The potential regression here is the spacing given to
the cells. However, the current calculation is totally
wrong (it needs to be based on per-cell, not on some
table-level default) and so it was "wrong" a different
way before. The unit test shows this visually.
Plus, MS Office version are inconsistent in handling this
unit test. The latest versions base the position on the
*last* row and not the first row.
The key reason for making this change is to support a
table that doesn't have a table default, needed to
fix tdf#92026.
Change-Id: I319513a2f5550c1527dc564fb8b3806d1251a0c2
Reviewed-on: https://gerrit.libreoffice.org/60988
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Change-Id: Icfe120db02a70d43e4120d1f4cbb2704640d3990
Reviewed-on: https://gerrit.libreoffice.org/59470
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
implemeent a reduction approach, which is good at finding virtual
methods that only themselves or their virtual partners.
The accessibility GetVisArea stuff is dead since
commit 891e41fac81fbd8d5cdb277b26639abfd25a7143
Date: Wed Apr 4 11:23:22 2018 +0200
dead code in AccessibleTextHelper_Impl::UpdateVisibleChildren
Change-Id: I78d9d8bca585ecec8394f2c3fe2baa93db0e58f5
Reviewed-on: https://gerrit.libreoffice.org/60912
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
"new" borderlines (from 2012, 2014) defined in
offapi/com/sun/star/table/BorderLineStyle.idl
where no supported in export yet!!! Sad.
const short FINE_DASHED = 14; (rtf)
const short DOUBLE_THIN = 15; (rtf/doc/docx)
const short DASH_DOT = 16; (rtf/docx)
const short DASH_DOT_DOT = 17; (rtf/docx)
Change-Id: I07fee08e17171db4baec489864fc47da107cee75
Reviewed-on: https://gerrit.libreoffice.org/60992
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Deciding when to and when not to deduplicate repeated direct formatting
of paragraph / character properties is stricky, this bug is about a case
when deduplication should happen and did not, since commit
1970a686273c5d4fc1eeb4430283e37085d9f647 (tdf#113408 RTF import style
dedup: separate paragraph and character handling, 2017-10-31).
Especially that deduplication works in both directions: it should remove
properties which are duplicated and also should insert explicit default
values for not repeated properties.
Fix the problem by making the getDefaultSPRM() aware of the context
(which style type it deals with), and then by making sure that only
default properties relevant for the given style type are inserted.
Change-Id: I35b6599cc47fa51b8754fd921c61a3b31a283547
Reviewed-on: https://gerrit.libreoffice.org/60946
Reviewed-by: Miklos Vajna <vmiklos@collabora.co.uk>
Tested-by: Jenkins
|