Age | Commit message (Collapse) | Author |
|
This reverts commit 4969f4c0d3e2581aaa8a5b5a5769840fa6b6f8ea.
Change-Id: I527e9366b05e8a20633720e334395b285991c524
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90473
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
GetWordCompatibilityMode() is an expensive call
that is called at least for every table row.
Caching will soon be even more beneficial
when LO also starts writing out this value
and thus quickly caches a non-negative result.
Anything created by MS Word since at least 2010 specifies
this optional setting, so most .docx files specify
this value - and thus will benefit from the cache.
One difference that I ignored is that Word replaces a
"number greater than what I know" with its own
compatibilityMode value. (tested in Word 2010 and 2016.)
That is probably only a concern at export time anyway...
Change-Id: I0ee71d34150c5f3d9858adb678814a10c7fe8959
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90434
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
I'm sure that these functions need to change because they
just seem very wrong. I think these are fall-back, safety
sections because usually the nEsc is already converted
away from _AUTO_ prior to reaching this point.
But this is all just based on code-reading.
It would be nice to have proof of this - especially
since the values for subscript are unexpected. I would
have expected it to use a value comparable to DFLT_ESC_SUB,
and not some random/magical -20.
(Note: the _AUTO_SUPER historically refered to an absolute 33
and not DFLT_ESC_SUPER until I recently changed that...)
Change-Id: I3bef0841f9a20f09994b2372e9a7db5941b29c9f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/88997
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
commit 24019d56fbc4ac94bd1b72b37a9bcabc7ef8164b seems to have introduced
this behavior.
This patch is a work around that skips code introduced by the above
commit when selection type includes number list type.
Change-Id: I7a74ad7262e3e008f6aa11fe04fb81b9604ef3e1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87795
Tested-by: Jenkins
Reviewed-by: Jim Raykowski <raykowj@gmail.com>
|
|
Change-Id: I2f1d5fdf976f9e787d64098a091022182156e401
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90408
Tested-by: Jenkins
Reviewed-by: andreas_kainz <kainz.a@gmail.com>
|
|
Revert "tdf#121607 defer other requests until password dialog is dismissed"
This reverts commit 205aeffda1a29865576fb865f23dfb6761ebfab1
as far as I can see the original problem doesn't reappear anyway
Change-Id: If5fd3383f3473662bc87a367fad2c1febe1c6a23
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90375
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: Ifa76e004128223460945d58d1c59c4e23db0f108
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90370
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Even if the row is set to not allow splitting across pages,
ignore that setting if the row is too big to fit on a single page.
Don't worry too much about compatibility, because there is no
sensible reason why anyone would have hidden content like this
intentionally.
An oversized row has always moved to start a new page. While that may
not strictly be necessary anymore, to approximate a bit of backward
compatibility, continue to do that. MS Word will do the same...
Word, prior to 2013, always tries to keep the whole row on
one page. In 2013 (compatibleMode == 15), native documents
will be treated like this patch. So, although this patch
throws away senseless compatibility with existing documents,
it is interoperable with anything authored by Word >= 2013.
Word 2013/2016 also opens .odt files this way.
HOWEVER, LO authored .docx files do not set compatibleMode=15,
so Word will treat them the old way - hiding all the content
on a single page.
Change-Id: I306e22230ed6fe21f6b66700ffd7615678859f5d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90005
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
Pending paragraphs were disposed before table
processing, so empty tables have left out of the
rewritten table style preference handling,
resulting for example different cell heights.
Regression from commit 4d5c0eaf3e0d3d3bcd9e691fffee19b75f3d6631
(tdf#118812 DOCX import: fix table style preference – part 2).
Change-Id: I5d5d29f37c8198484e24519ced5a50458c923064
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90347
Tested-by: Jenkins
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
Titel -> Title
Thema -> Subject
In "strings.hrc", translated the defines, but kept the their NC_(...)
translation references to prevent a need to update translations.
Change-Id: Ie8b4a94376495d5e148bc50dcc9902e805deff0d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90156
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@cib.de>
|
|
Change-Id: Ic8d049c7b4a9005473c91cdb6200ae24b14e1a58
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90306
Tested-by: Jenkins
Reviewed-by: Xisco Faulí <xiscofauli@libreoffice.org>
|
|
Fix layout of "auto" width tables with incomplete grids,
where table width is defined by cells of an arbitrary
table row, not necessarily the first row, and last cells
of the rows can be wider, than their saved values.
Change-Id: I68bc8f1a4f57f3c64d0e83c585f2be129d9b5a84
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90261
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
Change-Id: I83a8ab6d8d9f6b0143a51176fed4fa8f2f682cfb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90312
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>
|
|
Change-Id: I179db8fe1762a4931a4752c9c865f8c1e1b4fc4d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90309
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>
|
|
Change-Id: I9c04e15ae05b8b6e98d79cddd87dff55ebbf1f1a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90308
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>
|
|
Change-Id: I58f3d546981ad80f8eeabab6b023c4a985f7aecb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90173
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>
|
|
Change-Id: I51d7bbe527a722ddd8a603afe15e75ad4b8c737a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90167
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>
|
|
Change-Id: I0e41ac95532b8d1782c4a5148a2b70b1d8a13b16
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90289
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
When the http response is 404 or something like
that we shouldn't retry the download embeded image
constantly. This causes libreoffice to short freezes.
Change-Id: I7381d04f12e9fbea961dd0e3333ea0d39aa93d14
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90102
Tested-by: Jenkins
Reviewed-by: Gülşah Köse <gulsah.kose@collabora.com>
|
|
To be specific, this focuses on the case where a single cell has btlr
direction, but the containing row does not, and there is a row span.
The old code that took the logical right of the cell frame served two
purposes, it seems to me:
- in the rare case where the 1st row is wider then a subsequent row, we
make sure that as the cell frame is expanded downwards, we're still
inside the table polygon
- in the tb-rl ("Japanese") case, the logical right maps to physical
bottom, and this way the cell frame is OK to render not only into its
own row frame, but can expand also downwards.
Given that btlr maps left to bottom, this mechanism is broken there. Be
consistent with the working tbrl case, so just expand towards the
logical left in the btlr case.
The rest of the changes just make sure that SwAttrHandler::FontChg()
calls SwFont::SetVertical() with the correct bVertLayoutLRBT parameter,
which instantly fixes the actual position of the text.
Change-Id: I9032e7c6de72cec704843f3aae3c7848e139ebfa
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90241
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
resulting broken table layout with incomplete cell merging,
for example, content in extra table rows and in columns
of different lengths.
Change-Id: Ic5057e43d4c66e34ffc1373030be66815ff52563
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90136
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
Change-Id: I410deb7c23b9a253b4331b12a59abd1d47884d62
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90189
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>
|
|
Change-Id: I1ed505d07cd8b96ac7f683846ef2053077792a12
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90172
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>
|
|
Change-Id: I2616f651f13306dd78732937310f204789b470fd
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90213
Tested-by: Jenkins
Reviewed-by: Xisco Faulí <xiscofauli@libreoffice.org>
|
|
if we have enough space under the horizontal X axis.
Note: allow truncation of vertical X axis labels only if they
are text labels and the position is NEAR_AXIS or OUTSIDE_START.
Regressions from commit 35d062f7879d5414334643cb90bff411726b2168
(tdf#116163: Limit label height in chart if needed)
and commit 26caf1bc59c81704f11225e3e431e412deb8c475
(tdf#114179: Custom size and position of the chart wall)
Change-Id: Idf86bc2b5482bb50a266cda57cc502621c2e08ed
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/89829
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
noticed something that wasn't being picked up, wrote some tests,
and found an unhandled case in Plugin::getParentFunctionDecl
Change-Id: I52b4ea273be6614e197392dfc4d6053bbc1704de
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90141
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I191d8778d362cd28474eea6d18bfe40044887e30
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90086
Tested-by: Jenkins
Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de>
|
|
... SwUndoSplitTable and SwUndoTableAutoFormat
See tdf#94879 for motivation.
Change-Id: I5e40c9ba391456c95b8ef44aa4cf11ebbd681d80
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90200
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
Change-Id: Ibe6b2d2f534fad296c46ef6c3c18c63ccbf265e2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90170
Tested-by: Jenkins
Reviewed-by: Björn Michaelsen <bjoern.michaelsen@libreoffice.org>
|
|
This patch ignores outline tracking when other than Headings/outline
content is navigated to from the Navigator.
Change-Id: I3ecbe508b5d641e8f9bd8fe2e70d8ae23f0bf879
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/89928
Tested-by: Jenkins
Reviewed-by: Jim Raykowski <raykowj@gmail.com>
|
|
We added a single setup method to replaced setActive call in each
tests.
This commit drop all the replaced setActive call in each tests in
this tiledrendering module.
Signed-off-by: Andrew Lee (李健秋) <andrew.lee@collabora.co.uk>
Change-Id: If0d0b8f6cd03bfbb935bd8a0023530799f5cde34
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90158
Tested-by: Jenkins
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
|
|
Originally there are 'setActive' in the 1st call of each tests in the
tiledrendering module.
Here we add a single setActive to replace the 1st call of each tests.
Signed-off-by: Andrew Lee (李健秋) <andrew.lee@collabora.co.uk>
Change-Id: I7e8a6fa39d2d4bf0e4eab2e30cbee7437e59ef9a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90157
Tested-by: Jenkins
Reviewed-by: Michael Meeks <michael.meeks@collabora.com>
|
|
The default padding of five results in a busy loop for layout
values >= 100000. And since we read the record value from a file,
don't assert on larger values, by simply masking it.
Change-Id: I95329d03405ed5fbc749dda3f7d090a4739b0018
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90124
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Change-Id: Ia82d545e4c5d4507899d123eba9d4b2efded992d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90125
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Change-Id: Idae665c926856922a25a1228ed34f3cae96e4bdd
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90085
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I256c7a2257dd74b5650854f0231615353b43dcdb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90119
Tested-by: Jenkins
Reviewed-by: Tamás Zolnai <tamas.zolnai@collabora.com>
|
|
Change-Id: I483cb726d6abc7cad41244e26811fb6e98d9fb3b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90113
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Chicago numbering is not supported for paragraph numbering (same as
DOC), so focus on footnote/endnote export only.
There is markup in RTF to store doc-global footnote/endnote numbering
type and the same for per-section. DOC writes both, then Writer only
reads the global setting and Word only reads the per-section setting.
This means only export is needed here, import already handled the
doc-global markup, and that's enough.
Change-Id: I3590560ca913e04078988fe4784e50fa5cbf17bf
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90112
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
Word refuses to open a document that has a w:bookmarkEnd inside
w:sdtContent but with no text content following it.
It turns out that the bookmark position is wrong anyway, it should end
before the text according to Writer's model.
It shouldn't make a difference whether the end is inside the sdtContent
or preceding the SDT, so write the text content of the SDT from the
EndField_Impl().
Another idea would be to move the writing of bookmarks in EndRun()
before the StartField_Impl() but who knows what that would break.
(regression from d55b26a093bdbced08985dbc7113190b52a8bc66)
Change-Id: I476c0829814b061d80811cc6817923ee06013a26
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90100
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@cib.de>
|
|
Change-Id: I81caa059084bbe21d112c4ca49ef5d3275356cce
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90087
Tested-by: Jenkins
Reviewed-by: Xisco Faulí <xiscofauli@libreoffice.org>
|
|
After cursor jumped to start of first line or end of the last
line (with up/down arrow keys), when we clicked the arrow keys on
the same direction again, cursor movement column will be the
last or first column.
Change-Id: I9137adea14ce3fa7dd516a4e6fac5641df5a556b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90012
Tested-by: Jenkins
Reviewed-by: Gülşah Köse <gulsah.kose@collabora.com>
|
|
Change-Id: I3623d9424c95983e005c146728518bafcd6d2dee
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90053
Tested-by: Jenkins
Reviewed-by: Xisco Faulí <xiscofauli@libreoffice.org>
|
|
Change-Id: Idf6920278c82a1bdddeb4e1b43c444010fba1aa0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/89985
Tested-by: Jenkins
Reviewed-by: Xisco Faulí <xiscofauli@libreoffice.org>
|
|
Note that chicago numbering can't be used for paragraph numbering. It
would be possible technically, but the spec && bffvalidator && Word
agrees on forbidding that.
Change-Id: Ic3de51f9724d399542f4fe6ac48e70e94c6ea4ba
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90080
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
See also commit 19b2ed4dee1ec33ad55473d43bfcd0bfa194cbee
(related tdf#51671, add UI options for new "hide legend entry" feature)
Change-Id: If20ab167d867d2845fa414ba0bbc490bed09f61c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/89123
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
Change-Id: I55e3e76d7d21c7d0796ec9355f01479232018c66
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/82085
Tested-by: Jenkins
Reviewed-by: Marco Cecchetti <marco.cecchetti@collabora.com>
|
|
When a redline was splitted we ended up with 2 redlines with the same
id (note that this id is used only by lok clients).
When a redline is a multiline, the generated rectangles was wrong.
Change-Id: Iaa7599ab16f4c8939606b1da0dcddb4aac6e983f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/82010
Tested-by: Jenkins
Reviewed-by: Marco Cecchetti <marco.cecchetti@collabora.com>
|
|
Export tests used to do the following steps:
1) Import
2) Verify
3) Discard state
4) Import
5) Export
6) Import
7) Verify
Assuming that the cost of 2), 3) and 7) is almost zero, that leaves us
with 4 filter() invocations for every such "export" test. But it seems
to me, we can save one filter() call if we do it this way:
1) Import
2) Verify
3) Export
4) Import
5) Verify
The only trick is that verify should perform its verification by only
reading the document model. So adapt tests which mutated the doc model:
paste tests should be import tests, the rest only did mutation by
accident.
A 'make check' is 88 seconds faster for me with this, using -j8. (~12
minutes of CPU time saved.)
Change-Id: I40db8a9b9adaa99020acd718ba4ec3fd9bee3fb8
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90077
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
so we can remove the reliance on ::Paint to redraw in red missing links
Change-Id: I80db2d9212496f16ddb598971f9eaa0c4c39f692
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90050
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Only this was missing, paragraph numbering import/export and footnote
numbering import was already working.
Change-Id: Ia5966cc7f1308ba81bebc1bf628d8efb17acb713
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/90075
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|