Age | Commit message (Collapse) | Author |
|
Since we just handle one specific case, just initialize the
variable to something different, instead of checking the
GetFirstRedoInfo result.
From valgrind:
==10566== Conditional jump or move depends on uninitialised value(s)
==10566== at 0x270E4451: SwDoc::ChgPageDesc(unsigned long, SwPageDesc const&) (docdesc.cxx:431)
==10566== by 0x272E341F: SwFEShell::ChgPageDesc(unsigned long, SwPageDesc const&) (fedesc.cxx:122)
...
Change-Id: Ib95f81297959a535a08dc399f72b68c15e621a27
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/118017
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
Change-Id: Icbf1a9a6fed698345b8faabf3443216240c65069
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/118011
Tested-by: Jenkins
Reviewed-by: Andrea Gelmini <andrea.gelmini@gelma.net>
|
|
Change-Id: I333100fda7e181f68f36b03279b3fbb8cb768310
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117615
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
LO uses the bounding box of the shape in case of type 'Parallel'. Word
uses in the corresponding wrap 'square' a box based on the full size
of the shape. That will be very different in some cases, e.g. for an
arc. And Word exchanges width and height in case of rotation angle in
[45°;135°[ and [225°; 315°[. To get the same appearance as in Word,
the wrap margins are suitable expanded on import.
Word puts the additional space needed for fat strokes into effectExtent
in case of wrap 'inline', so there is no need to add the half strokes
width in addition. Word determines the area for the shape depending
on rotation angle. Both are now considered. Total same appearance is
not possible because it would need negative vertical wrap margins,
which are currently faulty in LO, see tdf#141880.
Patch solves in addition tdf#142486, tdf#142305
The export to Word would require negative values in effectExtent in
some cases. They are allowed in OOXML but not supported in Word. My
idea is to switch to wrap mode 'Tight' if needed. But export of wrap
has so many bad parts, that it needs separate work and is not
included here.
Handling of border width for export of own frames is missing.
Unittest changes
----------------
testDmlTextshapeB and TestDmlTextshape in ooxmlexport6.cxx are set to
current values. Import and Export still have large errors with these
shapes and correct value from file is unknown. So an exact value is
pointless. Only the original problem needs to be still fixed, which
is the case.
testWpsOnly in ooxmlexport10.cxx. I have removed the test for
LeftMargin equals 0. The test makes no sense, because the original
file has distL=114300.
testTdf124600 in ooxmlimport2.cxx.
I have added a tolerance. It would fail with Expected: 2029,
Actual:2028, likely a rounding problem somewhere.
testTdf124600 in ooxmlimport2.cxx
Word refers to outer edge of the border for align='left', LO aligns
at snap rectangle. The different intepretations become visible if a
thick line is used. LO needs a margin to get the same rendering as in
Word. So an expected value of 0 is wrong and I have disabled the test
for now. ToDo: tdf#142798. Get the correct margin and activate the
test then.
testTextframeGradient in ooxmlexport2.cxx. I didn't find any reference
for a default value. The test is not reliable, I get both 316 and 318
as actual value. Handling of shadow in VML shapes is buggy, the values
for margin and shadow are wrong anyway. Reports are e.g. tdf#142486,
tdf#142558. For now I have added a tolerance of 2.
testDMLGroupShapeChildPosition in ooxmlexport6.cxx. The accuracy has
become better. After reload we get the same values as before.
testEffectExtent in ooxmlexport.cxx. tdf#142805. I have disabled the
test, because the image is not loaded at all, and therefore it makes
no sense to test a margin of it. And you can only test the sum of
distL and effectExtent l, because LO has only one property
'LeftMargin' for it.
testEffectExtentInline in ooxmlexport.cxx. To get the same vertical
alignment as in Word, it would be necessary to have negative margins,
but those are not implemented yet. tdf#141880. Currently the patch
contains a heuristic to detect unchanged objects and write the
grab-Bag values then.
testEffectExtentLineWidth in ooxmlexport16.cxx. I have changed the
expected value from 508 to 506. That is still away from the to be
fixed faulty 561. It is a VML shape and import of connectors in
VML shapes is faulty.
testRelorientation in ooxmlexport4.cxx. Changed to 8662, the original
problem is still fixed. I don't know the reason for the difference to
the values from file.
Change-Id: I28f156637f6ae64975cf2917f0e5cc89e689aff5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/115668
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
|
|
... and SwWriteTableRow
See tdf#94879 for motivation.
Change-Id: I74edd0d6864fd7c2731d904c8bf7cfbd309a226c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117965
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
MS Word formats only have 9 list levels (0-8),
and listLevel 9 is used to indicate body-level
as a way to cancel the inheritance of a listLevel.
LibreOffice however has 10 levels (0-9), so it was treating
this as 10th level numbering. Nope - it needs to be
no numbering at all.
Change-Id: I3fd58ba518ba8bc7d15a08cf896fbeed8e6a38c2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117921
Tested-by: Justin Luth <justin_luth@sil.org>
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
after 8298aa62726312eee6f8fbb64a9fc9b12680447f
< uitest: guard execute_blocking_action >
Change-Id: I954c11c42620660afd48a585f5c59ff2196f4b6a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117738
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
Change-Id: Ifab4209e4481058046c95efc9efccf337b4badd5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117939
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
This reverts commit 066c4054f4a1078602aaab5516590628eaf6a47e.
As mst_ says:
there isn't just File->Open but also Insert->Text from File...
in that case the optimisation looks wrong to me
Change-Id: I9d0b4bc748842b8fac7ea4cd58192562f8fd82f2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117864
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Chapter numbering is a horrendous invention of LibreOffice.
It is not at all a part of MS Formats.
If there is anything that potentially can cause problems
by choosing one style to be the outline style,
invalidate it.
We won't totally ignore CN because it has some unique
properties in LO, and so it is worthwhile trying
to round-trip it (alwyas LFO 0?) or populate it
with the expected Heading styles if compatible.
(I tried looping through and re-setting the
matching styles with the "Outline" numrule first,
but lots of unit tests prove that doesn't work well.)
I couldn't believe the incredibly good effect this had on
my torture test example document. Hurray.
Then again, completely preventing CN at all also
passed all unit tests, so I added a preventative test earlier.
Change-Id: Iafa0d3606e558a3f17e49c3f1f6cebb7b25de962
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117920
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Change-Id: I7b6879d0e038f8c07ae5853d66c43096c25365e8
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117841
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
So chapter numbering is a complete scourge for importing numbering,
but after many unit tests proving my attempted solutions were
incompletely, I finally found one that worked great.
Unfortunately, I made a major coding blunder, and effectively
killed chapter numbering unconditionally. But it fixed every
unit test problem!!
So that pretty much proves that chapter numbering is trash,
but for LO it still has some important functions,
so ensure that at least on a clean, nice, LO-authored document
chapter numbering still round-trips nicely.
Change-Id: I37491ecb638aa79fd7c7621e9cf0ea6282397b83
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117919
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
to make sure the documents are closed in inverse order when
different documents are used in the same test
Change-Id: I97ff33f9f91423d06f4e74bba6fc6a5312b24a17
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117914
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
This looks like it was a logic error.
Probably they were only looking at the situation where
the paragraph style set the numbering style to "Outline".
Well, in this case, Heading 1 is actually not part
of Chapter Numbering, but it was the body sprm
that set the style.
In any case, whenever the LFO chosen as the outline
style is referenced, it should apply LO's special,
unique and extremely troublesome "Outline" numbering style.
Change-Id: I354d9b9f4a48abed3453ee046ddd63cda174f25d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117746
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Mostly done with a script
for motivation, see 89aaa17a0a4413f07da2bc5084b0164f15dc01ac
< UITest: introduce guarded context managers >
Change-Id: I8b0e4aa746f7973b1bd1343219820d8b2315cf19
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117703
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
In no_view.odt, the table 1340 does RemoveFollowFlowLine(), then splits
again, and by the time Split() returns, its flag
m_bRemoveFollowFlowLinePending is set and so in the next
SwTabFrame::MakeAll() it will do the same thing again, until eventually
it hits
layact.cxx:544: LoopControl_1 in SwLayAction::InternalAction
The flag is set during lcl_RecalcSplitLine(), in SwContentFrame::Cut(),
which operates under the assumption that there are no other cells in the
follow flow line that may contain content frames.
This only happens if "Verdana" and "Arial" fonts are available and not
substituted with "DejaVu Sans" etc.
(regression from c4af8b0f106f31b6e79a145bc4ab83bb6a55c808)
Unfortunately this uncovers another pagination bug in testTdf88496
which was hidden by this flag being set from text frames that shouldn't
exist anyway.
Change-Id: I3cc82c13bfc06960b827aefc64f5b5b12520955a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117851
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
|
|
Change-Id: I5f0342f07f9227e45ef12c3c76015aeee4c2a0c7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117833
Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
Tested-by: Jenkins
|
|
This function look a LOT different
way back at original import,
and some of the refactoring since then
left things a bit messy.
The impetus for changing this was to see how m_nListLevel
was being used, and having it assigned to a temporary
const variable for no reason was annoying.
Change-Id: Ia25397dbe4bce2fd7bdfbb1b3053bd271c08212a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/114404
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
The big problem was that hindiVowels was imported as
CHARS_NEPALI, but CHARS_NEPALI was not exported in DOCX.
Another easy win was to DOC export both LO forms of russianLower
and both forms of russianUpper - just like DOCX.
The page number field only allows a subset of formats.
LO doesn't have support for Hex or DollarText formats.
The Ordinal/Cardinal formats were an easy win
for export, even though they don't import.
[Import seems to be in ww8par5.cxx's GetNumTypeFromName(),
except that is being passed an empty string.]
Change-Id: I176a70e0e05967414337c17d8b4acc87e20ead73
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117793
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
If listLvl is undefined, it is treated as level 0,
so when testing that outlineLvl is equal to listLvl,
then consider MAXLEVEL to be equivalent to zero.
The opposite is not true. An undefined outlineLvl
is not considered to be level 0.
This doesn't fix bug 104239, but it is one step
in the right direction.
Two unit tests followed this path,
(transparent-text.doc and tdf80635_pageLeft.doc)
but neither one sets a numbering style,
so they don't make good unit tests.
Change-Id: I62cd7f5500fcd6dc06327900c33f12c129610a04
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117745
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Change-Id: I8c655c6ed1e1d05407dac27ef19f26c6ca8d8d54
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117692
Tested-by: Jenkins
Reviewed-by: Jim Raykowski <raykowj@gmail.com>
|
|
after commit 1feb59c3dd9e5f714ce894f8de31bbb6869c9f3b
Change-Id: I5dde1a776c11b9755e33cd7ccae54f29045b5ea1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117697
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
trust the import filter to have already de-duplicated styles
Shaves 5% off my load time
Change-Id: I36b5196762483f629a38e00156d2fe9a1cb28c97
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117759
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Good - an example document that proves that inheritance
needs to happen with more than the first 9 items which
contain the Heading styles that are most often
associated with Chapter Numbering.
Change-Id: I69ab1dd3fdab4d54695d9b321b0d2970509090a6
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/114403
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
This reverts commit 374baf308979306aa35575118c40ccd7caae1e29.
Many uitests are failing randomly in jenkins for no apparent
reason
Change-Id: I5960330fab4967518bfeea32b3b8c5f8bfbea57e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117752
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
Change-Id: I4e13ca9eac9c86f08c869c8fa9d3fb346c2aa2d2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117710
Tested-by: Jenkins
Reviewed-by: Heiko Tietze <heiko.tietze@documentfoundation.org>
|
|
RegisterNumFormat always ran earlier since it forced
the listLevel to be in the valid range.
So that can automatically be simplified.
Secondly, just because the listLevel was not specified
doesn't mean that it will be a level 1. It could
inherit a listlevel from the style. So just leave
it as undefined here, and let it be handled later on.
ww8export's testListNolevel already ensures that
SetStylesList must handle unspecified listLevel.
chapterNumberingTortureTest21.doc matches the assert,
so I have found my document, but this already works.
Change-Id: Idf8850980b92054104c16200b2dea6bd7dde1f46
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117481
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
I cross-checked with what was imported via DOCX
and took inspiration from tdf#141341 to consolidate
these two nearly-identical functions that otherwise
would need to be kept in sync (and never are)..
Change-Id: I7e021d7c9d68597da5b0ce1311ae243fc003e4f7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117736
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Change-Id: Ie3c938992a5c1e7aab3d7ae32f4169d04225195a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117707
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
can just call IsAlive
Change-Id: I7cd3b33520261b68d549783cabd5cf3f6f4847a3
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117696
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Most of the custom shapes were exported with DrawingML custom
geometry instead of DOCX preset shapes, losing their preset type,
adjust handles and text position. Add a preset exporter class to
handle all possible preset shapes later, and fix the export of
the following shapes now: "bevel", "blockArc", "bracePair",
"bracketPair", "hexagon", "octagon", "parallelogram", "plus",
"roundRect" and "triangle".
tdf#92525: keep text position in triangles with different
adjustments, too.
tdf#142398: part 1: export simple shapes as preset shapes.
Change-Id: I6aee74f7670bea8c1fe5909cbf307778ea728669
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/115606
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
Change-Id: Ia1d0c358868e899e121b8331fad17ce36d454f10
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117668
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
|
|
by re-arranging the boost::multi_index key, we can re-use
that index to find all formats matching a name
Change-Id: I3b3b6ceb7137926be3981578a64052f47f279af0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117676
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Mostly done by a script
for motivation, see 89aaa17a0a4413f07da2bc5084b0164f15dc01ac
< UITest: introduce guarded context managers >
Change-Id: I75ef7712af3676363a9a464acf83f6f68ffc4f85
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117617
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
in preparation for adding optimised implementations of it for subclasses
Change-Id: I6d8ff54864d2d3c605a1cd0b4da2c6136e2a21cb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117672
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
because the index it relies on, is a unique index.
So make this explicit in the method call by returning a normal
iterator.
Change-Id: I428d1b14d1274079955eb4f893ca91d1745264e7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117673
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
When LFO == 0, that is a special case that means that
numbering should be disabled. (This can similarly be
accomplished by setting Lvl to 9 - Body text,
except doing so here could mess up Lvl inheritance).
If the LFO is a negative number, that indicates something
different with formatting that we don't currently handle,
but it still means cancel numbering. So at this point
just treat them all the same as if they were LFO 0.
Change-Id: I1bab06c6b766175a3982bf3a7b19f7b321df807e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117480
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Mostly done by a script
for motivation, see 89aaa17a0a4413f07da2bc5084b0164f15dc01ac
< UITest: introduce guarded context managers >
Change-Id: Ia924293dc2a12230fd13f69fd734875ced86be8a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117593
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
This patch includes:
* A README.help.md as a general documentation for the different
help types, the LO code and help-related build options.
* Adds --disable-xmlhelp for removing the xmlhelp support from the
build. Disable for iOS, Android and Emscripten. This was partly
included in HAVE_FEATURE_DESKTOP before.
* Rename WITH_HELP define to WITH_HELPPACKS, which reflects the
actual usage AFAIK.
* Depend --with-omindex on --with-help=online and don't override
the --with-help setting. Error out on conflicting options.
* Depend --with-helppack-integration on build help variants, which
actually result in help packs.
Kind of reverts commit 2c38ea6d16b910294220cefaf8ae6a0683e6405a
("Building without --with-help is not supposed to disable help
functionality").
Change-Id: Ie4cb73905b3ed94e991d9f1bd75cfbd6de9da385
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/116222
Tested-by: Jenkins
Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
|
|
since...
commit 2e6e87decdb87221a7b832958e03195b1d8fc133
Date: Mon Jun 21 13:18:00 2021 +0200
replace dynamic_cast<SwLayoutFrame> with cheaper check
Change-Id: I4623e3eca5a82fa37f52697c189310a835fc76e0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117614
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I12613ed381e961fd36348eabd543b11a1960337d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117613
Tested-by: Caolán McNamara <caolanm@redhat.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
an intermediate horizontal GtkBox works to stop the propogation
Change-Id: I9fd51c3968c463c7b872b7a12bc3960602d0b025
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117596
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
to avoid expensive dynamic_casts in writer
Change-Id: Ic73ae7d4af985043f8533913b33939e3445ec7de
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117586
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
It just looks redundant. And all three formats
handle it very different.
RTF - never writes RES_PARATR_OUTLINELEVEL
or RES_PARATR_NUMRULE in styles,
only via this OutlineNumbering.
DOCX does nothing. That's because there always
seems to be a RES_PARATR_OUTLINELEVEL and
RES_PARATR_NUMRULE.
DOC does BOTH. It writes RES_PARATR_NUMRULE
(and now also writes RES_PARATR_OUTLINELEVEL)
and also writes them via OutlineNumbering.
So OutlineNumbering seems to always have
been partially redundant and now is fully
redundant. Thus it can be treated like DOCX.
Change-Id: Ib9fd32c440e0151b462d6499a60952f120887082
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113452
Tested-by: Jenkins
Reviewed-by: Justin Luth <justin_luth@sil.org>
|
|
Change-Id: I672be703fb097c71303434127d832681843607e2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117580
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
This nested table is an example that was broken by commit
91b2239783dc716bd71ce7962bfd7e341dfe4175 - if loaded with a wide window,
it goes into layout loop; if loaded with a tall window size where page
2 is visible, strangely it doesn't loop.
Change-Id: I5e73cfcd928ff1a321667c1a75b0ba7f348d4b77
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117587
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
|
|
in prep to add parent support
Change-Id: I2aa4b9343f895ae866f600dd3260b7fdc4e1efec
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117579
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: Ia84984e47c60eb8407bbddaf9e3365d9f7f52311
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117581
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I2102d27fe2b38e3787e39a2b7234991a57c30d99
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117578
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
On loading, this hits loop control
warn:legacy.osl:580715:580715:sw/source/core/layout/layact.cxx:543: LoopControl_1 in SwLayAction::InternalAction
This is because there's a fly frame 404 anchored at the last text
frame 353 on page 2 inside the nested table 347.
ShouldBwdMoved() sees that there is space on the bottom of page 2 and
hence the follow flow row joined, but then it immediately splits again
in the same way as before due to the fly with WrapTextMode_NONE.
But then the outer table's cell 273 (upper of 347) is invalidated
again, hence the loop.
Try to check for overlapping flys in SwTabFrame::ShouldBwdMoved()
by reusing CalcFlyOffsets(), which is ... not quite ideal, but perhaps
better than copy-pasting half of it to a new function.
This should have less side effects than the previous fix, but a problem
remains that clicking on the shape on bottom of page 2 causes the layout
to go wonky, but that was also the case with previous fix.
Note there's a check of SwLayouter::DoesRowContainMovedFwdFrame() there
already, but that doesn't help because it will only detect when the fly
itself was moved forward, but in this case the fly remains on the page.
Also likely it wouldn't be a good idea to move a text frame forward if
the only thing of it that fits on a page is an anchored fly (i.e. its
follow has mnOffset=0) because that can be intentional.
Change-Id: I0376f7dcb784c006990336233c97f5093aaccb77
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/117473
Tested-by: Jenkins
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|