Age | Commit message (Collapse) | Author | Files | Lines |
|
if we have to produce output data for a large range, we end
up with an O(n^2) loop because we loop over X, while looping over X
Change-Id: If78879c2781561dd8b965671370ccde274d98fd2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187752
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Jenkins
|
|
instead of a combination data of a bool param and an optional
data in a struct
Change-Id: I67338b55fc38bbe053e1e13dea92f6c76389663b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187773
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Jenkins
|
|
Change-Id: I1e9b85279abb8eb931ced4dfa941eef06585e496
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187744
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
It was too optimistic to think this worked on msvc
Change-Id: I66c85619d5ad1c93dbc2e03bdc228cf4d674ea88
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187778
Tested-by: Jenkins
Tested-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
Per conversation here: https://gerrit.libreoffice.org/c/core/+/187762
and some offline communcation.
Change-Id: If66b74299dd965d237f368c6c901057241226458
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187779
Tested-by: Jenkins
Reviewed-by: Kohei Yoshida <kohei@libreoffice.org>
|
|
This reverts commit 7c714428ff1a178378dec03cc287ab86fc35ec5a.
Reason for revert: Edit Style dialogs always start with the first tab
Change-Id: I3c4d882279b22597fb078f6fbe273309586e1cb1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187780
Reviewed-by: Heiko Tietze <heiko.tietze@documentfoundation.org>
Tested-by: Heiko Tietze <heiko.tietze@documentfoundation.org>
|
|
Change-Id: I049a020d1d58f6ed135a7c53c6624092104a1b78
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187658
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: I4a9df517f66ccdf394444c86cd783c201eb2ccae
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187678
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
Change-Id: I7c1be7cb7ce0684f65fbd073523282e50e4dccd0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187743
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Tested-by: Jenkins
|
|
after const clean up of the Slot* argument
objdump -t instdir/program/libsclo.so |grep -v data.rel.ro|grep \\\.data|c++filt|grep -v s_cd|wc
before: 96
after: 78
while:
pmap -px PID|grep 'rw.--'|grep -v anon|awk '{ sum+=$4 } END { print sum }'
for a --with-distro=CPLinux-LOKit build and spawned kit calc process from
2352 to 1860 pages
Change-Id: Ic0ea3e7934ac469f7c6225c252506adb5b62a4c2
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187742
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
it started to crash after
commit ca6b1677cc3d923f0c13b2253b48a0ea90485b41
Author: Xisco Fauli <xiscofauli@libreoffice.org>
Date: Tue Jun 3 11:59:57 2025 +0200
sw: no need to create a dummy attr pool
because that target pool is nullptr in SfxItemPool::GetUserOrPoolDefaultItem
now.
After some testing, there are some reasons why I believe it should be
dropped:
* Although it's an Writer/Web option it only works on Web ( File - New -
HTML document )
* Even if set in a web document, it won't be exported to html. To do so,
it has to be set in the Page Style dialog
* The document background can be changed from the Apperance tab so
basically this tab is duplicated now
This issue would have be caught if tdf#49895 had a test.
I will write a UItest in a follow-up commit
Change-Id: I3087a6942b9e5406de936fdde47d4c9143be55e8
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187754
Tested-by: Jenkins
Reviewed-by: Sahil Gautam <sahil.gautam@collabora.com>
|
|
Change-Id: Ia5c60a3595d452e0ead238975232d6ac2c353c5a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/186898
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Also change access to use GetTabNo() instead (if it makes sense).
Change-Id: I9249c4539a851e34b64d56faa1928d1fb3b8705a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187507
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Per Apple's documentation, the CFDataRef returned by the
CTFontCopyTable() function is "A retained reference to the
font table data as a CFDataRef object. The table data is not
actually copied; however, the data reference must be released."
So, instead of making a copy of the CFDataRef's data, just use
the CFDataRef's data pointer for the HarfBuzz blob and release
the CFDataRef in the blob's destroy function.
Change-Id: Ia670d036b20ca268b203eedcb1abbaad471d5384
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187761
Reviewed-by: Khaled Hosny <khaled@libreoffice.org>
Tested-by: Jenkins
|
|
tables"
This reverts commit af47a78e11d4819bb4b1ae1a1277babd10d815b7.
Reason for revert: seems to cause regression/crash tdf#167455
after the update to harfbuzz 11.2
See also https://github.com/harfbuzz/harfbuzz/issues/5386 that confirms the crash to be related to the patch
Change-Id: I359f716afbc8d94fbdb25e600827ac3c41c2ae0c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187698
Tested-by: Jenkins
Reviewed-by: Patrick Luby <guibomacdev@gmail.com>
Reviewed-by: Khaled Hosny <khaled@libreoffice.org>
|
|
Calculating text bounds can be a costly operation (e.g. for fonts with
CFF table since it has to be calculated from outlines), and leads to
loading potentially large font tables (which is especially an issue on
macOS where we copy font table data when loading, unlike other platforms
where HarfBuzz mmap’s the file). It seems that the text bounds are only
used in one place, so this change calculates the bounds only when
requested.
Change-Id: Ib9017a58dda30b4cc07ceec087cbf4ad45d56d00
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187683
Tested-by: Jenkins
Reviewed-by: Khaled Hosny <khaled@libreoffice.org>
|
|
Change-Id: I2dc35fc72923ee20b9deabf3239af3d859e2fe23
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187707
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Ibcc9a87c673701c69cb71e1fd9af835ce0fe7116
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187706
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Ica973f48c4a7114d84282c07e116a7f7dc6b5a54
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187659
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Hopefully this is a more reliable fix for loading a file that contains
a unicode segment on Windows.
Change-Id: Iabe3ea849858290e782123840d8756ed07159dec
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187762
Reviewed-by: Kohei Yoshida <kohei@libreoffice.org>
Tested-by: Jenkins
|
|
Extends the Writer script type hinting mechanism to natively support
DOC/DOCX script type hinting.
Implements the OOXML w:hint attribute, in both import and export.
Implements the DOC sprmCIdctHint property, in both import and export.
This change also removes the previously-implemented IdctHint emulation,
based on breaking the input text into smaller spans. Writer's script
type hinting mechanism is sufficient to replace this code.
Change-Id: I08f934358164dc5814bfac010d4fac9b3334f947
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187755
Tested-by: Jenkins
Reviewed-by: Jonathan Clark <jonathan@libreoffice.org>
|
|
This will hopefully avoid unnecessary memory spike when loading
a large file, by not loading the entire content of the file to an
in-memory buffer.
Change-Id: Ia46527f0835f8ddc7f67f26e36b289c4e9d66eed
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187760
Reviewed-by: Kohei Yoshida <kohei@libreoffice.org>
Tested-by: Jenkins
|
|
cairo_image_surface_create_for_data has an unfortunate signature,
same as underlying pixman_image_create_bits, which doesn't take
const data, but it's never changed.
Change-Id: I1ba7edbcf0955f8c76a436036c122c56436fed88
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187708
Tested-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
Change-Id: Ib85b8b4b3a2d00e7ef0ecb354fba0b862d60abdd
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187749
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
Change-Id: I8df0c99ea3ba40408e146d9065c2acdfc4c05cec
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187747
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Tested-by: Jenkins
|
|
Change-Id: Ifc319bdc9ed0cf095a99f6ea7cac068f52bc31b1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187748
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
Change-Id: I02777999608137e41cdd4a3182d9884db830b12e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187746
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Tested-by: Jenkins
|
|
enable moving them from .data section to .data.rel.ro.local section
Change-Id: Id986b801c5cb8ef05048752a2ff69edeacd27817
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187745
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Tested-by: Jenkins
|
|
Change-Id: I7f24868387672f4237a9333f91acada1d33b81d8
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187738
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
Change-Id: I7a85f9b189f56a09f724c157a620fa6fff0f92a5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187750
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
Change-Id: Iebd34475da0d005fe3e5fee8554133b448f46cca
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187617
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
Change-Id: Ief6211412cd73e59b7380d0f5ca536a38354e900
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187702
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: Ia2c5543af95abd98919c672a560e8e69c5c3a6f7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187701
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Tested-by: Jenkins
|
|
Open the bugdoc in Writer, looks OK, save to DOCX, open in Word: an
unexpected paragraph appears between the two floating tables.
What happens is that the DOCX import inserts dummy anchor paragraphs
between floating tables, so we can maintain the invariant that each text
node has at most one floating table anchored to it (which simplifies
layout code), but then these dummy text nodes are not filtered out on
the export side.
Fix the problem by scanning the nodes array for such dummy nodes once at
the start of the exporter, omitting such dummy nodes from the export
result and write the affected floating tables when the next text node is
written in the output.
An alternative I considered is to leave
MSWordExportBase::OutputContentNode() unchanged and just make some of
the m_pSerializer calls conditional, so the dummy anchor node is missing
from the export result. The problem is that it needed ~12 conditions and
it would be easy to change the code in the future in a way that some
part of the dummy node markup would be still emitted. So instead skip
the entire text node and write the table with the next node.
Change-Id: Ic15342d431a5c1e3086dc2029df6455ad675e13e
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187696
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
This filter currently does not have a UI, nor allow field type selection
or any other filter settings. The separator character must be ','. It
also does not perform any nice format related post processing at all.
Obviously this is very limiting feature-wise compared to the default CSV
filter, but it generally loads faster & may prove to be useful when you
need to open a very large CSV file.
To use this filter, you need to manually select 'Text CSV (orcus)'
filter from the file open dialog.
Change-Id: Ibeab7c7d2dda5df99c4efe6cfdce608847e4040a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187670
Tested-by: Jenkins
Reviewed-by: Kohei Yoshida <kohei@libreoffice.org>
|
|
Word and Powerpoint store data internal to the chart in .docx and .pptx
formats, respectively. In chartex this uses the <cx:chartData> and <cx:data>
tags, and unlike in pre-2016 charts, is at the chartSpace level rather than
internal to individual series. Provide support for these tags and data
structures.
Note that this does not yet allow .docx and .pptx files with chartex
contents to round-trip successfully (MSO -> LO -> MSO). Additional work
remains needed for this.
Change-Id: I1e91cd0e014787ce19303a906ff76cd6c7c4b0fa
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187426
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
Tested-by: Jenkins
|
|
Add support for style index export, and a CI test for this.
Change-Id: Idb109ef705a6bf8338e18040d447c2806c0781ee
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187425
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
objdump -t instdir/program/libsclo.so |grep -v data.rel.ro|grep \\\.data|c++filt|grep -v s_cd|wc
before: 115
after: 96
Change-Id: I880d1934d7794fb57567f1e542e8a4feff645a83
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187651
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
Amends eaa970c9b9d4d8e9e29e419791f7d0dceb7c4d8c
Chart dialogs with VT style had no default tab selected
Change-Id: Ie4db0bdb526d53bc37a821657415681a29744daa
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187734
Reviewed-by: Heiko Tietze <heiko.tietze@documentfoundation.org>
Tested-by: Heiko Tietze <heiko.tietze@documentfoundation.org>
|
|
Change-Id: I709d0aa7b5aeaaf52510d268df40270c2ff6c1d8
|
|
There were three methods left that use PaintOrStrip in
a mode to directly draw Text: DrawText_ToRectangle,
DrawText_ToPosition and DrawText_ToEditView.
I adapted all three to use PaintOrStrip in a Primitive
creating mode and render using a PrmitiveRenderer.
While StripPortions was already using the Primitive
creating mode the others needed adaptions ranging from
handling clipping over rotation to transformations.
I checked these as good as I could, but since errors
cannot be excluded, I added an ENV_VAR called
DISBALE_EDITENGINE_ON_PRIMITIVES
with which that change can be deactivated. Please use
this by setting it for the office when bugs pop up
whose bibisect point to this change - if it does not
happen with that var set, this change is responsible.
Offering this comes with a cost: Currently PaintOrStrip
*still* offers both modes, but when this change causes
no big problems I will be enabled to remove all direct
paint and stuff done extra there. This will then allow
in the future simpler changes since changes to EditEngine
output will no longer have to be done in two scopes.
For now this allows all remaining places that paint
EditEngine content still directly to continue to work.
This can in the future be extended to get the exact
same context as seq<Primitives> so offers the path to
potential re-use (not only paint, you can process and
transform Primitives in many ways).
A special case is the SD OutlinerView: It uses a set
callback (PaintingFirstLineHdl) to additionally paint
a PageNumber and a Graphic in that mode, but I managed
to replace/extend that to Primitive usage so this is
also prepared for simplifications in the future. It
would also be good to change that callback from it's
current form then.
Change-Id: Idc9233f0ce27c3e4153851f4e45faa24d15af36d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187628
Tested-by: Jenkins
Reviewed-by: Armin Le Grand <Armin.Le.Grand@me.com>
|
|
... property #2"
This reverts my 25.2 commits d4dad2e181fd2dc30c3916fd59c05745023a51c8
and 5a617f77ee803eb65051e93f659ef66433879052.
Obviously this entire thing is just too broken
for a simple guy like me to fix.
It seems like each kind of control has its own definition
of what should happen if the border is not specified.
So much for ODF standardization of default style values...
This makes it virtually impossible to force a setting at import time,
and since styles (without specified fo:borders)
could theoretically be shared between the different control types,
it might also make it impossible to force an export as well?
make CppunitTest_xmloff_style
Change-Id: I0182878a16213ad2c7f06524fe85388aae86551f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187654
Reviewed-by: Justin Luth <jluth@mail.com>
Tested-by: Jenkins
|
|
from commit 20667acbca5df17771a3581632a672d5db4e751b
I think it has served its purpose for now, flushing out the most obvious issues.
The remaining places this triggers are less likely to be real issues
Change-Id: I575da6b899c190e9f867238389d8f864aa18e04c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187695
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
* Update helpcontent2 from branch 'master'
to 5da8be137998369addd60a3aa9faf46f3ad74b8a
- Change argument separator from "," to ";" for LOOKUP
Change-Id: Iccd9c2cd5aa1a91bf6230e778cae65a5ee07ed3e
Reviewed-on: https://gerrit.libreoffice.org/c/help/+/187677
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
Reviewed-by: Olivier Hallot <olivier.hallot@libreoffice.org>
|
|
Previously,the problem arose due to headings of various levels aligned at 0.00"
being affected by paragraph numbering rules. In this patch, some checks to skip
indentation for styles such as headings of
various levels that are aligned at 0.00" are introduced.
Change-Id: I5513d78dc7f32a5caf56d93c4bedfc9bd8a8068d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/184290
Tested-by: Jenkins
Reviewed-by: Sahil Gautam <sahil.gautam@collabora.com>
|
|
Change-Id: I45cbc3ecbe0434ef7c852433266a8957c98ae365
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187449
Reviewed-by: Thorsten Behrens <thorsten.behrens@collabora.com>
Tested-by: Jenkins
|
|
- Replaces ValueSet with IconView widget in the ThemeDialog.
- Cleans up redundant code related to ValueSet and updates the themedialog UI file to support IconView.
Change-Id: Ia36aa6174cde2756aadadcbde4c5bdba2201d1a1
Signed-off-by: Parth Raiyani <parth.raiyani@collabora.com>
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/186644
Reviewed-by: Szymon Kłos <szymon.klos@collabora.com>
Tested-by: Jenkins
|
|
MSWordExportBase::NumberingLevel() would only replace each level's
placeholder once in the format string; rework this so it iterates the
whole format string.
Change-Id: I7f375ac1f96f4fcbac51f7cba1cc67378f4edbb8
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187580
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins CollaboraOffice <jenkinscollaboraoffice@gmail.com>
(cherry picked from commit 58db687fd96fcbb6a6b2909142df2fa01427add9)
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187635
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@collabora.com>
|
|
Change-Id: Ic3714b5b0b508508b27be3a980e404d8cd81d2d4
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187622
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Change-Id: Ia58b5cb0cbe2500a92c013d16ff960c4324c7712
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/187644
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|