Age | Commit message (Collapse) | Author | Files | Lines |
|
speed up the matching of duplicates in CellAttributeHelper by using
std::set and partial sorting by name
Takes my time from 33s to 6s
Change-Id: I06376c1e253981cb5a3020142d24fa5776513d4d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/166262
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
the comment when calling MakeCellStyleSheet does not match the
implementation, so make the implementation match the comment.
i.e. if there is an existing stylesheet with that name, just return it.
Reduces load time for me from 47s to 33s
Change-Id: If1bd08baf8515933b87c075d9eef04bc0a125357
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/166241
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Reduce the work we do in IndexedStyleSheets::Reindex
takes my test document from 117s to 48s
Change-Id: I2e23b05684d0f2e3a9dc05c0a0fc4e9bbea7008d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/166180
Tested-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Change-Id: I095178b1b3c61829abf5c08b8b81cd7415e3e795
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/162747
Tested-by: Caolán McNamara <caolan.mcnamara@collabora.com>
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
we can use an unordered_map here
Change-Id: I445521d9dec2d49d9df7d2a733a43330d0996ecb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/164924
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
During export from Calc to XLS, ensure that the print range
is converted to a valid XLS print range. If the print range
exceeds XLS format limitations, it will be appropriately cropped
to fit within the constraints.
Change-Id: I2b233e4d57dfbe30f491448d2ee2972b8b717ea3
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/164469
Tested-by: Jenkins
Reviewed-by: Andreas Heinisch <andreas.heinisch@yahoo.de>
|
|
If a document contains a "DRMContent" section that we don't know
how to decrypt, show a warning before showing the un-encrypted
content.
Change-Id: Id7f48b68e1d65ce46434906b2fb3c7912dcc5f49
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/163513
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
If we can't decrypt the document, don't set the SotStorage to
an empty SvRef. It is possible the document contains a "fallback"
un-encrypted document, which we can show. This is typical for some
Excel add-ons, which encrypt the document, but also add a part
that is un-encrypted to write the instructions on how to get the
add-on.
Change-Id: I6079786b1d0dce3a819a70e057699d0a461a5be5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/163511
Tested-by: Jenkins
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
|
|
I tried to create this test as a Cppunittest but it fails with
Run: 1 Failure total: 2 Failures: 1 Errors: 1
warn:vcl:428912:428912:vcl/source/app/svmain.cxx:494: DeInitVCL: some top Windows are still alive
text = "" type = "14FloatingWindow", ptr = 0x5608715e49e0
text = "tdf159707.xls (read-only) — LibreOfficeDev Calc" type = "10WorkWindow", ptr = 0x5608709ab420
Change-Id: I1b9e16d57050fe07f74c632d441e3a06f49e9438
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/163363
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
This patch depends on the previous patch for this bug report,
which allows each sheet to hold its own default row height.
The given height for the row might not be enough to fit
the content. If the row is set to use optimal height,
that will be corrected as soon as the row is edited.
Obviously, it should not require an edit to be correct
on FILEOPEN, so this patch recalculates after loading the document.
This might have a very negative effect on the time needed
to open a file. I couldn't duplicate the XLS method
because Library_scfilt.mk doesn't link to ScSizeDeviceProvider.
The existing UpdateAllRowHeights wasn't designed to
allow any performance improvements, so I made a new one.
The new one is based on the newer ScDocRowHeightUpdater
class - so perhaps the older method can be phased out.
This new UpdateAllRowHeights could replace the
XLS bSetRowHeights clause - with hopefully some performance
benefit.
I'm not sure I'm ready for the regression hate
that would come from the inevitable performance implications.
Testing however doesn't suggest a huge slowdown. I tested with
time make sc.check
before the fix I was getting 16m 4s +- 10s
after the fix I was getting 16m 25s +- 10s
Specific test showing the need for these patches:
make CppunitTest_sc_subsequent_filters_test2 \
CPPUNIT_TEST_NAME=testTdf123026_optimalRowHeight
Impacted unit tests (without the previous patch)
are documented in earlier patchsets.
make CppunitTest_sc_subsequent_export_test \
CPPUNIT_TEST_NAME=testMiscRowHeightExport
make CppunitTest_sc_subsequent_export_test2
make CppunitTest_sc_jumbosheets_test CPPUNIT_TEST_NAME=testTdf134553
Change-Id: I6d020c1a5137dd4f05e20e82b1764a102b7f56d0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/140260
Reviewed-by: Justin Luth <jluth@mail.com>
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
and others, e.g.
ooo101471-3.odf fdo68105-1.ods fdo66241-1.ods
fdo68105-2.ods fdo64646-4.xls ooo101471-2.sxc
Change-Id: I5a639cc6539130bc6a7ada00652d02529544f650
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/153800
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolan.mcnamara@collabora.com>
|
|
This does not fix the actual bug, but is a workaround for a bug
somewhere in the Notebookbar code that was exposed by commit
5b5fb5e9c0580ecb70cf65882865bdaf16498fa5 "sfx2: do reload on SID_EDITDOC
in SfxViewFrame::ExecReload_Impl()".
For an ODF file, the Notebookbar is switched to editable by the call to
SetReadOnlyUI() from ReadOnlyUIGuard, but the XLS file takes a different
path and SetReadOnlyUI() is called in line 819.
Here the SfxViewFrame is not yet connected to the SfxObjectShell so
unsurprisingly it does not affect the UI; more surprising is that moving
the call to around line 859 leads to the SfxHintId::ModeChanged event
being handled in SfxViewFrame::Notify() with lots of things invalidated,
but the NotebookBar is entirely unimpressed by that.
Change-Id: I16f321bb3444eeae76348e20816af02938c3f46b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/151089
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
|
|
... that don't have a style assignment. Typically in ods files
created by older versions or by 3rd party.
- For hidden comments this should make no difference, as we used
to apply the default comment formatting as DF underneath their
defined DF, just now we do that as a style assignment instead.
- Same for comments from xlsx and xls files.
- For visible comments from ods files created by OOo/LO, there
should be no difference either, as such files typically include
the shape formatting in them.
- For visible comments from ods files created by Excel, there
will be a difference, as Excel used to not include the full shape
formatting (known to be the case with Excel 2007 and 2016; can't
test any other version). This resulted with a weird look, e.g.
a line instead of an arrow, a default blue fill color and too
distant shadow, which clearly not how comments supposed to look.
Moreover, the comment will turn to transparent after hiding or
copying the cell, and will revert to the default look anyway
with resaving. Given that we were already enforcing the default
formatting for hidden comments and for foreign formats, I think
it's reasonable to do that for visible comments too (and in
general it's unclear to me why the ODF import treats visible
comments differently than hidden ones).
The main motivation of this change is to aid solving the shadow
issue - see the next commits.
Regarding the comment height change in the testCommentSize test:
This does *not* mean there is a change in that comment's import.
What this test does is to replace the existing comment with
a new comment, and that use the default formatting instead of
inheriting the formatting of the old one. But while the current
default formatting is whatever defined in the Note style, the old
default formatting was actually the draw pool defaults. This is
because we used to apply the comment formatting as DF, and the
relevant svx code (in SdrTextObj::NbcSetText) wasn't aware of the
fact that part of the DF was considered as default in this case.
Which means that this test was actually asserting a buggy behavior.
Change-Id: I37723cce3c719ecaa9c57bef25bcb168e353c55c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/150489
Tested-by: Jenkins
Reviewed-by: Maxim Monastirsky <momonasmon@gmail.com>
|
|
and others, probably since:
commit f02ec0d8489404966596cb0e949bf969e8e66fa2
Date: Wed Mar 22 08:50:15 2023 +0200
move hasUnhiddenData inside chart2::DataSeries
Change-Id: Icf59439f00a40ef036decaa36f3d5af68787c864
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/149688
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
Change-Id: I2082eb3d09ce5f85432807124a80d591e33574d8
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/147344
Tested-by: Jenkins
Reviewed-by: Julien Nabet <serval2412@yahoo.fr>
|
|
seen on importing forum-mso-en4-253817.xls and others
Error: attempt to advance a dereferenceable (start-of-sequence) iterator
which falls outside its valid range.
presumably since:
commit 7674399aac661eb503d7badc53b9a4d68bd9839d
Date: Fri May 27 19:51:40 2022 +0200
try to range-reduce even COUNTIFS if not matching empty cells
bodge this to fill in 0 for missing ranges
Change-Id: If77d8ab887859f11b240975bb837a27785f500a1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/146604
Tested-by: Jenkins
Tested-by: Caolán McNamara <caolanm@redhat.com>
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
|
|
So the actual name is more important than I knew.
For example, VBA maps click macros based on the button name.
So use the MS-provided name and use the SPID to connect to the shape id.
Of the existing unit tests that were affected by this:
-macro-button-form-control.xlsm: button now launches hello macro
-tdf117266_macroButton.xlsm - same file as ^^^
-button-form-control.xls nearly same file as ^^^
-tdf134769.xlsx - checkbox still exists on round-trip
-tdf106181.ods - checkbox still exists, but link to cell still lost.
-checkbox-form-control.xlsx - nearly same as ^^^
All looks good.
Change-Id: If83cf17d60b73c46d55a21b1a46ed320513044cb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/145961
Tested-by: Jenkins
Reviewed-by: Justin Luth <jluth@mail.com>
|
|
For "#FMT" test the original format type for not being NUMBER, the current
format type may be different from finding an edit format.
This constellation did not matter until
commit 4fd1333ba4bb4f2311e9098291154772bd310429
CommitDate: Thu Mar 25 15:00:31 2021 +0100
tdf#140968 tdf#140978 XLSX import: fix lost rounded filters
introduced the bFiltering parameter in which case the edit format is not
used (and bFiltering is also set when collecting the filter entries for
.xls export). In fact the logic should be changed such that obtaining
the edit format isn't even executed in that case. For now just fix the
"#FMT" case to be backported.
Also, the bool return of
SvNumberformat::GetOutputString(double,OUString&,Color**) does not indicate
success or failure, but whether the "star" asterisk spreading was inserted or
not. Contrary to SvNumberformat::GetOutputString(double,sal_uInt16,OUString&)
... cough.
Change-Id: Ic8cbbd283a80c654a9ff22ea36897c9b72b8837d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/144110
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
|
|
Unlike Word, ThisWorksheet cannot hold auto-running
subroutines for Open/Close/New.
This fixes a LO 7.4 regression caused by
commit beb6c62e990599d91ac5d9183164c94d269027d3.
Change-Id: Idb8f72775d9392b306cb924ee776821272b12f3b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/141127
Tested-by: Jenkins
Reviewed-by: Justin Luth <jluth@mail.com>
|
|
regression from
commit 258a8d133d0548c7af51f4852260e9969df288fc
Author: Noel Grandin <noel.grandin@collabora.co.uk>
tdf#150034 very slow opening pathological chart
if nNewLen is < 10, we can get stuck in an infinite loop
because the calculation will not actually decrement nNewLen
Change-Id: I98920deaa087713c44531a7f6506d1d5d32b2e98
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/140949
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
A cell with a line-break specified is no longer a simple string,
so send the string to editeng where it can be properly displayed.
This already works fine for XLS and XLSX.
This change effectively means that text:line-break
will disappear, and become two separate paragraphs.
Given the nature of spreadsheets, the precidence in
various other patches and the failure to properly
display without this, I think I'm fine with
unilaterally changing one paragraph into two.
Change-Id: I7abcbc6f698a4cb319aea5465a8b0393cf63f31f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/140262
Tested-by: Jenkins
Reviewed-by: Justin Luth <jluth@mail.com>
|
|
This reverts commit 613f6e75fbfcc024fc9c3732261cbf2cda6f4ad8.
Reason for revert: it seems all charts in XLS files are affected.
At least I could reproduce it with all the XLS files I tested it with.
Change-Id: I68d13d6f609521ce6af759cf25249656176b66cd
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/139557
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
|
|
There's a xlsx export crash on *2nd* export with data pilot data
of forum-en-38962.ods. We export without crash this .ods to xlsx
*once*, but on saving again to xlsx it crashes. It also crashes with an
export to .xls and then an export to .xlsx. Repeated exports to xls are
crash free.
Already on the first export there are loads of warning messages of
ScDPMember::GetItemData: what data? nDim 0, mnDataId #
ScNameToIndexAccess getByIndex failed com.sun.star.container.NoSuchElementException message: "at sc/source/core/data/dptabsrc.cxx:2267
The pivot tables are broken on both exports.
This smells like some sort of data pilot cache
corruption/insufficiency during .xlsx export but the root cause is
yet unknown. This only fixes the crash symptom.
Change-Id: Ie479f7f0abc4af284e61f6aa5a69943ccdd72eab
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/139611
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
|
|
In current build some links to external xls files are not saved as relative even if
"Save URLs relative to file system" option is checked. This patch aims to
solve this issue.
Change-Id: I6d0984bdcdeef57b227c8ab1353e002fa4355fc9
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/131711
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
revert part of
commit 10c934147d469965dba6abc78efd02759a010b8e
Author: Noel Grandin <noel.grandin@collabora.co.uk>
Date: Tue Apr 23 09:26:46 2019 +0200
tdf#113266 slow opening XLS with 45 MB drawing
in this situation, avoiding setting all the rects dirty speeds up the
load from 70s to 1s.
Change-Id: Ia35119ff20ef5bded9a34f12cce453434dea4383
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137688
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
Value2 ignores date formatting in a way that Value does not.
Extended vba.xls to test this too.
Change-Id: Id75a3dafae2166221731847b7c81e93e8ad835fe
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/133645
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Exported XLS documents lost their formulas opened in MSO.
Regression from commit 12ee423c7549ddd2b86dfc3fc6fed2c617dcca7f
"tdf#144397 tdf#144636 XLSX: cache external named ranges and their formulas".
Change-Id: I009630a41b29bb21349711cc07d0ed181f816374
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/134075
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
When opening CVE-2012-4233-4.xls or ofz5527-1.xls in LO UI, they
show a warning about too many columns (presumably because Excel
as of now supports 16k and LO only 1k columns by default), but
besides stripping the extra columns they otherwise actually load
normally. But ScFiltersTest::testCVEs() fails loading them
just because of the warning. And if 16k columns are forced
for the tests, the files unexpectedly load fine, making the test
fail.
ScDocShell::ConvertFrom() is rather inconsistent on whether
only a warning results in a failed load or not, but most seem
not to return failure on only warnings, and I think that makes
sense, so make code paths for all formats consistently handle
it that way.
Change-Id: I202f29f5a5a44aecd34b84cf2f8180222e8d870a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/130500
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
|
|
Inserting a column messed up formulas in the test
document of tdf#142010. This was a regression
resulted by the tdf#89281 fix for a performance
regression in XLS import related to shared formulas.
Revert of commit ba686b9bd2596811141e4028947334f10799c356
"tdf#89281 fix performance regression of XLS import - cleanup"
and commit b18b5b7edf3d14ef5f0efe53e367f88a423088c4
"tdf#89281 fix performance regression of XLS import".
Change-Id: I96636fb1d84939385efbe7054a4271ff10b88907
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/124151
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
Remove the m_wThis field, and instead just check the ref count to know
if we are being called from our own destructor.
Also re-order ScCellRangesBase::Notify so we do the cheap checks first.
This shaves 2% off the load time
Change-Id: Ie95a4ff46bacba583ec4b3a934f98a9aac3551b9
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/120385
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
This builds on top of commit 1e3263a677b61c718d0fd1be15c066b933f7de18
(XLSX export: handle button form controls, 2021-07-01).
The binary XLS export already had code to turn Calc macro names into
Excel ones, reuse that for XLSX purposes.
Also fix the unwanted named range on export,
oox::xls::FormulaParser::importMacroName() mentions how XLSX doesn't
have matching named ranges for vba macros (while XLS has), mirror this
on the export side as well.
Change-Id: I877b6ba2c2e834a2327482da5cadcddf1b4672bb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/118485
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
Create but not insert/add hidden FILTER_CRITERIA (_xlnm._FilterDatabase)
to ScRangeName, because no need for them, so avoid of creating
extra named ranges by autofilter re-range (shown in Sheet ->
Named Ranges and Expressions -> Manage...).
See also commit 3a2263a20629b0a3e9b4935dd43dc59817d0241c
"mib17: #163146# XLS import: do not skip hidden names[...]".
Change-Id: Ibfd86942a99b251a59394955bc890e6dbe0390f3
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/115189
Tested-by: László Németh <nemeth@numbertext.org>
Reviewed-by: László Németh <nemeth@numbertext.org>
|
|
ja-JP Gengou calendar is an implicit secondary calendar switched
to and from by format codes.
Change-Id: Iea268310ac5e917f2168fec0166754424baa925d
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112511
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
|
|
E.g. ja-JP
GGGE [~gregorian]YYYY
saved as OOXML ended up as
[$-411]ggge\ gregorianyyyy
which when reloaded became
GGGE GREGoRianYYYY
displaying
令和3 R033Ro03ian2021
(whatever Xcl might have done with that).
Change-Id: If8a7d5b837b69c32afacc7a8d7646fedc84ab87a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/112510
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
|
|
* Update helpcontent2 from branch 'master'
to a097662d6678263bb91b6a1dc1f5494fbf47bf38
- tdf#95147 Add note to use Optimal Height with wrapping in XLS files
Change-Id: Ia7a0f3a1cdbb514f95a19cf1496b0b82bd1626cd
Reviewed-on: https://gerrit.libreoffice.org/c/help/+/107308
Tested-by: Jenkins
Reviewed-by: Ilmari Lauhakangas <ilmari.lauhakangas@libreoffice.org>
|
|
A 0-byte ("empty") pptx file is obviously junk input, so it's not
surprising if the catch-all generic_Text filter is chosen to open it in
Writer at the end.
But we can do better: if we really get an empty file URL with an
extension we can recognize, that we can fake the filter type / filter
name, so the empty "presentation" opens in Impress, and also a re-save
works as expected.
This builds on top of commit 8a201be240b6d408d15166be7ffc576b9e123634
(fdo#68903 Import .tsv and .xls plain text files in Calc by default,
2013-10-27), just the new way works for all supported file extensions
and also with filters which would not handle empty input (e.g. pptx
refuses the import if the ZIP storage is broken).
Change-Id: Ie01650a5eb6ca42c35e090133965467b621bb526
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/104939
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
|
|
This partially reverts fixes for i#63105 and i#71033, i.e. parts of
commits 6ca96e7aee937affef218e0bb892e6904f4ef358 and
35777f0b401ba7bee21a24c32739e86598ba608b.
Instead of using hard-coded Win-1252 for Excel files by default, use
LibreOffice locale language or default document language (as used for
old WMF, DXF and RTF files - see
20f6a6b159c69771dc0e087f63b6c701908e32e2,
1a9a77f84cac68bd5374df3e9ee4df88dc87a0ac, and
f26bbdb386f9e66b1a154d0589b1a446ea2dd8d5).
When a font in XLS specifies ANSI_CHARSET (0), allow using the default
encoding, not the corresponding Win-1252. This mimics Excel behaviour
for such fonts.
Do not use first font's encoding as file default encoding, as Excel
doesn't do that.
Do not consider font-specified encoding in ImportExcel::Rstring to
mimic Excel behaviour for cell number formats (but keep using it in
ImportExcel::ReadLabel for cell content).
Change-Id: I9860013a1244dc12f06f30c81a67909012c04913
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/93612
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
Testing on Windows with UI scaling set to 150%, having screen size
1920x1080, ScMacrosTest::testVba failed in window2.xls (SplitColumn)
because it tried to split at pixel position 486, while maximum split
position was 460.
This changes tested SplitColumn value in the test from 243 to 193.
Change-Id: Iec9b25dc7b2a700adf23d7ec8983f393ba603191
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/85983
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
by locking the underlying SdrModel while the comments are added, so we
avoid the reformatting that happends when the model is unlocked
Takes the time from 3m30 to 1m for me.
Change-Id: Ibc3d2b1ded548d5eee7715d9a655a0af61b0eb2b
Reviewed-on: https://gerrit.libreoffice.org/85523
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
XLS stores default column width as 16-bit unsigned integer in the range
from 0 to 255 inclusive ([MS-XLS] 2.4.89 DefColWidth), unlike in OOXML
where any floating-point value is valid. Additionally, some correction
is used, introduced in commit 555d702903fb0857122024e1ab78a72d122d3f16
(basis of empirical formula in XclTools::GetXclDefColWidthCorrection
is unclear). So in XLS, when the default is calculated, we need to take
into account if the resulting stored value will actually represent our
calculated value. If not, then just ignore the calculated value, and
use arbitrary 8 as the default.
With that, following IsDefWidth will correctly check if passed width
is represented by the stored default or not. All widths that can't be
represented as integral count chars in DefColWidth will have fUserSet
set in corresponding ColInfo records, thus correctly keeping widths.
Change-Id: Iedcc5583c861f5b18a422a9b279c48cff729cbc5
Reviewed-on: https://gerrit.libreoffice.org/83613
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
|
|
... in RTFDocumentImpl::checkUnicode(); see ooo86460-1.xls [sic]
for an example.
There is another caller of text() in rtfdispatchdestination.cxx:311 but
it turns out that buffered text was created by text() in the first
place.
This shouldn't be a problem for DOCX because XML 1.0 doesn't allow the
bad control characters anyway so the sax parser should report an error
in that case.
Change-Id: Ice45e1c3c8c7db668a4cfb8364e42addea1777ce
Reviewed-on: https://gerrit.libreoffice.org/81697
Reviewed-by: Michael Stahl <michael.stahl@cib.de>
Tested-by: Michael Stahl <michael.stahl@cib.de>
|
|
extend the find-map so we can do partial-match searching.
Makes it 10x faster for me.
Change-Id: I1952a221a919707af078fac9fd1eb63781d9188d
Reviewed-on: https://gerrit.libreoffice.org/81488
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
|
|
a) treat shared/Scripts equivalently to document scripts
This doesn't automatically warn/block running those scripts when used in a
freshly loaded document on its own however
because DocumentMacroMode::checkMacrosOnLoading will see at...
if ( m_xData->m_rDocumentAccess.documentStorageHasMacros() || hasMacroLibrary() )
that the document contains no macros and flip the allow macros flag to true so
that potentially new uses of macros added by the user during the edit are
allowed to run
b) so, add an additional flag to indicate existence of use of macros in a document
c) for odf import, set it when a script:event-listener tag is encountered
d) for html import when registerScriptEvents or SwFormatINetFormat::SetMacroTable is called
e) for doc import when Read_F_Macro or StoreMacroCmds is called as well for good measure
f) for xls import when registerScriptEvent or ScMacroInfo::SetMacro is called
g) for oox import when VbaProject::attachMacros is called
Change-Id: Ic1203d8ec7dfc217aa217135033ae9db2888e19b
Reviewed-on: https://gerrit.libreoffice.org/77131
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
Tested-by: Caolán McNamara <caolanm@redhat.com>
|
|
Automatic axis labels in xls format are created on a different code
path than the ones in OOXML, translate these too
Change-Id: Id1e8a10accf765d40d71c4fcdfec2c1acee832a2
Reviewed-on: https://gerrit.libreoffice.org/75202
Tested-by: Jenkins
Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
|
|
The same string for XLS format is on a different code path
localize it too, just in case
Change-Id: Iacaf59080cc320498867680fc13836edbe9afe49
Reviewed-on: https://gerrit.libreoffice.org/74820
Tested-by: Jenkins
Reviewed-by: Andras Timar <andras.timar@collabora.com>
|
|
...when loading xls/tdf117254-4.xls as obtained by
bin/get-bugzilla-attachments-by-mimetype (i.e., the attachment at
<https://bugs.documentfoundation.org/show_bug.cgi?id=117254#c5>):
> chart2/source/tools/PolynomialRegressionCurveCalculator.cxx:189:32: runtime error: division by zero
> #0 in chart::PolynomialRegressionCurveCalculator::recalculateRegression(com::sun::star::uno::Sequence<double> const&, com::sun::star::uno::Sequence<double> const&) at chart2/source/tools/PolynomialRegressionCurveCalculator.cxx:189:32
> #1 in chart::VSeriesPlotter::createRegressionCurvesShapes(chart::VDataSeries const&, com::sun::star::uno::Reference<com::sun::star::drawing::XShapes> const&, com::sun::star::uno::Reference<com::sun::star::drawing::XShapes> const&, bool) at chart2/source/view/charttypes/VSeriesPlotter.cxx:1343:22
> #2 in chart::AreaChart::impl_createSeriesShapes() at chart2/source/view/charttypes/AreaChart.cxx:539:17
> #3 in chart::AreaChart::createShapes() at chart2/source/view/charttypes/AreaChart.cxx:965:5
> #4 in chart::ChartView::impl_createDiagramAndContent(chart::CreateShapeParam2D const&, com::sun::star::awt::Size const&) at chart2/source/view/main/ChartView.cxx:1608:25
> #5 in chart::ChartView::createShapes2D(com::sun::star::awt::Size const&) at chart2/source/view/main/ChartView.cxx:3037:41
> #6 in chart::ChartView::createShapes() at chart2/source/view/main/ChartView.cxx:2506:5
[...]
Leaving aRSquared initialized to 0.0 when the divisor is zero is in line with
the code prior to 00cb825ab3f16a94f1e8311ba0c24f72588e788e "fdo#75538 R^2
calculation for trendline similar to LINEST function" and with the recent nearby
change f44d14e5f3909a4271034ba02da92ed2dee7b89c "Avoid
-fsanitize=float-divide-by-zero".
Change-Id: If2c17ad178788982729f647b4c695d3788fad500
Reviewed-on: https://gerrit.libreoffice.org/74068
Tested-by: Jenkins
Reviewed-by: Laurent BP <laurent.balland-poirier@laposte.net>
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
...as happens with `--convert-to pdf xls/rhbz499466-2.xls` with
xls/rhbz499466-2.xls as obtained by bin/get-bugzilla-attachments-by-mimetype
(i.e., the attachment at
<https://bugzilla.redhat.com/show_bug.cgi?id=499466#c1>):
> chart2/source/tools/PolynomialRegressionCurveCalculator.cxx:73:14: runtime error: division by zero
> #0 in chart::PolynomialRegressionCurveCalculator::recalculateRegression(com::sun::star::uno::Sequence<double> const&, com::sun::star::uno::Sequence<double> const&) at chart2/source/tools/PolynomialRegressionCurveCalculator.cxx:73:14
> #1 in chart::VSeriesPlotter::createRegressionCurvesShapes(chart::VDataSeries const&, com::sun::star::uno::Reference<com::sun::star::drawing::XShapes> const&, com::sun::star::uno::Reference<com::sun::star::drawing::XShapes> const&, bool) at chart2/source/view/charttypes/VSeriesPlotter.cxx:1341:22
> #2 in chart::AreaChart::impl_createSeriesShapes() at chart2/source/view/charttypes/AreaChart.cxx:539:17
> #3 in chart::AreaChart::createShapes() at chart2/source/view/charttypes/AreaChart.cxx:965:5
> #4 in chart::ChartView::impl_createDiagramAndContent(chart::CreateShapeParam2D const&, com::sun::star::awt::Size const&) at chart2/source/view/main/ChartView.cxx:1608:25
> #5 in chart::ChartView::createShapes2D(com::sun::star::awt::Size const&) at chart2/source/view/main/ChartView.cxx:3037:41
[...]
Change-Id: I1ef70aa90256a5fcc84151d869d8ffc6acfbbb0e
Reviewed-on: https://gerrit.libreoffice.org/74055
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
...as happens when loading xls/fdo76137-3.xls as obtained by
bin/get-bugzilla-attachments-by-mimetype (i.e, the attachment at
<https://bugs.documentfoundation.org/show_bug.cgi?id=76137#c6>):
> chart2/source/view/charttypes/PieChart.cxx:1110:81: runtime error: signed integer overflow: 2025540036 + 157276681 cannot be represented in type 'int'
> #0 in chart::PieChart::rearrangeLabelToAvoidOverlapIfRequested(com::sun::star::awt::Size const&) at chart2/source/view/charttypes/PieChart.cxx:1110:81
> #1 in chart::ChartView::impl_createDiagramAndContent(chart::CreateShapeParam2D const&, com::sun::star::awt::Size const&) at chart2/source/view/main/ChartView.cxx:1651:23
> #2 in chart::ChartView::createShapes2D(com::sun::star::awt::Size const&) at chart2/source/view/main/ChartView.cxx:3037:41
Change-Id: I52a8363db2ca37dbcf56e1c136d9ee2223add945
Reviewed-on: https://gerrit.libreoffice.org/74048
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
...as happens with `--convert-to pdf xls/fdo53482-2.xls` with xls/fdo53482-2.xls
as obtained by xls/fdo53482-2.xls (i.e., the attachment at
<https://bugs.documentfoundation.org/show_bug.cgi?id=53482#c25>):
> chart2/source/tools/PolynomialRegressionCurveCalculator.cxx:190:38: runtime error: division by zero
> #0 in chart::PolynomialRegressionCurveCalculator::recalculateRegression(com::sun::star::uno::Sequence<double> const&, com::sun::star::uno::Sequence<double> const&) at chart2/source/tools/PolynomialRegressionCurveCalculator.cxx:190:38
> #1 in chart::VSeriesPlotter::createRegressionCurvesShapes(chart::VDataSeries const&, com::sun::star::uno::Reference<com::sun::star::drawing::XShapes> const&, com::sun::star::uno::Reference<com::sun::star::drawing::XShapes> const&, bool) at chart2/source/view/charttypes/VSeriesPlotter.cxx:1341:22
> #2 in chart::AreaChart::impl_createSeriesShapes() at chart2/source/view/charttypes/AreaChart.cxx:539:17
> #3 in chart::AreaChart::createShapes() at chart2/source/view/charttypes/AreaChart.cxx:965:5
> #4 in chart::ChartView::impl_createDiagramAndContent(chart::CreateShapeParam2D const&, com::sun::star::awt::Size const&) at chart2/source/view/main/ChartView.cxx:1608:25
> #5 in chart::ChartView::createShapes2D(com::sun::star::awt::Size const&) at chart2/source/view/main/ChartView.cxx:3037:41
The division aSumError/aSumTotal is there ever since the code's introduction in
e6a0cc2d6cb37bf4e04861173c7e55b307513778 "fdo#35712 polynomial and moving
average regression lines", and quikee suggested this fix on IRC.
Change-Id: Ic9cd7b8aaa8451e1230fca8137bad29e28f6477f
Reviewed-on: https://gerrit.libreoffice.org/74036
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|
|
`--convert-to pdf xls/fdo37212-1.xls`, with xls/fdo37212-1.xls as obtained by
bin/get-bugzilla-attachments-by-mimetype (i.e., the attachment at
<https://bugs.documentfoundation.org/show_bug.cgi?id=37212#c0>) needs a timeout
of more than 5 min for my plain --enable-dbgutil build and a timeout of more
than 10 min for my ASan+UBSan --enable-dbgutil build.
Change-Id: Ia20140049a2f95c4ac006860599e07687cc594eb
Reviewed-on: https://gerrit.libreoffice.org/73997
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
|