After 5c9ba1f47d, the failures on Jenkins
(see 21191d0d89) made it apparent that
(at least some of) Jenkins builds also run in RDP sessions. Since the
tests excluded in commit 9c6142ec26 never
failed in Jenkins builds before, it is wrong to exclude all those tests
in all RDP sessions: our CI would not test those on Windows.
In the meanwhile, I discovered that the system that failed the tests
actually had 16-bit color depth, despite RDP being configured to use
32-bit colors; that was the reason why the colors were modified on
roundtrip. So it is better to test the actual problem to exclude tests.
This reimplements the check that was introduced in commit
9c6142ec26 to test default virtual device
color depth.
Change-Id: I329a3e2d8eca21732c77dcacf15394d1246b2e18
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/113460
Tested-by: Mike Kaganski <mike.kaganski@collabora.com>
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
MacrosTest::loadFromDesktop itself asserts on its return value.
Thus, the additional checks in unit tests are redundant, and only create
noise unrelated to the tested functionality.
SwModelTestBase::load calls SwModelTestBase::loadURL which calls loadFromDesktop from within.
Change-Id: I30061bbc6e13e05f0cfbf17a7910926070b996f3
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/111547
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
...by re-enabling the code temporarily #if'ed-out in
a528392e71 "Fixed/improved
loplugin:cppunitassertequals" (and which then triggers lots of other
lopglugin:cppunitassertequal CPPUNIT_ASSERT -> CPPUNIT_ASSERT_EQUAL warnings).
For two css::uno::Reference equality comparisons in cppu/qa/test_any.cxx, it
was more straightforward to rewrite them with an explicit call to operator ==
(which silences loplugin:cppunitassertequal) than to adapt them to
CPPUNIT_ASSERT_EQUAL's requirement for arguments of identical types.
In sc/qa/unit/ucalc_pivottable.cxx, ScDPItemData needs toString, which has been
implemented trivially for now, but might want to combine that with the
DEBUG_PIVOT_TABLE-only ScDPItemData::Dump.
Change-Id: Iae6d09cf69bd4e52fe4411bba9e50c48e696291c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/110546
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
...for LIBO_INTERNAL_ONLY. These had been missed by
1b43cceaea "Make many OUString functions take
std::u16string_view parameters" because they did not match the multi-overload
pattern that was addressed there, but they nevertheless benefit from being
changed just as well (witness e.g. the various resulting changes from copy() to
subView()).
This showed a conversion from OStringChar to std::string_view to be missing
(while the corresponding conversion form OUStringChar to std::u16string_view was
already present).
The improvement to loplugin:stringadd became necessary to fix
> [CPT] compilerplugins/clang/test/stringadd.cxx
> error: 'error' diagnostics expected but not seen:
> File ~/lo/core/compilerplugins/clang/test/stringadd.cxx Line 43 (directive at ~/lo/core/compilerplugins/clang/test/stringadd.cxx:42): simplify by merging with the preceding assignment [loplugin:stringadd]
> File ~/lo/core/compilerplugins/clang/test/stringadd.cxx Line 61 (directive at ~/lo/core/compilerplugins/clang/test/stringadd.cxx:60): simplify by merging with the preceding assignment [loplugin:stringadd]
> 2 errors generated.
Change-Id: Ie40de0616a66e60e289c1af0ca60aed6f9ecc279
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107602
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
...to "..."+OUStringToOString(...), in preparation of making OUStringToOString
take a std::u16string_view
Change-Id: Ic01ca249d4d0eb58ac959ffdf96471308a3715dc
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107600
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
...in preparation of potential future changes from using OString to using
std::string_view, where OString has an undocumented feature of allowing
construction from a null pointer.
This is mostly the result of a manual audit of potentially problematic getenv
calls across the code base. But there can be other problematic places too, like
the xmlGetProp call in tools/source/xml/XmlWalker.cxx. To identify those,
rtl_{string,uString}_newFromStr aborts now in non-production debug builds when a
null pointer is passed(and all places that hit with a full `make check
screenshot` have been addressed here). Once we are confident that all
problematic places have been identified, we should drop support for the
undocumented feature (see the TODO in sal/rtl/strtmpl.cxx).
Change-Id: I595cc6d4f1cda74add2a3db171323f817d362b08
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/107430
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
It should be used to conditionally disable tests that are unreliable
in non-default-DPI environments. The workaround implemented by use
of the function should be temporary, until a proper fix is found.
Change-Id: Ie236226fcfd0ffb054149efc2b7a1727506c4ad0
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106661
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
which means that some call sites have to change to use
unicode string literals i.e. u"foo" instead of "foo"
Change-Id: Ie51c3adf56d343dd1d1710777f9d2a43ee66221c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/106125
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
The patch contains a larger rework of RecalcPos and connected areas
and covers several bugs. Essentials in short:
Move initialization from RecalcPos to own method and use it in
ScXMLImport::endDocument
Do not change hidden objects, which includes not setting width or
height to zero, and be consistent in using object visibility.
Special handling of vertical flipped customshapes.
Repair anchor on import of line and measure line.
ODF conformance: Create logical rectangle from anchor instead using
size.
Details:
tdf#137044
ScDrawLayer::SetPageSize is called several times while loading a
document. It includes a call to ScDrawLayer::RecalcPos for all cell
anchored objects. An object gets initialized with the first call.
Problem was, that the row heights were not finished at that time and
anchor cells and offsets were partly calculated based on default cell
height. That results in wrong height and offset of objects.
The solution separates initialization from RecalcPos and puts it into
an own method ScDrawLayer::InitializeCellAnchoredObj. This is then
called from ScXMLImport::endDocument when row height settings are
finished.
The call to RecalcPos is not totally removed from SetPageSize but only
excluded while loading, because it is needed for size changes after
the document is loaded.
tdf#137576 partly
For measure lines and ordinary lines, which were anchored 'To cell
(resize with cell)', LibreOffice has written wrong end-cell info to
file. So reopening results in wrong lines. The geometry of lines is
based on two points. Fortunatelly the combination of position of the
cell, which contains the shape, and start and end points gives correct
absolute position of these points.
Solution is, to regenerate the initial ScDrawObjData infos from these
points and do not use the stored end-cell info. For a total fix
implementation of NbcSetSnapRect for SdrMeasureObj is needed, which is
not included here.
tdf#137020
Cell anchored shapes are contained in a cell in file. To determine
size and position of the shape a rectangle is used, so defined, that
after applying transformation you get the desired shape. In case of
custom shapes, a vertical flip is not contained in the transformation
but it is an attribute inside the shape and flip is done at the shape
center and will not change the rectangle.
This rectangle determines start and end addresses and offsets in
ScDrawObjData in rNoRotatedAnchor. The info is used directly in XML
export. It is correctly build while loading the file.
But in case of vertical flipped custom shapes the logical rectangle
of the shape has an additional 180deg rotation. Changing that behavior
is currently out of scope. Therefore special handling of vertical
flipped custom shapes was added.
tdf#99549
ODF specifies that in case of existence of end-cell attribute, size
attributes have to be ignored. But LO has based the logical rectangle
on size. In addition it has written zero width and height in case of
hidden row and cols. Result was, that objects are 'lost' on opening
although they still exists in the file.
With the patch the object size is recalculated from anchor on opening.
tdf#137355, tdf#137044, tdf#115655
The old solution has recalculated the snap rectangle based on current
state of hidden row or column. That has produced shapes of zero width
or height and loss of offset in case start or end cell of the shape
was hidden. In running LO it was partly offset by using cached infos
in ScDrawObjData. That failed in case of save and reload.
Solution is, to only change visible shapes. It is enough to adapt the
shape when it becomes visible. That is introduced in RecalcPos and
SetCellAnchoredFromPosition.
tdf#137216
Shapes anchored to cell were not hidden, if the column of its anchor
was hidden, and undo of hiding an image in a cell by hiding its column
didn't work. Reason was, that the shapes were not set to hidden in the
shapes geometry.
Solution is to copy a similar part from SetRowHidden to SetColHidden.
without bugreport, but detected while debugging
LO has used a cell reference with bHiddenAsZero as true in shape
export. That has resulted in wrong offsets.
Unittest changes:
Test::testGraphicsInGroup()
ScShapeTest::testCustomShapeCellAnchoredRotatedShape()
Set expected values so, that they correspond to anchor in file.
ScFiltersTest::testLegacyCellAnchoredRotatedShape()
FIXME is solved now and the test is adaped to reflect that.
XSheetAnnotationShapeSupplier::testGetAnnotationShape()
Expected value is adapted to the fact, that now annotation shape
gets its position after optimal row height is applied.
Change-Id: Iffee996054ebf79e04044da5520f8d1a8a48b7c1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/104643
Tested-by: Jenkins
Reviewed-by: Regina Henschel <rb.henschel@t-online.de>
do more like
commit 121771e37f
Date: Mon Sep 21 09:17:54 2020 +0200
Make some OUStringLiteral vars constexpr
cause coverity can live with that
Change-Id: I9efd7f848289c4865997a44c6780373068422227
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/103147
Tested-by: Jenkins
Reviewed-by: Caolán McNamara <caolanm@redhat.com>
This is a prerequisite for making conversion from OUStringLiteral to OUString
more efficient at least for C++20 (by replacing its internals with a constexpr-
generated sal_uString-compatible layout with a SAL_STRING_STATIC_FLAG refCount,
conditionally for C++20 for now).
For a configure-wise bare-bones build on Linux, size reported by `du -bs
instdir` grew by 118792 bytes from 1155636636 to 1155755428.
In most places just a u"..." string literal prefix had to be added. In some
places
char const a[] = "...";
variables have been changed to char16_t, and a few places required even further
changes to code (which prompted the addition of include/o3tl/string_view.hxx
helper function o3tl::equalsIgnoreAsciiCase and the additional
OUString::createFromAscii overload).
For all uses of macros expanding to string literals, the relevant uses have been
rewritten as
u"" MACRO
instead of changing the macro definitions. It should be possible to change at
least some of those macro definitions (and drop the u"" from their call sites)
in follow-up commits.
Change-Id: Iec4ef1a057d412d22443312d40c6a8a290dc6144
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/101483
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
So that the UNO constructor work can continue - where we need the
desktop to be disposed properly so that all UNO constructors objects
have their dispose() called, and they can clean up their global state.
We detect this case by changing a SAL_WARN to an assert in
Desktop::disposing()
(*) in ~ScTabViewShell, don't call EnterHandler, because that tries to
create EditEngine's and other stuff, which crashes
(*) Need a fake singleton so that the servicemanager calls dispose()
on the AnalysAddIn and we can clear the global variable there.
Change-Id: Id13b51e17afc16fcbbc65d64281cdf847e4a58cf
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/99640
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
... when the first page has a heading
Regression from commit 17e51f427b (DOCX
import: first page header should always set default headers as well,
2014-11-21), the problem is around how to split a first + follow page
style on import, and then do the opposite on export.
This is described using a single section in OOXML, but Writer has 2 page
styles for this (unlike in case of the DOC filter). This means the
header margin has to be taken from one of these page styles. The above
commit tweaked the import, so the follow page style has the wanted
header margin, but this leads to incorrect layout.
Fix the problem by tweaking the export instead: it has random access to
the doc model, so it can take the header margin from the first page
style if needed, and then the import side can be reverted, leading to
correct layout.
Also remove some leftover debug code in test/, which was added in commit
5352d45dd4 (convert AnimationImport to
fast-parser APIs, 2020-02-18).
Change-Id: I4bbf7271f3a437e8432399bd1e32e9d24190a501
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/94013
Tested-by: Jenkins
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
The line shape itself didn't really have a height, rather it had a
stroke. For some reason, the SVG mask then decides that nothing has to
be painted there, so unless the line is entirely opaque, the line shape
gets lost on export.
Fix the problem by handling transparency similar to the PDF export,
which detects if the whole purpose of the transparency gradient is to
pass around a transparency percentage. We don't need a mask in that
case, we can just use opacity as described at e.g.
<https://developer.mozilla.org/en-US/docs/Web/SVG/Attribute/opacity>.
Change-Id: I0355b9b09b6dd48bbacc5b7cc54fb71866304ef1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/91932
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
The validator cmdline is fairly nontrivial, log it so that in case it
fails, one can invoke it manually without having to hunt down the string
in a debugger.
Change-Id: Id18c6d67f1b3b78da31e8be8c60f703623b80309
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/89875
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
GpgComponentFactory is unused since:
commit 06d7dbb356
Date: Sun May 21 14:28:57 2017 +0200
gpg4libre: share static xmlsec lib between nss and gpg
tableautoformatfield.cxx should have been removed in:
commit 5990beed9a
Date: Fri Aug 29 16:37:43 2014 +0200
Dead code
xmlRow is unused since initial import
Change-Id: Ief186e9ef46238cc8cdb49f4adde52a45a98cbc4
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/89830
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
And...
(*) space out the namespace constant values so I dont
keep forgetting and making them overlap.
(*) Remove CreateDocumentContext from SvXMLImport since it is
now unused.
Change-Id: I30f54bfc1389e91b18e4fee8b83e1b297419899b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/88938
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
"Find explicit casts from signed to unsigned integer in comparison against
unsigned integer, where the cast is presumably used to avoid warnings about
signed vs. unsigned comparisons, and could thus be replaced with
o3tl::make_unsigned for clairty." (compilerplugins/clang/unsignedcompare.cxx)
o3tl::make_unsigned requires its argument to be non-negative, and there is a
chance that some original code like
static_cast<sal_uInt32>(n) >= c
used the explicit cast to actually force a (potentially negative) value of
sal_Int32 to be interpreted as an unsigned sal_uInt32, rather than using the
cast to avoid a false "signed vs. unsigned comparison" warning in a case where
n is known to be non-negative. It appears that restricting this plugin to non-
equality comparisons (<, >, <=, >=) and excluding equality comparisons (==, !=)
is a useful heuristic to avoid such false positives. The only remainging false
positive I found was 0288c8ffec "Rephrase cast
from sal_Int32 to sal_uInt32".
But which of course does not mean that there were no further false positivies
that I missed. So this commit may accidentally introduce some false hits of the
assert in o3tl::make_unsigned. At least, it passed a full (Linux ASan+UBSan
--enable-dbgutil) `make check && make screenshot`.
It is by design that o3tl::make_unsigned only accepts signed integer parameter
types (and is not defined as a nop for unsigned ones), to avoid unnecessary uses
which would in general be suspicious. But the STATIC_ARRAY_SELECT macro in
include/oox/helper/helper.hxx is used with both signed and unsigned types, so
needs a little oox::detail::make_unsigned helper function for now. (The
ultimate fix being to get rid of the macro in the first place.)
Change-Id: Ia4adc9f44c70ad1dfd608784cac39ee922c32175
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/87556
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Noel Grandin