building installation sets takes a long time, and is a .PHONY action.
Having a way to skip that when running make is esp. useful with IDEs
that have no nice way of separating the "run/debug" action from the
build action. (XCode, looking at you and your awkward schemes…)
Use with »make build PKGFORMAT=""«
Change-Id: Iab0c03a4096fe5571a5496b52504b3393e9f3cf6
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137111
Tested-by: Jenkins
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
Beehive and Blue_Curve Impress templates have same bug
as Piano template: presentation:class was missing to
date-time, footer and page-number
Change-Id: I5d4f2463ce7dee3e20eea65206cd4cd91020b989
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137088
Tested-by: Jenkins
Reviewed-by: Adolfo Jayme Barrientos <fitojb@ubuntu.com>
In SF_PythonHelper._PythonDispatcher():
Force a hardcoded call of the methods in the Dialog
service which may potentially be invoked while the
dialog is displayed, bypassing the generic CallByName()
Basic builtin function:
Activate
Center
EndExecute
Resize
(Controls is already hardcoded because returning an array)
Change-Id: I4931f43b37ae5f7337740d1c65b255f4a14ca069
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137084
Tested-by: Jean-Pierre Ledure <jp@ledure.be>
Reviewed-by: Jean-Pierre Ledure <jp@ledure.be>
Tested-by: Jenkins
when opening a large writer file, shaves off about 5% off load time
Change-Id: I501d0ebfbaec572e7c93c1dbc3f0cfaab3938f20
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137099
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
* Update translations from branch 'master'
to c875ee9259eeb47a0ec1767d8d602f86bb793a40
- update translations for 7-4/master
and force-fix errors using pocheck
Change-Id: Ie4a055202e93e4ac5563df42235db5cf2b1c2b16
if the style attributes are set like
<rect x="10" y="10" width="100" height="100" fill="#00D000"
clip-path="url(#myClip)"/>
it works, however, if it uses a css style like
<rect x="10" y="10" width="100" height="100" style="fill:#00D000"
clip-path="url(#myClip)"/>
it fails to get the clipPath from the parent, because the css style
is the direct parent, thus, check one level up
Change-Id: Iff6df95c9fa9da4c2f1a986cca0ad82ab1494353
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137094
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
even when you are a verified non-profit, you still cannot just point to
a donation site/you still would have to offer apple-pay...
the helpmenu one is ugly since it changes the endresult compared to
what is assembled in instdir, but is the least intrusive way for now...
Change-Id: Id348d69371048bdd04d2961cad564fc73f11fd45
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137050
Tested-by: Jenkins
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
The unit test depends on the two previous fixes in this bug report.
-initial load tests highlight import
-export tests saving the character background
-reload tests this shade import
Change-Id: Iceb405841efc6814654061659e5551ac89f3c7b5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137091
Tested-by: Jenkins
Reviewed-by: Justin Luth <jluth@mail.com>
ATK (and AT-SPI) allows to specify 3 different coordinate
types [1]:
1) `ATK_XY_SCREEN`: coordinates relative to the screen.
2) `ATK_XY_WINDOW`: coordinates relative to the widget's
top-level window.
3) `ATK_XY_PARENT` coordinates relative to the widget's
immediate parent.
The `XAccessibleComponent` interface provides equivalents
for 1) and 3), but not 2), and the gtk3 VCL plugin wasn't
really handling 2) so far either.
This adds handling for `ATK_XY_WINDOW`. The position
in the window is calculated by recursively walking up the a11y
hierarchy and summing up the positions relative to the
parent, until a window is reached (or there's no parent,
or none implementing `AtkComponent`).
Also add an explicit check for `ATK_XY_PARENT` if
none of the other two constants is used and warn
and return if that isn't used either, to cover the case
that new constants will be introduced in newer ATK versions.
(That check needs to be conditional, because `ATK_XY_PARENT`
is only defined from ATK 2.30 on, while our minimum
required version is currently 2.28.1.)
The coordinates received when testing this with a
selected Writer paragraph in Accerciser's IPython console
look generally plausible now with this change in place:
In [108]: acc.queryComponent().getPosition(pyatspi.component.XY_SCREEN)
Out[109]: (1939, 417)
In [109]: acc.queryComponent().getPosition(pyatspi.component.XY_WINDOW)
Out[110]: (19, 245)
In [110]: acc.queryComponent().getPosition(pyatspi.component.XY_PARENT)
Out[111]: (19, 113)
In [111]: acc.queryText().getCharacterExtents(0, pyatspi.component.XY_SCREEN)
Out[112]: (2015, 417, 5, 19)
In [112]: acc.queryText().getCharacterExtents(0, pyatspi.component.XY_WINDOW)
Out[113]: (95, 245, 5, 19)
In [113]: acc.queryText().getCharacterExtents(0, pyatspi.component.XY_PARENT)
Out[114]: (76, 0, 5, 19)
(Previously, requesting window-relative coordinates gave the
same result as coordinates relative to direct parent.)
[1] https://docs.gtk.org/atk/enum.CoordType.html
Change-Id: Idf8f7d08f9054a8df28d1ee8fccc791a803de045
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137027
Tested-by: Jenkins
Reviewed-by: Michael Weghorn <m.weghorn@posteo.de>
It was not possible to select an entry from a content control dropdown
using the keyboard, which breaks accessibility.
This had the benefit that the mouse handler code could contain the popup
start calls, but Word can do this with alt-down arrow, so make sense to
add it on our side as well.
Fix the problem by adding SwContentControl::ShouldOpenPopup(), which
knows that dropdowns want a popup with alt-down, and then connecting
SwEditWin::KeyInput() to it.
Date content controls probably will want something similar, but that's
not yet done in this commit.
Change-Id: I6853d3a1661e4d826b96b1b5cb938909875af2ad
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137102
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
Consistently with other places such as Formatting toolbar
and Format Cells -> Font Effects tab page
Change-Id: I7ec5733f01182b25b5d823a43a3639b21ee874e7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137087
Tested-by: Jenkins
Reviewed-by: Gabor Kelemen <kelemeng@ubuntu.com>
EditEngine's character background is a SvxColorItem,
while Writer uses a SvxBrushItem. Thus we have no automatic
transition possible between the two, and during export it
was just being ignored as > RES_UNKNOWNATR_BEGIN.
A unit test will have to wait until import (of shd) works.
P.S. Import of "highlight" was fixed in an earlier commit.
Setting Options - Save - MS Office Compat - As Highlighting
would fully work now.
Change-Id: I8674173a3d1bf1581c09458c3087d6fd39e48a1c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137090
Tested-by: Justin Luth <jluth@mail.com>
Reviewed-by: Justin Luth <jluth@mail.com>
Testing:
- create a new validation list in the A1 cell with some value.
- add a new tab while you are on the A1 cell on the original tab.
or
- switch to an existing tab from the original A1 cell.
TODO: UNIT test
Change-Id: Ic13afdca45cf31fb50a7446b6eb6d64069d30200
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137040
Tested-by: Jenkins
Tested-by: Gabor Kelemen <kelemeng@ubuntu.com>
Tested-by: Balazs Varga <balazs.varga.extern@allotropia.de>
Reviewed-by: Balazs Varga <balazs.varga.extern@allotropia.de>
There currently is only lv-LV.
Change-Id: I39aa38fbf34c44d914aeb6af0f55d820b5567a7a
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137083
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
which is cheaper than copying it when querying from UnoPolyPolygon
Change-Id: If4661dfdff7e277d5ef25a8f0c937e549f8be9f1
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137080
Tested-by: Jenkins
Reviewed-by: Noel Grandin <noel.grandin@collabora.co.uk>
Toggling a checkbox content control was only possible by clicking on it
with the mouse, which breaks accessiblity.
Trying to type into a checkbox content control triggered the read-only
popup (which is good), but there was no special handling for the space
character, which is meant to toggle the checkbox.
Fix the problem by adding a way to query if the current keycode is meant
to interact with the content control, and if so, invoke
GotoContentControl() from SwEditWin::KeyInput(), similar to how the
click handler already did this already.
This only handles checkboxes, but other types can be addressed in a
follow-up commits similarly.
Change-Id: I5c88f2e2f1c2d0f4b28f2ce0b6b1c75b14b7d67c
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137082
Reviewed-by: Miklos Vajna <vmiklos@collabora.com>
Tested-by: Jenkins
... and not only accept for DateOrder::MDY.
Change-Id: Ic17efbdfee5aac1e00d3cee7b14c16875b3dd292
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137076
Reviewed-by: Eike Rathke <erack@redhat.com>
Tested-by: Jenkins
...that were present ever since the surrounding code was added in
95b27dd2e5 "tdf#45705 rework zoom in and zoom out
UI commands in Calc", but for no apparent reason (when MAXZOOM and MINZOOM were
macros expanding to unadorned int literals). But even after
ac2a6ee961 "basegfx: zoomIn() and zoomOut() should
be sal_uInt16" changed MAXZOOM and MINZOOM to be constants of type sal_uInt16,
contemporary compilers are unlikely to emit warnings for these nominally
signed-vs.-unsigned comparisons.
Change-Id: If4481a2e539632dbdca2ce4b13aab63fb84225b7
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137075
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
Also skip spaces before the #, Something like
xlink:href=" #target" is still valid
Change-Id: Ia62e58ca31bfecd283776ec84fcc4bba1836afab
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137074
Tested-by: Jenkins
Reviewed-by: Xisco Fauli <xiscofauli@libreoffice.org>
...<bdc6974f92>
"[clang] Implement ElaboratedType sugaring for types written bare".
For one, it caused diagnostics to now emit 'OString' instead of 'rtl::OUString'
etc., which required adapting a number of tests.
For another, some tests started to fail because the relevant plugins didn't
expect ElaboratedType sugar in places where it now occurs:
> error: 'error' diagnostics expected but not seen:
> File compilerplugins/clang/test/redundantcast.cxx Line 297: redundant cstyle cast from 'Enum1' to 'Enum1' [loplugin:redundantcast]
> File compilerplugins/clang/test/redundantcast.cxx Line 308: redundant cstyle cast from 'Enum1' to 'Enum1' [loplugin:redundantcast]
> error: 'error' diagnostics expected but not seen:
> File compilerplugins/clang/test/referencecasting.cxx Line 25 (directive at compilerplugins/clang/test/referencecasting.cxx:24): the source reference is already a subtype of the destination reference, just use = [loplugin:referencecasting]
> File compilerplugins/clang/test/referencecasting.cxx Line 37 (directive at compilerplugins/clang/test/referencecasting.cxx:36): the source reference is already a subtype of the destination reference, just use = [loplugin:referencecasting]
> File compilerplugins/clang/test/referencecasting.cxx Line 48 (directive at compilerplugins/clang/test/referencecasting.cxx:47): the source reference is already a subtype of the destination reference, just use = [loplugin:referencecasting]
> File compilerplugins/clang/test/referencecasting.cxx Line 100 (directive at compilerplugins/clang/test/referencecasting.cxx:99): the source reference is already a subtype of the destination reference, just use = [loplugin:referencecasting]
> File compilerplugins/clang/test/referencecasting.cxx Line 120 (directive at compilerplugins/clang/test/referencecasting.cxx:119): the source reference is already a subtype of the destination reference, just use = [loplugin:referencecasting]
> File compilerplugins/clang/test/referencecasting.cxx Line 188 (directive at compilerplugins/clang/test/referencecasting.cxx:187): the source reference is already a subtype of the destination reference, just use = [loplugin:referencecasting]
> File compilerplugins/clang/test/referencecasting.cxx Line 200 (directive at compilerplugins/clang/test/referencecasting.cxx:199): the source reference is already a subtype of the destination reference, just use = [loplugin:referencecasting]
> File compilerplugins/clang/test/referencecasting.cxx Line 206 (directive at compilerplugins/clang/test/referencecasting.cxx:205): the source reference is already a subtype of the destination reference, just use = [loplugin:referencecasting]
> error: 'error' diagnostics seen but not expected:
> File compilerplugins/clang/test/typedefparam.cxx Line 42: function param 1 at definition site does not match function param at declaration site, 'FooT *' (aka 'test2::Foo *') vs 'struct Foo *' [loplugin:typedefparam]
> File compilerplugins/clang/test/typedefparam.cxx Line 55: function param 1 at definition site does not match function param at declaration site, 'FooT *' (aka 'test3::Foo *') vs 'Foo *' [loplugin:typedefparam]
> error: 'note' diagnostics seen but not expected:
> File compilerplugins/clang/test/typedefparam.cxx Line 40: declaration site here [loplugin:typedefparam]
> File compilerplugins/clang/test/typedefparam.cxx Line 53: declaration site here [loplugin:typedefparam]
Hopefully, there are not too many places in our plugins left that similarly
don't expect ElaboratedType sugar in certain places, but which are not covered
by tests. At least, a full build didn't turn up any further false positives,
but there may of course be false negatives now that would go undetected.
Change-Id: I9bfb1cfb57df5f6e228b512c19c664d48285b675
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137049
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
For one, Javadoc (unlike Doxygen) doesn't support @attention, so this caused a
> Generating workdir/CustomTarget/odk/docs/java/ref/help-doc.html...
> Note: Custom tags that could override future standard tags: @attention. To avoid potential overrides, use at least one period character (.) in custom tag names.
> Note: Custom tags that were not seen: @attention
warning during the build. And for another, that paragraph was phrased in poor,
broken English and didn't add much value anyway.
Change-Id: I603170ec262d042e92bdff38ae415dda0bcc28bb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137053
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
The problem is that in SwpHints::TryInsertHint() there is a check for
IsInXMLImport() that is presumably an optimization to avoid the
potentially expensive call to BuildPortions().
While LO would only produce 1 text:span referencing an automatic style
around any given character content, this is not required by ODF, and so
other producers may legitimately produce such nested text:span elements.
Unfortunately the current SwpHints::Insert() isn't deterministic, the
RES_TXTATR_AUTOFMT with same start/end will be compared by address in
CompareSwpHtStart() (whereas RES_TXTATR_CHARFMT has a sort number for
this), so the result is going to be a random order.
Change-Id: Id62a7ff5fb85dbe42b7e1a27b0d8b36b74cf1100
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137033
Tested-by: Jenkins
Reviewed-by: Michael Stahl <michael.stahl@allotropia.de>
...<0b89d1d59f>
"[Sema] Add deprecation warnings for some compiler provided __has_* type
traits", which hits in Boost include files,
> In file included from libreofficekit/qa/tilebench/tilebench.cxx:27:
> In file included from external/boost/include/boost/property_tree/json_parser.hpp:31:
> In file included from workdir/UnpackedTarball/boost/boost/property_tree/json_parser.hpp:14:
> In file included from external/boost/include/boost/property_tree/ptree.hpp:31:
> In file included from workdir/UnpackedTarball/boost/boost/property_tree/ptree.hpp:16:
> In file included from workdir/UnpackedTarball/boost/boost/property_tree/string_path.hpp:15:
> In file included from workdir/UnpackedTarball/boost/boost/property_tree/id_translator.hpp:16:
> In file included from external/boost/include/boost/optional.hpp:31:
> In file included from workdir/UnpackedTarball/boost/boost/optional.hpp:15:
> In file included from workdir/UnpackedTarball/boost/boost/optional/optional.hpp:42:
> workdir/UnpackedTarball/boost/boost/type_traits/has_nothrow_constructor.hpp:27:84: error: builtin __has_nothrow_constructor is deprecated; use __is_nothrow_constructible instead [-Werror,-Wdeprecated-builtins]
> template <class T> struct has_nothrow_constructor : public integral_constant<bool, BOOST_HAS_NOTHROW_CONSTRUCTOR(T)>{};
> ^
> workdir/UnpackedTarball/boost/boost/type_traits/intrinsics.hpp:199:48: note: expanded from macro 'BOOST_HAS_NOTHROW_CONSTRUCTOR'
> # define BOOST_HAS_NOTHROW_CONSTRUCTOR(T) (__has_nothrow_constructor(T) && is_default_constructible<T>::value)
> ^
etc.
Change-Id: I08376710e25013b44279532d6e5fc256ed95cb76
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137046
Tested-by: Jenkins
Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
of course it should only be added when the bluetooth part of the
sdremote is enabled, not the other way round...
fixes c78b49ec13
Change-Id: I0085c61a4bafece80be92c56928095c2726e11b4
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137025
Tested-by: Jenkins
Reviewed-by: Christian Lohmaier <lohmaier+LibreOffice@googlemail.com>
Updated XRange to include XFind
Change-Id: If9570fbb5de8d14063a72a4c4c55fcfd491dd496
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137004
Tested-by: Tomaž Vajngerl <quikee@gmail.com>
Reviewed-by: Tomaž Vajngerl <quikee@gmail.com>
Switched the fixed and variable/current field
Change-Id: I18136feebed225f7645b6869d3600ee78f945bfb
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137030
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Tested-by: Jenkins
Reviewed-by: Heiko Tietze <heiko.tietze@documentfoundation.org>
The ctor does not take 'new T', now it takes arguments that
are passed to T ctor.
Change-Id: I04b330d98dd84c86e00ec937a00d1097f21c55c5
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137039
Tested-by: Jenkins
Reviewed-by: Luboš Luňák <l.lunak@collabora.com>
The true hyperlink boundaries are available as SwpHints starts/ends, which are used
in DOC(X) export (see SwWW8AttrIter::OutAttrWithRange).
However, I don't see a reasonable way to expose this information to xmloff, so I
decided instead to just merge the identical hyperlink properties of adjacent ranges
into a single hyperlink. This will allow to fix already split hyperlinks saved in
previous versions. The downside is that this disallows to have separate adjacent
identical hyperlinks - I hope that this would not be a real issue.
Change-Id: I901e6035a5e89bc515b5742c6a5f564c77faf05b
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137013
Tested-by: Jenkins
Reviewed-by: Mike Kaganski <mike.kaganski@collabora.com>
Although it is a highlight, LO doesn't have such a silly
duplicate thing in the editeng code. So just map this as
normal char background instead of char highlight.
As of LO 7.x, we default to exporting as char background anyway,
so highlight is on the way out in LO.
P.S. Highlight is one of 17-ish colors. It is the background
button in the Char panel in MS Word. Background is on Word's
Para panel (even though it is a character property),
and so it can fairly easily be removed in MS Word.
I didn't add a unit test because it doesn't round-trip,
and an import-only test is basically useless.
I'm having a surprisingly hard time finding why it doesn't export.
Change-Id: Iad279b503b9f307994f1d9b96e7d984d6d8b44fc
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/137036
Tested-by: Jenkins
Reviewed-by: Justin Luth <jluth@mail.com>
All zoom functions use sal_uInt16 values. For some reason, basegfx used
long when zoomIn and zoomOut were created in 2012 (see commit
315d2ddc16: "optimized zoom to use more common intervals"), this then
got mass converted to tools::Long in commit 387a88fa25: "use tools::Long
in basegfx..chart2".
So fix is to change zoomIn/Out() to use sal_uInt16.
Change-Id: I2a56d6f58e14f77aeb8741d332fe9bc282eb969f
Reviewed-on: https://gerrit.libreoffice.org/c/core/+/135715
Tested-by: Jenkins
Reviewed-by: Thorsten Behrens <thorsten.behrens@allotropia.de>
