mir/libreoffice
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
83e957bfbfa22980461b6c0f07231902549504d4
libreoffice/vcl/qt5/Qt5Widget.cxx

432 lines
13 KiB
C++
Raw Normal View History

KF5 add QWidget to Kf5Frame Change-Id: I53845519d0dda324c9544f057b18c6afd4cf858a
2017-10-27 13:15:37 +02:00
/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
/*
* This file is part of the LibreOffice project.
*
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/.
*
* This file incorporates work covered by the following license notice:
*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed
* with this work for additional information regarding copyright
* ownership. The ASF licenses this file to you under the Apache
* License, Version 2.0 (the "License"); you may not use this file
* except in compliance with the License. You may obtain a copy of
* the License at http://www.apache.org/licenses/LICENSE-2.0 .
*/
Various trivial loplugin fixes in --enable-qt5 Change-Id: I9539eb77f663e1174919ae801495801f81571710
2018-06-01 15:28:26 +02:00
#include <Qt5Widget.hxx>
QT5 rename from KF5 Move out of unx, as this will eventually compile on other OS platforms. At least currently it doesn't contain platform dependant code. Change-Id: Iea0bebf574201881ea158381fe7ba8af2a9a6488
2017-10-30 18:45:46 +01:00
#include <Qt5Widget.moc>
KF5 add QWidget to Kf5Frame Change-Id: I53845519d0dda324c9544f057b18c6afd4cf858a
2017-10-27 13:15:37 +02:00
Various trivial loplugin fixes in --enable-qt5 Change-Id: I9539eb77f663e1174919ae801495801f81571710
2018-06-01 15:28:26 +02:00
#include <Qt5Frame.hxx>
#include <Qt5Graphics.hxx>
#include <Qt5Tools.hxx>
KF5 minimal graphics support This way we pass the "No mpGraphics set" message. Change-Id: I097a823c295b251304dfbec8a71eba31b3d2b9ef
2017-10-31 01:07:06 +01:00
Qt5 fix the clang plugin build ... and fix the global Qt* and wrong included headers. Change-Id: I436628d3a2d4469d5cc72f9fe7f394426d86fb2e Reviewed-on: https://gerrit.libreoffice.org/65023 Tested-by: Jenkins Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
2018-12-12 12:33:26 +00:00
#include <QtCore/QMimeData>
#include <QtGui/QDrag>
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
#include <QtGui/QFocusEvent>
KF5 minimal graphics support This way we pass the "No mpGraphics set" message. Change-Id: I097a823c295b251304dfbec8a71eba31b3d2b9ef
2017-10-31 01:07:06 +01:00
#include <QtGui/QImage>
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
#include <QtGui/QKeyEvent>
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
#include <QtGui/QMouseEvent>
KF5 minimal graphics support This way we pass the "No mpGraphics set" message. Change-Id: I097a823c295b251304dfbec8a71eba31b3d2b9ef
2017-10-31 01:07:06 +01:00
#include <QtGui/QPainter>
#include <QtGui/QPaintEvent>
tdf#120777 KDE5: Dialogs get blank when they are resized On window resize qt5 only redraws changed parts of widgets. If resize is minor (i.e. height has increased by 1 pixel, for example), qt5 may consider most parts of widget not changed and skip redrawing them and redraw only certain widget elements. But if cairo is used for drawing, on resize previously drawed image of widget is discarded and new one of different size is created. New image is empty, but qt5 doesn't issue redraw for whole widget. To mitigate this issue, data from old image of widget should be copied over to image of new widget, qt5 will redraw it partially or fully if necessary. Change-Id: Id950074efece9072bbfc002dfcb6ead813d5aeff Reviewed-on: https://gerrit.libreoffice.org/62698 Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de> Tested-by: Jenkins
2018-10-31 15:13:42 +03:00
#include <QtGui/QResizeEvent>
QT5 queue paint update on show Something like gtk_widget_set_redraw_on_allocate. On resize of dialogs, we still get artefacts, but at least you see something without forcing an event. And since there isn't any Qt5 theming yet, we let VCL paint the menus with a border. Change-Id: Ia2ff6b199ee25e481e3fb29a921b0d3d490f1535
2017-11-07 13:13:10 +01:00
#include <QtGui/QShowEvent>
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
#include <QtGui/QWheelEvent>
Templatize Qt5Widget class to distinguish between QMainWindow (top-level SalFrame) and QWidget (everything else). The former are going to receive a native menu Change-Id: I9023d1e464345f96f13967f5e2c8b0716890fbc9
2018-05-14 15:39:30 +02:00
#include <QtWidgets/QMainWindow>
Qt5 fix the clang plugin build ... and fix the global Qt* and wrong included headers. Change-Id: I436628d3a2d4469d5cc72f9fe7f394426d86fb2e Reviewed-on: https://gerrit.libreoffice.org/65023 Tested-by: Jenkins Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
2018-12-12 12:33:26 +00:00
#include <QtWidgets/QToolTip>
#include <QtWidgets/QWidget>
KF5 minimal graphics support This way we pass the "No mpGraphics set" message. Change-Id: I097a823c295b251304dfbec8a71eba31b3d2b9ef
2017-10-31 01:07:06 +01:00
QT5 implement cairo rendering path Instead of QImage, this uses cairo_surface_t internally and just blits the composed image in the Qt5Widgets paint function. To enable this rendering path set SAL_VCL_QT5_USE_CAIRO. Change-Id: Ieddda9bad2596ce46d7d07d4d7060e40d44997db
2017-10-24 19:49:45 +02:00
#include <cairo.h>
#include <headless/svpgdi.hxx>
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::paintEvent(QPaintEvent* pEvent)
KF5 add QWidget to Kf5Frame Change-Id: I53845519d0dda324c9544f057b18c6afd4cf858a
2017-10-27 13:15:37 +02:00
{
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
QPainter p(this);
kde5: add more missing pieces to Qt5Frame - some cleanup - ShowFullScreen - ToTop/SetAlwaysOnTop - CaptureMouse/SetPointerPos - Beep/Flush - ClipRegion functions Change-Id: Ieafdbdae3f808879b9ad0ebed0fa99a2e4f7ee2e Reviewed-on: https://gerrit.libreoffice.org/60008 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-09-04 09:17:11 +02:00
if (!m_pFrame->m_bNullRegion)
p.setClipRegion(m_pFrame->m_aRegion);
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
if (m_pFrame->m_bUseCairo)
QT5 implement cairo rendering path Instead of QImage, this uses cairo_surface_t internally and just blits the composed image in the Qt5Widgets paint function. To enable this rendering path set SAL_VCL_QT5_USE_CAIRO. Change-Id: Ieddda9bad2596ce46d7d07d4d7060e40d44997db
2017-10-24 19:49:45 +02:00
{
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
cairo_surface_t* pSurface = m_pFrame->m_pSurface.get();
cairo_surface_flush(pSurface);
QT5 implement cairo rendering path Instead of QImage, this uses cairo_surface_t internally and just blits the composed image in the Qt5Widgets paint function. To enable this rendering path set SAL_VCL_QT5_USE_CAIRO. Change-Id: Ieddda9bad2596ce46d7d07d4d7060e40d44997db
2017-10-24 19:49:45 +02:00
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
QImage aImage(cairo_image_surface_get_data(pSurface), size().width(), size().height(),
Qt5_DefaultFormat32);
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
p.drawImage(pEvent->rect().topLeft(), aImage, pEvent->rect());
QT5 implement cairo rendering path Instead of QImage, this uses cairo_surface_t internally and just blits the composed image in the Qt5Widgets paint function. To enable this rendering path set SAL_VCL_QT5_USE_CAIRO. Change-Id: Ieddda9bad2596ce46d7d07d4d7060e40d44997db
2017-10-24 19:49:45 +02:00
}
else
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
p.drawImage(pEvent->rect().topLeft(), *m_pFrame->m_pQImage, pEvent->rect());
KF5 minimal graphics support This way we pass the "No mpGraphics set" message. Change-Id: I097a823c295b251304dfbec8a71eba31b3d2b9ef
2017-10-31 01:07:06 +01:00
}
tdf#120777 KDE5: Dialogs get blank when they are resized On window resize qt5 only redraws changed parts of widgets. If resize is minor (i.e. height has increased by 1 pixel, for example), qt5 may consider most parts of widget not changed and skip redrawing them and redraw only certain widget elements. But if cairo is used for drawing, on resize previously drawed image of widget is discarded and new one of different size is created. New image is empty, but qt5 doesn't issue redraw for whole widget. To mitigate this issue, data from old image of widget should be copied over to image of new widget, qt5 will redraw it partially or fully if necessary. Change-Id: Id950074efece9072bbfc002dfcb6ead813d5aeff Reviewed-on: https://gerrit.libreoffice.org/62698 Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de> Tested-by: Jenkins
2018-10-31 15:13:42 +03:00
void Qt5Widget::resizeEvent(QResizeEvent* pEvent)
KF5 minimal graphics support This way we pass the "No mpGraphics set" message. Change-Id: I097a823c295b251304dfbec8a71eba31b3d2b9ef
2017-10-31 01:07:06 +01:00
{
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
if (m_pFrame->m_bUseCairo)
QT5 implement cairo rendering path Instead of QImage, this uses cairo_surface_t internally and just blits the composed image in the Qt5Widgets paint function. To enable this rendering path set SAL_VCL_QT5_USE_CAIRO. Change-Id: Ieddda9bad2596ce46d7d07d4d7060e40d44997db
2017-10-24 19:49:45 +02:00
{
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
int width = size().width();
int height = size().height();
Set Qt5Widget to be a central widget of QMainWindow this is meant to solve the problem of native menu bar overlapping w/ non-native, as well as the inability to place an object or select text dragging the mouse cursor w/ LMB pressed Change-Id: I29f590ebf79d1ecc7e17b402125384cf13774bf3 Reviewed-on: https://gerrit.libreoffice.org/58171 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-09 17:14:14 +02:00
if (m_pFrame->m_pSvpGraphics)
{
cairo_surface_t* pSurface
= cairo_image_surface_create(CAIRO_FORMAT_ARGB32, width, height);
cairo_surface_set_user_data(pSurface, SvpSalGraphics::getDamageKey(),
&m_pFrame->m_aDamageHandler, nullptr);
m_pFrame->m_pSvpGraphics->setSurface(pSurface, basegfx::B2IVector(width, height));
tdf#120777 KDE5: Dialogs get blank when they are resized On window resize qt5 only redraws changed parts of widgets. If resize is minor (i.e. height has increased by 1 pixel, for example), qt5 may consider most parts of widget not changed and skip redrawing them and redraw only certain widget elements. But if cairo is used for drawing, on resize previously drawed image of widget is discarded and new one of different size is created. New image is empty, but qt5 doesn't issue redraw for whole widget. To mitigate this issue, data from old image of widget should be copied over to image of new widget, qt5 will redraw it partially or fully if necessary. Change-Id: Id950074efece9072bbfc002dfcb6ead813d5aeff Reviewed-on: https://gerrit.libreoffice.org/62698 Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de> Tested-by: Jenkins
2018-10-31 15:13:42 +03:00
UniqueCairoSurface old_surface(m_pFrame->m_pSurface.release());
Set Qt5Widget to be a central widget of QMainWindow this is meant to solve the problem of native menu bar overlapping w/ non-native, as well as the inability to place an object or select text dragging the mouse cursor w/ LMB pressed Change-Id: I29f590ebf79d1ecc7e17b402125384cf13774bf3 Reviewed-on: https://gerrit.libreoffice.org/58171 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-09 17:14:14 +02:00
m_pFrame->m_pSurface.reset(pSurface);
tdf#120777 KDE5: Dialogs get blank when they are resized On window resize qt5 only redraws changed parts of widgets. If resize is minor (i.e. height has increased by 1 pixel, for example), qt5 may consider most parts of widget not changed and skip redrawing them and redraw only certain widget elements. But if cairo is used for drawing, on resize previously drawed image of widget is discarded and new one of different size is created. New image is empty, but qt5 doesn't issue redraw for whole widget. To mitigate this issue, data from old image of widget should be copied over to image of new widget, qt5 will redraw it partially or fully if necessary. Change-Id: Id950074efece9072bbfc002dfcb6ead813d5aeff Reviewed-on: https://gerrit.libreoffice.org/62698 Reviewed-by: Samuel Mehrbrodt <Samuel.Mehrbrodt@cib.de> Tested-by: Jenkins
2018-10-31 15:13:42 +03:00
int min_width = qMin(pEvent->oldSize().width(), pEvent->size().width());
int min_height = qMin(pEvent->oldSize().height(), pEvent->size().height());
SalTwoRect rect(0, 0, min_width, min_height, 0, 0, min_width, min_height);
m_pFrame->m_pSvpGraphics->copySource(rect, old_surface.get());
Set Qt5Widget to be a central widget of QMainWindow this is meant to solve the problem of native menu bar overlapping w/ non-native, as well as the inability to place an object or select text dragging the mouse cursor w/ LMB pressed Change-Id: I29f590ebf79d1ecc7e17b402125384cf13774bf3 Reviewed-on: https://gerrit.libreoffice.org/58171 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-09 17:14:14 +02:00
}
QT5 implement cairo rendering path Instead of QImage, this uses cairo_surface_t internally and just blits the composed image in the Qt5Widgets paint function. To enable this rendering path set SAL_VCL_QT5_USE_CAIRO. Change-Id: Ieddda9bad2596ce46d7d07d4d7060e40d44997db
2017-10-24 19:49:45 +02:00
}
else
{
tdf#120777 KDE5: Update initialization of QImages Make sure that created empty QImages are filled with transparent pixels. Copying data from previous QImage on widget resize removes blanking on window resize when qt5 vcl plugin is used. Change-Id: If072a4b8b334bc87dbe4aaea9aa8774bb5e202ee Reviewed-on: https://gerrit.libreoffice.org/63029 Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de> Tested-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-11-07 17:48:06 +03:00
QImage* pImage = nullptr;
if (m_pFrame->m_pQImage)
pImage = new QImage(
m_pFrame->m_pQImage->copy(0, 0, pEvent->size().width(), pEvent->size().height()));
else
{
pImage = new QImage(size(), Qt5_DefaultFormat32);
pImage->fill(Qt::transparent);
}
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
m_pFrame->m_pQt5Graphics->ChangeQImage(pImage);
m_pFrame->m_pQImage.reset(pImage);
QT5 implement cairo rendering path Instead of QImage, this uses cairo_surface_t internally and just blits the composed image in the Qt5Widgets paint function. To enable this rendering path set SAL_VCL_QT5_USE_CAIRO. Change-Id: Ieddda9bad2596ce46d7d07d4d7060e40d44997db
2017-10-24 19:49:45 +02:00
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
m_pFrame->maGeometry.nWidth = size().width();
m_pFrame->maGeometry.nHeight = size().height();
Qt5 set and update the frame geometry The geometry is used for mouse hit tests for mapping LO mouse events to popup positions. This way (popup) menus become at at least usable. Change-Id: Ic2b409d7a8937320c70b9f377374fe7d51e4ebf3 Reviewed-on: https://gerrit.libreoffice.org/47276 Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Jan-Marek Glogowski <glogow@fbihome.de>
2017-12-08 11:37:18 +00:00
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
m_pFrame->CallCallback(SalEvent::Resize, nullptr);
KF5 minimal graphics support This way we pass the "No mpGraphics set" message. Change-Id: I097a823c295b251304dfbec8a71eba31b3d2b9ef
2017-10-31 01:07:06 +01:00
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::handleMouseButtonEvent(QMouseEvent* pEvent, bool bReleased)
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
{
SalMouseEvent aEvent;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
switch (pEvent->button())
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
{
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
case Qt::LeftButton:
aEvent.mnButton = MOUSE_LEFT;
break;
case Qt::MidButton:
aEvent.mnButton = MOUSE_MIDDLE;
break;
case Qt::RightButton:
aEvent.mnButton = MOUSE_RIGHT;
break;
default:
return;
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
}
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
aEvent.mnTime = pEvent->timestamp();
More loplugin:cstylecast: vcl auto-rewrite with <https://gerrit.libreoffice.org/#/c/47798/> "Enable loplugin:cstylecast for some more cases" plus solenv/clang-format/reformat-formatted-files Change-Id: I363c01a1ae9e863fca4fb4589829492d7280d711
2018-01-12 20:10:40 +01:00
aEvent.mnX = static_cast<long>(pEvent->pos().x());
aEvent.mnY = static_cast<long>(pEvent->pos().y());
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
aEvent.mnCode = GetKeyModCode(pEvent->modifiers()) | GetMouseModCode(pEvent->buttons());
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
SalEvent nEventType;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
if (bReleased)
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
nEventType = SalEvent::MouseButtonUp;
else
nEventType = SalEvent::MouseButtonDown;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
m_pFrame->CallCallback(nEventType, &aEvent);
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::mousePressEvent(QMouseEvent* pEvent) { handleMouseButtonEvent(pEvent, false); }
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::mouseReleaseEvent(QMouseEvent* pEvent) { handleMouseButtonEvent(pEvent, true); }
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::mouseMoveEvent(QMouseEvent* pEvent)
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
{
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
QPoint point = pEvent->pos();
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
SalMouseEvent aEvent;
aEvent.mnTime = pEvent->timestamp();
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
aEvent.mnX = point.x();
aEvent.mnY = point.y();
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
aEvent.mnCode = GetKeyModCode(pEvent->modifiers()) | GetMouseModCode(pEvent->buttons());
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
aEvent.mnButton = 0;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
m_pFrame->CallCallback(SalEvent::MouseMove, &aEvent);
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
pEvent->accept();
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::wheelEvent(QWheelEvent* pEvent)
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
{
SalWheelMouseEvent aEvent;
aEvent.mnTime = pEvent->timestamp();
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
aEvent.mnX = pEvent->pos().x();
aEvent.mnY = pEvent->pos().y();
aEvent.mnCode = GetKeyModCode(pEvent->modifiers()) | GetMouseModCode(pEvent->buttons());
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
int nDelta = pEvent->angleDelta().x();
aEvent.mbHorz = true;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
if (!nDelta)
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
{
nDelta = pEvent->angleDelta().y();
aEvent.mbHorz = false;
}
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
if (!nDelta)
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
return;
nDelta /= 8;
aEvent.mnDelta = nDelta;
aEvent.mnNotchDelta = nDelta > 0 ? 1 : -1;
aEvent.mnScrollLines = 3;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
m_pFrame->CallCallback(SalEvent::WheelMouse, &aEvent);
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
pEvent->accept();
}
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
void Qt5Widget::startDrag()
{
tdf#120775: Implement DnD between 2 SalFrames still only internal DnD (between multiple frames within 1 LibO instance) is supported Change-Id: I151162ee1075c8b2d71520aae950fc7214d752b7 Reviewed-on: https://gerrit.libreoffice.org/64477 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-12-03 16:40:27 +01:00
// internal drag source
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
QMimeData* mimeData = new QMimeData;
tdf#120775: Implement DnD between 2 SalFrames still only internal DnD (between multiple frames within 1 LibO instance) is supported Change-Id: I151162ee1075c8b2d71520aae950fc7214d752b7 Reviewed-on: https://gerrit.libreoffice.org/64477 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-12-03 16:40:27 +01:00
mimeData->setData(m_InternalMimeType, nullptr);
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
QDrag* drag = new QDrag(this);
drag->setMimeData(mimeData);
drag->exec(Qt::CopyAction | Qt::MoveAction, Qt::MoveAction);
}
void Qt5Widget::dragEnterEvent(QDragEnterEvent* event)
{
tdf#120775: Implement DnD between 2 SalFrames still only internal DnD (between multiple frames within 1 LibO instance) is supported Change-Id: I151162ee1075c8b2d71520aae950fc7214d752b7 Reviewed-on: https://gerrit.libreoffice.org/64477 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-12-03 16:40:27 +01:00
if (event->mimeData()->hasFormat(m_InternalMimeType))
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
event->accept();
tdf#120775: Implement DnD between 2 SalFrames still only internal DnD (between multiple frames within 1 LibO instance) is supported Change-Id: I151162ee1075c8b2d71520aae950fc7214d752b7 Reviewed-on: https://gerrit.libreoffice.org/64477 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-12-03 16:40:27 +01:00
// else FIXME: external drag source
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
}
void Qt5Widget::dragMoveEvent(QDragMoveEvent* event)
{
Kick-start beginning of drag event Change-Id: Iaad25a7acdc7d64013bc3dd0d9410e7d2d5c6762 Reviewed-on: https://gerrit.libreoffice.org/61791 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-15 14:44:29 +02:00
QPoint point = event->pos();
m_pFrame->draggingStarted(point.x(), point.y());
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
QWidget::dragMoveEvent(event);
}
void Qt5Widget::dropEvent(QDropEvent* event)
{
Now that initiating drag works, let's also accept drops Change-Id: Iab328edd799dd4ce04312db4e640f86a8f7fda77 Reviewed-on: https://gerrit.libreoffice.org/61897 Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de> Tested-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-18 01:01:14 +02:00
QPoint point = event->pos();
m_pFrame->dropping(point.x(), point.y());
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
QWidget::dropEvent(event);
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::moveEvent(QMoveEvent*) { m_pFrame->CallCallback(SalEvent::Move, nullptr); }
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::showEvent(QShowEvent*)
QT5 queue paint update on show Something like gtk_widget_set_redraw_on_allocate. On resize of dialogs, we still get artefacts, but at least you see something without forcing an event. And since there isn't any Qt5 theming yet, we let VCL paint the menus with a border. Change-Id: Ia2ff6b199ee25e481e3fb29a921b0d3d490f1535
2017-11-07 13:13:10 +01:00
{
Make clang-format happy, remove vcl/unx/kde5 from blacklist just between us girls, I hate you clang-format Change-Id: I4cd041eff7a9fa01eb9192cf357e3f07a26fa6ad
2018-05-31 12:14:55 +02:00
QSize aSize(m_pFrame->GetQWidget()->size());
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
SalPaintEvent aPaintEvt(0, 0, aSize.width(), aSize.height(), true);
m_pFrame->CallCallback(SalEvent::Paint, &aPaintEvt);
QT5 queue paint update on show Something like gtk_widget_set_redraw_on_allocate. On resize of dialogs, we still get artefacts, but at least you see something without forcing an event. And since there isn't any Qt5 theming yet, we let VCL paint the menus with a border. Change-Id: Ia2ff6b199ee25e481e3fb29a921b0d3d490f1535
2017-11-07 13:13:10 +01:00
}
QT5 implement some mouse handling Scrollwheel handling seems to work with mouse, but not correct when using a touchpad - at least for me. Change-Id: I4f1b32205516912e31f9c52605ba2bf4ec6059a8
2017-10-30 19:55:18 +01:00
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::closeEvent(QCloseEvent* /*pEvent*/)
Frames can now be closed by clicking top-right corner [x] Change-Id: I4f841bfd8fe819a322af1ecaa95d17273d3031ec
2018-04-16 13:30:23 +02:00
{
Set Qt5Widget to be a central widget of QMainWindow this is meant to solve the problem of native menu bar overlapping w/ non-native, as well as the inability to place an object or select text dragging the mouse cursor w/ LMB pressed Change-Id: I29f590ebf79d1ecc7e17b402125384cf13774bf3 Reviewed-on: https://gerrit.libreoffice.org/58171 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-09 17:14:14 +02:00
m_pFrame->CallCallback(SalEvent::Close, nullptr);
Frames can now be closed by clicking top-right corner [x] Change-Id: I4f841bfd8fe819a322af1ecaa95d17273d3031ec
2018-04-16 13:30:23 +02:00
}
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
static sal_uInt16 GetKeyCode(int keyval)
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
{
sal_uInt16 nCode = 0;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
if (keyval >= Qt::Key_0 && keyval <= Qt::Key_9)
nCode = KEY_0 + (keyval - Qt::Key_0);
else if (keyval >= Qt::Key_A && keyval <= Qt::Key_Z)
nCode = KEY_A + (keyval - Qt::Key_A);
else if (keyval >= Qt::Key_F1 && keyval <= Qt::Key_F26)
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
nCode = KEY_F1 + (keyval - Qt::Key_F1);
else
{
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
switch (keyval)
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
{
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
case Qt::Key_Down:
nCode = KEY_DOWN;
break;
case Qt::Key_Up:
nCode = KEY_UP;
break;
case Qt::Key_Left:
nCode = KEY_LEFT;
break;
case Qt::Key_Right:
nCode = KEY_RIGHT;
break;
case Qt::Key_Home:
nCode = KEY_HOME;
break;
case Qt::Key_End:
nCode = KEY_END;
break;
case Qt::Key_PageUp:
nCode = KEY_PAGEUP;
break;
case Qt::Key_PageDown:
nCode = KEY_PAGEDOWN;
break;
case Qt::Key_Return:
tdf#120784 KDE5: Process "Enter" key from numpad like normal "Enter" key Change-Id: Id4a721aa30e92a9396f93342cf42d5790207bb89 Reviewed-on: https://gerrit.libreoffice.org/63095 Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de> Tested-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-11-08 15:21:14 +03:00
case Qt::Key_Enter:
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
nCode = KEY_RETURN;
break;
case Qt::Key_Escape:
nCode = KEY_ESCAPE;
break;
case Qt::Key_Tab:
Traversing dialog controls backwards (Shift-Tab) works now Change-Id: If0e8d7307dd37436751e524d7ed24e89d5b72db3
2018-06-26 11:48:52 +02:00
// oddly enough, Qt doesn't send Shift-Tab event as 'Tab key pressed with Shift
// modifier' but as 'Backtab key pressed' (while its modifier bits are still
// set to Shift) -- so let's map both Key_Tab and Key_Backtab to VCL's KEY_TAB
case Qt::Key_Backtab:
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
nCode = KEY_TAB;
break;
case Qt::Key_Backspace:
nCode = KEY_BACKSPACE;
break;
case Qt::Key_Space:
nCode = KEY_SPACE;
break;
case Qt::Key_Insert:
nCode = KEY_INSERT;
break;
case Qt::Key_Delete:
nCode = KEY_DELETE;
break;
case Qt::Key_Plus:
nCode = KEY_ADD;
break;
case Qt::Key_Minus:
nCode = KEY_SUBTRACT;
break;
case Qt::Key_Asterisk:
nCode = KEY_MULTIPLY;
break;
case Qt::Key_Slash:
nCode = KEY_DIVIDE;
break;
case Qt::Key_Period:
nCode = KEY_POINT;
break;
case Qt::Key_Comma:
nCode = KEY_COMMA;
break;
case Qt::Key_Less:
nCode = KEY_LESS;
break;
case Qt::Key_Greater:
nCode = KEY_GREATER;
break;
case Qt::Key_Equal:
nCode = KEY_EQUAL;
break;
case Qt::Key_Find:
nCode = KEY_FIND;
break;
case Qt::Key_Menu:
nCode = KEY_CONTEXTMENU;
break;
case Qt::Key_Help:
nCode = KEY_HELP;
break;
case Qt::Key_Undo:
nCode = KEY_UNDO;
break;
case Qt::Key_Redo:
nCode = KEY_REPEAT;
break;
case Qt::Key_Cancel:
nCode = KEY_F11;
break;
case Qt::Key_AsciiTilde:
nCode = KEY_TILDE;
break;
case Qt::Key_QuoteLeft:
nCode = KEY_QUOTELEFT;
break;
case Qt::Key_BracketLeft:
nCode = KEY_BRACKETLEFT;
break;
case Qt::Key_BracketRight:
nCode = KEY_BRACKETRIGHT;
break;
case Qt::Key_Semicolon:
nCode = KEY_SEMICOLON;
break;
case Qt::Key_Copy:
nCode = KEY_COPY;
break;
case Qt::Key_Cut:
nCode = KEY_CUT;
break;
case Qt::Key_Open:
nCode = KEY_OPEN;
break;
case Qt::Key_Paste:
nCode = KEY_PASTE;
break;
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
}
}
return nCode;
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
bool Qt5Widget::handleKeyEvent(QKeyEvent* pEvent, bool bDown)
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
{
SalKeyEvent aEvent;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
aEvent.mnCharCode = (pEvent->text().isEmpty() ? 0 : pEvent->text().at(0).unicode());
aEvent.mnRepeat = 0;
aEvent.mnCode = GetKeyCode(pEvent->key());
aEvent.mnCode |= GetKeyModCode(pEvent->modifiers());
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
bool bStopProcessingKey;
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
if (bDown)
bStopProcessingKey = m_pFrame->CallCallback(SalEvent::KeyInput, &aEvent);
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
else
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
bStopProcessingKey = m_pFrame->CallCallback(SalEvent::KeyUp, &aEvent);
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
return bStopProcessingKey;
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::keyPressEvent(QKeyEvent* pEvent)
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
{
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
if (handleKeyEvent(pEvent, true))
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
pEvent->accept();
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::keyReleaseEvent(QKeyEvent* pEvent)
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
{
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
if (handleKeyEvent(pEvent, false))
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
pEvent->accept();
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::focusInEvent(QFocusEvent*) { m_pFrame->CallCallback(SalEvent::GetFocus, nullptr); }
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
void Qt5Widget::focusOutEvent(QFocusEvent*)
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
{
Retrofit "KeepEmptyLinesAtTheStartOfBlocks: false" into .clang-format ...even if that can cause reformatting of already formatted code. The problem I came across is that without this something like > namespace { > > void f1(); > > void f2(); > > } (which is quite a common style in the current code base) would be changed to > namespace > { > > void f1(); > > void f2(); > } instead of > namespace > { > void f1(); > > void f2(); > } and I found no other clang-format style option that would result in the presence or absence of an empty line be identical at the start and end of the namespace block. vmiklos asked to reformat the existing new (i.e., non-blacklisted) files at the same time, so this commit includes that. Some of those new files had not been formatted at all, so this commit includes their full reformatting changes. Change-Id: I54daf0c11098d07d02c802104cf7f56372e61f7c Reviewed-on: https://gerrit.libreoffice.org/44450 Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de> Tested-by: Jenkins <ci@libreoffice.org> Reviewed-by: Stephan Bergmann <sbergman@redhat.com>
2017-11-08 11:28:04 +01:00
m_pFrame->CallCallback(SalEvent::LoseFocus, nullptr);
QT5 initial keyboard support No idea, if we can / should map those vendor specific keys, like Gtk+ does. This also prefixes the non-virtual functions with "handle". Change-Id: Id8c42651e07d33728ff6deced06a82de29aa3fad
2017-11-07 13:18:05 +01:00
}
tdf#120454: Implement native rendering of tooltips side-step the entire misery of proper positioning of non-native tooltips just like in gtk3 Change-Id: I09a75fae672d3d999c946c50c547d5f2cfa3ec14 Reviewed-on: https://gerrit.libreoffice.org/64956 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-12-11 12:12:52 +01:00
void Qt5Widget::showTooltip(const OUString& rTooltip)
{
QPoint pt = QCursor::pos();
QToolTip::showText(pt, toQString(rTooltip));
}
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
Qt5Widget::Qt5Widget(Qt5Frame& rFrame, Qt::WindowFlags f)
: QWidget(Q_NULLPTR, f)
, m_pFrame(&rFrame)
Templatize Qt5Widget class to distinguish between QMainWindow (top-level SalFrame) and QWidget (everything else). The former are going to receive a native menu Change-Id: I9023d1e464345f96f13967f5e2c8b0716890fbc9
2018-05-14 15:39:30 +02:00
{
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
create();
setMouseTracking(true);
Implement drag'n'drop-related qt5 events, no-op so far Change-Id: I67b8cecdc85ad77d7ab48d229459a79bc168072d Reviewed-on: https://gerrit.libreoffice.org/61721 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-10-12 16:19:07 +02:00
setAcceptDrops(true);
Revert templatization of Qt5Widget class it was not a step in the right direction in 1st place. We distinguish between main window frame (QMainWindow) and everything else (QWidget) differently now Change-Id: I331c9760f7616b2f199e63e5e521df74b08acc62 Reviewed-on: https://gerrit.libreoffice.org/58853 Tested-by: Jenkins Reviewed-by: Thorsten Behrens <Thorsten.Behrens@CIB.de>
2018-08-10 18:29:10 +02:00
setFocusPolicy(Qt::StrongFocus);
No parent-child relationship between Qt5Widgets QWidget destructor always deletes all its children, which is fatal in case not-yet-deleted frame (e.g. because it's in lazy delete queue) points to one of those children There's parent-child relationships between the frames though, so perhaps no need to have it double. For native modality this seems enough Change-Id: I556ace8d7949bcdfb208170cc08181ac1f2622e8 Reviewed-on: https://gerrit.libreoffice.org/57560 Tested-by: Jenkins Reviewed-by: Katarina Behrens <Katarina.Behrens@cib.de>
2018-07-17 15:57:41 +02:00
}
KF5 add QWidget to Kf5Frame Change-Id: I53845519d0dda324c9544f057b18c6afd4cf858a
2017-10-27 13:15:37 +02:00
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */
Reference in New Issue Copy Permalink