GitLab

We're assigning it in the initializers list

The method qApp->devicePixelRatio() should only be used under
special circumstances.  The documentation of QGuiApplication says:

  "Returns the highest screen device pixel ratio found on the system.
   [...] Use this function only when you don't know which window
   you are targeting. If you do know the target window,
   use QWindow::devicePixelRatio() instead."

The Okular code used qApp->devicePixelRatio() a few times without
reason.  Few people noticed this, because apparently only Windows
and Wayland can set per-screen dprs.  This patch replaces those
calls with the correct QWindow call.

Oliver Sander authored Feb 19, 2021 and Albert Astals Cid committed Mar 08, 2021

The PixmapRequest constructor expected width and height in logical
pixels, and scaled them to device pixels.  However, not knowing what
screen the request is for, it had to use qApp->devicePixelRatio()
for the scaling.  That value may not be correct if not all screens
use the same scaling.

Fix this by introducing a new constructor that takes the device
pixel ratio as an additional argument, and deprecating the old
constructor.

This fixes the issue that it's not obvious how to stop the next page
timer if a pdf has defined a given duration.

BUGS: 432416

Source files are no longer separated by UI and non-UI and similar,
but only by their build target.

* ui/ -> part/
* Move all source files from conf/ to part/
* Keep config skeleton definitions in conf/, needed for the mobile target too
* Move editdrawingtooldialogtest.h from conf/autotests/ to autotests/
* ui/data/icons/ -> icons/
* Move /part.cpp, /part.rc and similar files to part/
* Adapt include paths in source files
* Adapt CMakeLists.txt files (in / and in subdirectories)
* Adapt /Messages.sh

David Hurka authored Aug 09, 2020 and Albert Astals Cid committed Aug 09, 2020

* QDesktopWidget -> QScreen

Simplifications:
* After setting new screen/geometry, forget them and let resizeEvent() handle any updates.
* Specifically, don’t connect to screen size change signals.
* Specifically, setScreen() only sets the new widget position and then waits for resizeEvent().
* This also means applyNewScreenSize() and repositionContent() can be collapsed into resizeEvent().
* Replace recalcGeometry() by setScreen(defaultScreen()), where defaultScreen() only checks the configuration.

Screen switching from inside presentation mode currently does not work correctly.
The reason is that the widget geometry() shall be moved
*while* the widget is configured as fullscreen.
This patch temporarily disables fullscreen, so the screen can be switched.

You can check it with release 20.04:
Screen switching from within presentation mode will not work.
Set the widget to windowed (e. g. Alt+F3 or right-click in task manager),
screen switching will work now.

!233 shall port PresentationWidget to QScreen in release 20.12,
so this is just a compact workaround without much intelligence.
The drawback is that custom window rules might get confused when the
screen geometry changes for some reasons.

This adds a KToggleAction which sets annotation tools to constrain angle mode.
It provides an alternative user interface to the Shift button, which is used to constrain angles since MR !210.
The action and the Shift button are XOR-ed, i. e. if constrain angle mode is activated, pressing Shift temporarily disables it.

The action state is remembered accross sessions, for consistency with most other actions. It should be difficult to check this action without knowing of its existence, since it is not in any toolbar or menu, just in the action collection.

The meaning of AnnotatorEngine::Modifiers was generalized a bit, moving the responsibility about whether to constrain angles back to PageViewAnnotator, because AnnotatorEngine does not know about the action.

FEATURE: 353560
FIXED-IN: 1.11

Luca Citi authored Jul 10, 2020 and Albert Astals Cid committed Jul 13, 2020

When drawing a line (or polygon) annotation, holding shift constrains the new segment to form an angle with the horizontal axis which is an integer multiple of 15°.
This is helpful when drawing, e.g., horizontal, vertical or 45 degree lines.

find . \( -name "*.cpp" -or -name "*.h"  -or -name "*.c"  -or -name "*.cc" \) -exec clang-format -i {} \;

If you reached this file doing a git blame, please see README.clang-format (added 2 commits in the future of this one)

Ahmad Samir authored Dec 09, 2019 and Albert Astals Cid committed Dec 09, 2019

The code compiles, tests pass (except for two tests but they fail
on master too). Okular loads and seems to work.

Subscribers: okular-devel

Tags: #okular

Differential Revision: https://phabricator.kde.org/D25553

Oliver Sander authored Jul 04, 2019 and Albert Astals Cid committed Aug 01, 2019

The previous code used two consecutive calls to the method
slotChangeDrawingToolEngine to reset the m_drawingEngine
member (i.e., to delete it and to create a new one from scratch).
That worked; however, as a side effect the calls to
slotChangeDrawingToolEngine also set the cursor to an arrow,
which happens every time one continuous stroke is finished.
This went unnoticed because when drawing in presentation mode
the cursor is an arrow anyway.  However, in anticipation
of future patches that change the cursor to a cross-hair
when drawing, this explicit setting of the arrow shape
needs to go.  Therefore this patch replaces the two calls
to slotChangeDrawingToolEngine by the code they contain,
minus the code to set the cursor shape. The new code is
only one line longer, and more readable too.

QRegion::rects is deprecated

QFontMetrincs::horizontalAdvance is only available in Qt 5.11

This reverts commit 69ec1755.

width -> horizontalAdvance

background -> window

Previously, when using a screen scaling factor larger than 1, the
drawings by mouse or stylus in presentation mode were slightly blocky.
The underlying cause was
a) a few integer types used for non-integer coordinates
b) the intermediate use of a QPixmap without the correct
   devicePixelRatio

Additionally, this patch uses a QPainterPath to draw the path,
rather than a sequence of individual lines.  This improves
the line quality a tiny bit, because the lines are properly
joined.

BUG: 384143

Differential Revision: https://phabricator.kde.org/D21836

Summary: fix clazy warning

Reviewers: aacid

Reviewed By: aacid

Subscribers: nicolasfella, okular-devel

Tags: #okular

Differential Revision: https://phabricator.kde.org/D20832

Oliver Sander authored Mar 26, 2019 and Albert Astals Cid committed Mar 26, 2019

Summary: As the title says.  In my humble opinion this makes the code *much* easier to read.

Test Plan: No functional changes.

Reviewers: #okular

Subscribers: aacid, davidhurka, okular-devel

Tags: #okular

Differential Revision: https://phabricator.kde.org/D19539

Oliver Sander authored Mar 26, 2019 and Albert Astals Cid committed Mar 26, 2019

Summary: Previously this was inconsistent: left button clicks triggered page changes at release, but right button and forward/backward button clicks triggered directly at the mousePressEvent.

Reviewers: #okular, ngraham, aacid

Reviewed By: ngraham

Subscribers: aacid, okular-devel

Tags: #okular

Differential Revision: https://phabricator.kde.org/D19702

This patch makes switching presentation slides using touch screen taps
more flexible.  It introduces a new configuration option
'Touch navigation' with three possible values:
* 'Tap left/right side to go back/forward':
   Tapping on the left(right) half of the screen
  makes the presentation go to the previous(next) slide.
* 'Tap anywhere to go forward':
  Tapping on the screen makes the presentation go to the
  next slide, no matter where the screen is being tapped
* 'Disabled':
  Screen tapping doesn't do anything.

Previously, only 'Tap anywhere...' was implemented.

This patch does not change the behavior of mouse clicks.

Differential Revision: https://phabricator.kde.org/D18118

they are unused(except the PageViewItem one), but if anyone would use
them things would go wrong, so protect us from it

Actually fixes a bug in PageView::slotFitWindowToPage in which we were
copying constructing PageViewItem and that's bad

Simone Gaiarin authored Dec 21, 2018 and Oliver Sander committed Dec 21, 2018

It is more logical to have the drawing tools first, then the eraser,
and then the action to delete everything. Also the two actions to erase
things are close to each other now.

The icon of "erase drawing" has been changed to draw-eraser-delete-objects
to distinguish it from the "Eraser" tool.

BUG: 399734

Differential Revision: https://phabricator.kde.org/D16171

Summary:
We request a inhibit lock of DBus type UNIX_FD from systemd logind. It's wrapped into (and owned by) a QDBusUnixFileDescriptor object of automatic storage. The file descriptor will be closed in QDBusUnixFileDescriptor Dtor, and may be reused by some other facility (e.g. pulseaudio).

If we want to store the lock longer than QDBusUnixFileDescriptor lifetime, we have to dup the file descriptor. If we don't dup, and close the original fd later in PresentationWidget::allowPowerManagement, bad things may happen.

Besides that, what we get from systemd is really a file descriptor, not a "cookie". So I renamed the m_sleepInhibitCookie to m_sleepInhibitFd and changed initial state to -1 accordingly.

BUG 393478
BUG 398720

Test Plan:
- bugs don't occur any longer
- inhibiting sleep during presentation mode still works

Reviewers: aacid

Reviewed By: aacid

Subscribers: ngraham, anthonyfieroni, okular-devel

Tags: #okular

Differential Revision: https://phabricator.kde.org/D15574