GitLab

Moved executeScript to document private and support to display options for JavaScript

Added support for the setIcon/getIcon defined in Poppler > 0_78

Implemented more unit tests to cover the OCGs

When refreshing, widgets must check if they should be visible

Implemented unit tests to the JavaScript functions.

Seems to make tests more stable

The code that draws a thin gradient at the right and bottom
boundaries of a page implicitly assumed that the display was
not scaled, i.e., that logical pixel where identical to
physical pixels.  Nobody really noticed because that gradient
was so thin.  Anyway, this patch fixes the gradient for
scaled displays.

I considered using a QLinearGradient, but decided against it:
The code would have been longer, and the manual code in the
patch can exploit the fact that all lines are horizontal
or vertical.  Therefore, expensive scan-conversion can be
avoided.

Pages get a black rectangle around them as 'outline' after
being rendered.  This rendering was not pixel-perfect at
fractional scalings: apparently, fractional line widths and
positions get rounded differently.  The result was that
sometimes there was a small gap between outline and page.

This pages fixes the problem by making the pen for the
outline 'cosmetic', i.e., one physical pixel wide.
With that, its position can be such that it encloses
the page without any gap.

Okular renders a simple 'shadow' at the right and the bottom of each
page.  The code for this is a bit fragile:  After the page is rendered,
Okular paints a black outline, and then the shadow.  The shadow is
a hand-implemented gradient, painted line by line.  Finally, the
remaining area is painted in the background color.  No pixel is
ever touched twice.  Unfortunately, the code is buggy, and in hidpi /
fractional scaling situations, some pixels are never touched.
This results in rendering glitches as reported in

  https://bugs.kde.org/show_bug.cgi?id=383943

Instead of trying to fix the previous approach, this patch makes the
code more robust by changing the order of the painting operations.
After painting the page, the code now first paints the background,
which is now enlarged to cover everything off the page.  Finally,
the black outline and shadow are drawn on the previously drawn
background.  This makes sures that no pixel is ever left uninitialized.
It also allows to simplify the shadow drawing operation a bit.

BUG: 383943

According to

  https://doc.qt.io/qt-5/qregion-obsolete.html#rects

the method QRegion::rects is obsolete and should be avoided.
Let's do that by moving to a range-based for over the QRegion.
That is easier to read anyway.

Summary:
This patch adds an "Open Containing Folder" menu item to the file menu, similar to
other KDE apps like Gwenview that have it there. This action is especially useful
for the case when you've downloaded a PDF from the internet that opens itself in
Okular. The location of this file may not be clear or easy to find without this
feature, and on several occasiona I have found myself wishing for it when this
happens.

Test Plan:
- Delete or move aside `~/.local/config/kxmlgui5/okular/part.rc`
- Menu item is enabled and works when there is an open document: {F6940449, size=full}
- Menu item is disabled when there is no open document: {F6940448, size=full}

Reviewers: #okular, aacid, sander

Reviewed By: #okular, aacid

Subscribers: shubham, davidhurka, okular-devel

Tags: #okular

Differential Revision: https://phabricator.kde.org/D22145

It's what KF5 uses since a while and we kind of agreed it was old-enough
looking at https://repology.org/metapackage/cmake/badges

It matches quite well our min Qt and KF5 requirements

It's LTS and reading https://repology.org/project/qt5/badges it's not
like many distros ship Qt 5.8 anyway

QFontMetrincs::horizontalAdvance is only available in Qt 5.11

This reverts commit 69ec1755.

We require Qt 5.8 and there (and up) it returns true unconditionally and
it's now marked as deprecated

width -> horizontalAdvance

background -> window

When not using frameworks integration we only get "false" in
shouldBeShownYesNo when the saved value is yes/no so set it to yes.

Also it was a bit weird before because we were using an uninitialized
value, so this is much better :)

BUGS: 409430

Summary:
The config dialog of each annotation tool is now a form layout without group boxes. Everything is aligned.

Reviewers: #okular, #vdg, ngraham, aacid

Reviewed By: #okular, #vdg, ngraham, aacid

Subscribers: sander, davidhurka, aacid, okular-devel, knambiar, ngraham

Tags: #okular

Maniphest Tasks: T8076

Differential Revision: https://phabricator.kde.org/D10859

otherwise it doesn't get shown on Plasma Mobile as well, most
distribution doesn't enable mobile app or splits it in seperate package
so this should not be problem

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:
On high res displays we end up using
the tile manager on single pages, that doesn't work
(for reasons that aren't yet clear) - but it
doesn't make sense if we most of the page anyway.

BUG: 401339

Reviewers: #okular, aacid

Subscribers: okular-devel

Tags: #okular

Differential Revision: https://phabricator.kde.org/D20882