Okular was checking for Q_WS_X11 definition (which is not available since Qt 5) to detect CUPS, so detection always fails making CUPS parameters ignored.
REVIEW: 127036
Drops the x11 fallback. We were falling back to QX11Extras that were in
fact using QScreen internally.
Also remove the QDesktopWidget fallbacks for the same reason.
Drops some API for fetching the dpi and unifies it with ::realDpi(QWidget).
Based on Sebas's patch.
Reviewed by Albert.
REVIEW: 126913
Try fixing mainshelltest: one more passes now
QUrl is not automatically exposed to DBus so the DBus call would fail.
Changed the parameter to QString instead so that DBus invocation works.
mainshelltest: set QStandardPaths to test mode
Parse command line flags with a single dash as a long option
Unlike KCmdLineArgs QCommandLineParser treats options starting with a
single minus as multiple short options by default.
Previously okular -unique would fail with the following error:
Unknown options: u, n, i, q, u, e.
Also changed mainshelltest to use two dashes in case this behaviour
should change in the future.
mainshelltest failures have been reduced from 15 to 4 by this commit.
Fix docdata saving and added a warning message if it fails
If the ~/.local/okular/docdata directory didn't exist previously
creating the docdata file would fail as there are missing paths.
It seems that KStandardDirs used to create the okular/docdata directory
automatically, with QStandardPaths we have to create it manually.
mainshelltest is down to one failed test now
Fix final test case in mainshelltest
as we call QProcess::terminate the exit code will not be 0
REVIEW: 126192
That patch extracts the video file, which is defined in a
rich media annotation as parameter for the flash player,
and uses the normal multimedia player, to playback the video
file.
This feature requires poppler-qt5 in version 0.36.
FEATURE: 326230
REVIEW: 124612
Process the operations of the rendition link independently of whether
there is a JS script or not. Most of the time the JS script won't be
executed correctly anyway, since Okular is missing most of the JS API,
so we should fall back on the given operations.
BUGS: 324167
FIXED-IN: KDE Applications 15.08.0
Allows us to not use a dummy printer for knowing page size
Which is nice since it won't block because Qt4 QPrinter is a bit weird sometimes
BUGS: 334708
FIXED-IN: KDE Applications 15.08
Allows us to not use a dummy printer for knowing page size
Which is nice since it won't block because Qt4 QPrinter is a bit weird sometimes
BUGS: 334708
FIXED-IN: KDE Applications 15.08
This commit fixes the error message "<url> is not a local file" when processing
source references. 'giveAbsolutePath(QString)' returns a local file name, so we
create the corresponding url using 'QUrl::fromLocalFile'.
The old auto_ptr was used for simple RAII. None of the auto_ptr flaws
were apparent in that context, but given that that class is actively
deprecated and that compilers warn about its use, it's better to switch
to its drop-in replacement.
My reason for not using QScopedPointer is that the C++11 version works
well enough, Okular is being built in C++11 mode anyway, unique_ptr
works on the KF5-minimal-mandated-compilers according to their docs, and
therefore there's no point in using something with a different API
(QScopedPointer) just because its name begins with a Q.
REVIEW: 124027
Wow, I have no idea how this happened. I wrote an big test case and spent a lot of time playing with tabs. Stephan Binner even told me about this a month ago, but I didn't quite understand the problem...
Broken embed mode detection means that the tabbed interface is disabled. It actually made it into release... ugh
REVIEW: 123680
Replace both the package version file by a generated file, as well as
OkularConfig.cmake package configuration file.
Note that with this change the compatibility CMake variable OKULAR_FOUND
is not set anymore, use Okular_FOUND instead.
REVIEW: 123314
I've no idea why i need this for it to compile now, but kind of makes sense to follow the
same pattern for all the phonon includes, so unify with the rest of includes in this file
Somehow i ended up commiting a partial patch and not what was in reviewboard, this amends the previous commit to be what i really wanted to commit
Sorry for the mess :/
This way dvi and any other potential user gets it for free
The diff is huge, but the synctex files are just moves.
And the code in core/ is also mostly just a move from the generator_pdf.cpp code
Acked by Luigi
REVIEW: 120311
