The test only tests what is currently supported so it does
not yet test radio buttons and action chains. But
the visibilitytest.pdf already provides these, too.
Differential Revision: https://phabricator.kde.org/D11597
Summary:
This adds the hidden property to JavaScript fields and
uses it to implement support for HideAction.
Test Plan: Unit test in the next commit.
Reviewers: #okular
Subscribers: aacid
Tags: #okular
Maniphest Tasks: T8274
Differential Revision: https://phabricator.kde.org/D11596
Summary:
Okular saves a new file with permissions 0600 completely ignoring the umask value. This is because it
makes use of QTemporaryFile which creates file with the said permissions and which then Okular copies
to the new location.
So to overcome this generate new file permissions using old mask value and change permissions of the
new file.
BUG: 392682
Test Plan:
Set umask to 0040
Open a pdf file from the same shell and save it under a new name.
Before patch:
file permisions -> 0600
After patch:
file permisions -> 0606
Reviewers: #okular, aacid
Tags: #okular
Differential Revision: https://phabricator.kde.org/D12049
Summary:
Some distros seem to have the option.
BUG: 393119
Test Plan: Still builds, can disable build
Reviewers: #okular, mart, aacid, lbeltrame
Tags: #okular
Differential Revision: https://phabricator.kde.org/D12352
Summary:
Mark the annotations that contain a non-empty popup. Especially useful to identify annotations that are not meant to contain an annotation as shapes or highlights.
BUG: 389836
Reviewers: #okular, aacid
Reviewed By: #okular, aacid
Subscribers: rkflx, aacid, ngraham
Tags: #okular
Differential Revision: https://phabricator.kde.org/D10797
Summary:
Include an AndroidManifest.xml file so when it's installed it has a
proper icon and name.
Include appdata information so it can be listed
Reviewers: #okular, mart, aacid
Reviewed By: #okular, mart, aacid
Tags: #okular
Differential Revision: https://phabricator.kde.org/D12299
Summary:
The highlights were removed when rotating the page, also the RotaionJob::rotationMatrix function was not rotating the Highlihgts correctly.
I removed the deletion and modified the rotaionMatrix function by making it shorter and adding a translation after the rotatin. I tried different rotations and they are working fine.
BUG: 387282
Reviewers: #okular
Subscribers: aacid, ngraham, #okular
Tags: #okular
Differential Revision: https://phabricator.kde.org/D11901
Summary:
Makes for straightforward Android testing and will work on normal
installations as well.
Reviewers: #kirigami, #okular, mart
Reviewed By: #kirigami, mart
Subscribers: plasma-devel
Tags: #kirigami, #okular
Differential Revision: https://phabricator.kde.org/D12176
Does not make sese to ask the user if he wants to go fullscreen on the
kile preview widget (or the milou preview widget, the other user of this
feature), and besides not making sense, it's crashing :D
BUGS: 390383
Summary:
in drawPixmap(source, sourceRect) the source rect is in native pixels.
In this code the size was scaled by the DPR, but the location was not.
blendRect is the contents of a selection rectangle, offset from the
bounding darker blue rectangle. Hence we haven't really noticed.
This completely fixes the rectangle when the scale == 2.0/3.0
Some smaller artifacts remain on fractional scale factors, which is a
separate code bug with the same symptoms.
CCBUG: 386111
Reviewers: aacid, rkflx
Reviewed By: rkflx
Subscribers: sander, rkflx, anthonyfieroni, #okular
Tags: #okular
Differential Revision: https://phabricator.kde.org/D8980
Some more tweaks to path resolving.
Don't prepend a / in entryPath if path already starts with /
Fonts can also have relative paths, so use the absolutePath before
calling getFontByName
BUGS: 392826
Summary:
Seems like someone forgot to consistently apply a change. Maybe the path
construction should be factored into a reusable lambda?
Test Plan:
Open Okular, open settings > configure Okular. Without the patch, you'll get a
"QFSFileEngine::open: No file name specified" warning, with the patch it
disappears. Note that there is also an unrelated "Illegal icon group: 7"
warning from KIconThemes, but I'm 99% certain that this is on KIconThemes doing
a bogus check in KIconEffect::apply, not on Okular
Reviewers: #okular, hetzenecker, aacid
Reviewed By: #okular, aacid
Subscribers: aacid, #okular
Tags: #okular
Differential Revision: https://phabricator.kde.org/D12001
Summary:
Seems like someone forgot to consistently apply a change. Maybe the path
construction should be factored into a reusable lambda?
Test Plan:
Open Okular, open settings > configure Okular. Without the patch, you'll get a
"QFSFileEngine::open: No file name specified" warning, with the patch it
disappears. Note that there is also an unrelated "Illegal icon group: 7"
warning from KIconThemes, but I'm 99% certain that this is on KIconThemes doing
a bogus check in KIconEffect::apply, not on Okular
Reviewers: #okular, hetzenecker, aacid
Reviewed By: #okular, aacid
Subscribers: aacid, #okular
Tags: #okular
Differential Revision: https://phabricator.kde.org/D12001
There's no need to continue generating something that won't be useful.
Also fixes a crash if the request would report partial updates since at that stage it uses the observer and that is gone
This fix smells a lot like a workaround for a bug in Qt but i could not
write down a sufficiently small testcase for it to report it to Qt, and
this is not wrong anyway since we eventually add playerPage to a
layout for this. What I tracked down was that since playerPage had no
parent, at some point a nativeId for it was being created and that
broke sizing back to the default 100x30 used in Qt instead of the
size we had given it.
For some reason i needed a PDF file with more than one video to
reproduce this problem, the first video was correctly sized but the
second one was 100x30.
The crash would happen because we were setting partial updates to true on requests
that were synchronous, because that sync request was cancelling an async one (prerender)
so now we only set newRequest partial updates to true when cancellingm another one if it's asynchronous
And a sync but with partial updates request would make stuff crash because the request
would all be processed on the main thread but generator_pdf.cpp cakks signalPartialPixmapRequest
with a queued connection so we that function was called the request would have already gone
