| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
Commit 83e790dae14725347180bb79e463c5f242616b8d removed the test, but
not the configure.json files, so the test always failed because the test
failed to compile.
Change-Id: I8b4ab98c34b94afd94b6fffd15516f06ec21c791
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
Latin-1 is a standard fallback for HTTP headers, but no one that
HTTPContentDisposition::Parse considers by default.
Task-number: QTBUG-70288
Change-Id: Ife36fffe31da5492cde49a4d447cfc42b3f64b95
Reviewed-by: Zakor Tamas <ztamas@inf.u-szeged.hu>
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
|
|
|
|
|
| |
Task-number: QTBUG-70247
Change-Id: Ic0898966d9a88a0bbcdf749c6ea7e80e054d68a3
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
| |
WebEngineView supports getting the http status code (error code) for
failed loading requests only. This patch lets the user access the
status code for successfully loaded pages as well.
Change-Id: Ib8dbdfe94eed4d62e731c736c13f60ebd62a23fa
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
| |
Fixes missing dom_key issue when ControlModifier is used on Linux and macOS.
Task-number: QTBUG-69442
Change-Id: Icd644d019494fd3c95269b42c71efd747507a3ea
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
Usually this field should be computed from the native scan code, but sometimes
we do not have a native scan code (in tests, for example). Add a fallback case
to compute this field from windows_key_code, with test.
Task-number: QTBUG-69442
Change-Id: I046094032576f4215cc8b3a6ced699a86ea148a9
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
Currently this field is always computed from QKeyEvent::nativeScanCode, but this
is not correct on macOS where nativeVirtualKey should be used instead (there is
no nativeScanCode on macOS). Fix by using native_key_code which already takes
this into account.
Task-number: QTBUG-69442
Change-Id: I7c99d7d147ab68bac7ab91101de173eb04fa5c03
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
|
| |
Currently this field is filled from QKeyEvent::nativeVirtualKey, but this is
correct only on macOS (cocoa). Other platforms should use nativeScanCode.
Task-number: QTBUG-69442
Change-Id: I097ec4c726a9d249c0adfbc620d2585983f6246c
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
|
| |
Make sure that the platform specific code paths are only taken when the relevant
QPA plugin is actually used. Also detect if eglfs plugin is using Xkb key codes.
Task-number: QTBUG-69442
Change-Id: I9546b61ad8d9f15020a96f8c0006d056aa157fcc
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
|
| |
Move swapping logic to separate helper functions and use them also for filling
out windows_key_code and dom_key.
Task-number: QTBUG-69442
Change-Id: Id3b883b28b877d54315c5e2fb0d77bad250f2462
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
| |
Change-Id: I327c2459e803602dae107defec8d03d75f363cd7
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|
|
|
|
|
|
|
| |
QTimer::singleShot wants to copy the lambda but LoadURLParams is not copyable.
Change-Id: I300b70eef0041294080a15166a50321bb4292cfd
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
| |
Change-Id: Icf0508baf0884afe7591b35e13733b17c9fdfe90
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
| |
Change-Id: I5ab52de1eb6bd7a25e8b680f4779a505e3f1fcf0
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
.qmake.conf
configure.json
examples/webenginewidgets/markdowneditor/resources/3rdparty/marked.js
examples/webenginewidgets/markdowneditor/resources/3rdparty/qt_attribution.json
examples/webenginewidgets/markdowneditor/resources/markdowneditor.qrc
mkspecs/features/platform.prf
src/3rdparty
src/core/media_capture_devices_dispatcher.cpp
src/core/net/url_request_context_getter_qt.cpp
src/core/net/url_request_context_getter_qt.h
src/core/web_contents_adapter.cpp
Change-Id: I467133ba455b1f85f6bb61793794c31cb1094541
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
--enable-webgl-software-rendering is supposed to be used together
with mesa llvmpipe / opengl32sw.dll. However, Chromium did still use
the built-in GPU blacklist to disable features based on the hardware
GPU driver.
Avoid this by always passing --ignore-gpu-blacklist for this mode.
Task-number: QTBUG-69236
Change-Id: I430d101f6eac64478585de54f705e76859c80597
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Not needed and triggers race condition.
Task-number: QTBUG-69007
Change-Id: Id57ba527387e5dbe44a8dd6c5a49e7278403ce64
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
IODevice::close and therefore aboutToClose is called by io thread,
connecting it to deleteLater will register deferred delete
on event loop of IODevice instance, which is on ui thread.
This could be racy since deletion on ui thread can happen before job
is done.
Change-Id: I895e6a71649ba65944d069f254d119cc60aada6c
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When an HTML select box was clicked inside of a QWebEngineView and
the parent QWebEngineView window was moved using the mouse
(via window decoration toolbar for example) the popup window
would stay around instead of being closed.
This happened because of the usage of the Qt:Tool window flag for the
popup window, which implies a tool that floats near its parent window.
The fix is threefold:
1) Use Qt::Popup instead, similarly to how QMenu does it.
Whenever the parent window is moved, the popup will now get a
CloseEvent.
2) Handle the CloseEvent by telling Chromium to close and destroy
the popup.
3) On Windows the OS might send mouse move events to the popup RWHVQD
instance after its parent QWebEngineView, RWHVQD,
QWebEnginePagePrivate (client adapter) is destroyed. We need to
guard the mouse forwarding code not to access the client adapter
if it has already been destroyed.
The second point is done by telling Chromium that its popup lost focus
which it interprets as a sign to hide it, and automatically destroy
it. This will destroy the underlying RWHVQtDelegateWidget and
RWHVQt instances.
Task-number: QTBUG-59891
Change-Id: I47f94a93c495a6caa5de92a6022eaca154994eda
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-69605
Change-Id: I863db484ff2bcf558585f75c73963097fb43148a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The empty URL is used both for representing a missing origin (browser-initiated
navigation request) and a unique/opaque origin. This is problematic since the
security implications are very different in these two cases: browser-initiated
requests usually should have high security clearance, while requests from unique
origins should be restricted.
Task-number: QTBUG-69372
Change-Id: Iff73fd1c9a29f1c5c281a8945536333081ff2d6b
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
- Use OpenURL on the devtools WebContents.
- Guard against self-destruction (otherwise quicknanobrowser crashes).
- Change quicknanobrowser to open links in new tabs.
Task-number: QTBUG-69359
Change-Id: I4db379731c6fa855124d38c5066b0aad622861d2
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
DragTargetDragEnter expects the 'int key_modifiers' parameter to have values
from the enum blink::WebInputEvent::Modifiers and not ui::EventFlags.
Also, contrary to the name, mouse modifiers are also expected.
Task-number: QTBUG-69231
Change-Id: I2369609775243fded563dde7675c4bc2dfc81021
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-69231
Change-Id: I35b503dae7e2d90b26b6e61a4c7c260e45df2b62
Reviewed-by: Filipe Azevedo <filipe.azevedo@kdab.com>
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Because windows.pri is included from a replace function gnArgs,
fatal does not work:
src/core/config/windows.pri:60: 'fatal' is not a recognized test function.
Change-Id: Ib831e1ffa7e8e345477692d453bbcc568864b98c
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
This should enable excluding such snippets from
documentation in some cases.
Change-Id: I46854412546e3774889e09831254828d18362f29
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I7e03d6e3347ee9b81a7414574098827b5773b7b4
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Directly accessing IOProfileData's m_requestInterceptor
from NetworkDelegateQt is not thread safe, unfriend
the class.
This change does not fix the issue of thread safety of
accessing the requestinterceptor from NetworkDelegateQt,
it is just a cleanup for following commit.
Task-number: QTBUG-69844
Change-Id: I488f8968b38ad200178cea2cb0304e1c93aec711
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
- Make private methods private.
- Use DCHECK_CURRENTLY_ON instead of just DCHECK (better error message).
- Use auto for iterators.
- Remove duplicate queue.empty() checks.
- Remove unused arguments for getDevicesForDesktopCapture.
Change-Id: I929f5d5882a2c1924159f4e001fe477e69a5f1a1
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Never used for anything?
Change-Id: Ib7facf264d3a286a8c2ea1d1584ab86f11f4ea33
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
It's a std::unique_ptr<content::MediaStreamUI> that is always null.
Change-Id: Id43be71b20dc3331692591af9fecbc73a8e521d6
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If webengine-printing-and-pdf is disabled we do not compile
print_view_manager_qt, therefore print_view_manager_qt.h
should never be included.
Add missing QT_CONFIG check in web_contents_adapter instead.
Change-Id: Ib0477c1f12025cd60ee9799ac06a6b664c65018c
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Add webengine-webchannel feature.
Change-Id: I600572180f8169aafe79cf0408527cc087d9a007
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Adds printRequested() signal for web content. This signal
is emitted on javascript window.print().
This change updates also qt printview manager implementation,
fixes a corner case and warnings about ipc unconsumed attachments.
Task-number: QTBUG-69237
Task-number: QTBUG-53745
Change-Id: I0c47b732e27e929ac6db237fb562b7d5f9b959c2
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-69550
Change-Id: I38fb8a32ad1967fc8a8568ad6c02ca3ca3000354
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This reverts commit a5bdc6236ad52586579a9a3da6a059835ed60761
and 04355f4d6f311adebfc354b094a5f79c6327e86f.
CI still crashes on tst_QWebEngineProfile::urlSchemeHandler tests,
issue is not reproducible locally and it will take time to investigate.
Task-number: QTBUG-69844
Change-Id: I7ddb6a32383400d553fe692c24effb6dfb7b6102
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Makes it possible to read extra headers added to the request.
Change-Id: I4af0ba8a5b28488b8f12a2cff29807575c24f026
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If default profile is not needed do not create
by accident one in ~QWebEngineProfilePrivate.
Task-number: QTBUG-66068
Change-Id: Ib4131268686178560f3a89dce41e59eb57613e17
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: Ic9b10175b48f993deda77a77bbfcf22f41e31598
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Public API for the new url/url_util_qt extension to Chromium, which allows to
integrate custom schemes into Chromium's url parsing library and security model.
Previously custom schemes would be treated as 'unknown' schemes and rely on
fallback behavior in Chromium.
[ChangeLog][Custom Schemes] Added the QWebEngineUrlScheme class for configuring
how custom schemes are parsed and which security restrictions should apply.
Task-number: QTBUG-62536
Change-Id: I7d8b9da3ad742f568b82ccc6a2456ad35e84069b
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Moves last two remaining classes to ozone.
Now all the context and surface handling is in ozone subdir.
Change-Id: I87d4570f866b905e8f79a73d70f04bd8b21f6e1b
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Switch to use ozone also for x11/glx. This way we have a
more elegant approach to hook in our gl surface classes
and we use ozone for all linux qpa backends. Moreover,
we have now also more fine grain control to reimplement
gl surface handling logic.
This also aims to fix currently broken wayland backend,
but instead of patching the broken x11 calls, we simply reuse
ozone which does not use x11 for ozone egl.
Add webengine-x11 check to handle ozone glx plugin compilation.
For ozone egl backend reimplement missing functions.
Task-number: QTBUG-65682
Task-number: QTBUG-51631
Change-Id: I8b45c6f4d09a1c703efb12b6452d78ab385371ab
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
To be able to determine where a download was triggered.
[ChangeLog][DownloadItem] Added a page/view accessor to tell were
the download was triggered.
Change-Id: I21843a545a3e0eb66f5e5fa8a50e77564f2118a7
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Turns out we did no have DNS prefetching. It requires an extra
component. The same component can also do pre-connect, but we do not
enable that, as it requires a predictive browsing backend.
[ChangeLog][Settings] DnsPrefetchEnabled added, but disabled by default.
Change-Id: Ieb036435b9f1a72a7be302e38e89e0c347c7176b
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
On little-endian rgba8888 is abgr32, however we still
swapped bytes resulting in argb32. Use simply correct
format from beginning, which avoids swapping.
Change-Id: I6578d6dfe626d5afdcefcae6dea3ca44b23d4412
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It is where the active logic of UpdateCursor should now be. It doesn't
matter for us yet, but we might as well follow how it is supposed
to be done.
Change-Id: I9d7f2ed767c3eef03e637680166b75307081f057
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
We do not have a cursor manager.
Change-Id: Idf4b89f12a0480d7f522879ba07ee807aaef641f
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Specify exactly how we parse using QUrl, and try parsing invalid URLs
from GUrl as well.
Change-Id: If6e95d2600315180f36f9ad767f97936b438cd84
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Otherwise, setting a WebEngineProfile to non-off-the-record would assert.
Therefore, now it is safe to set the quicknanobrowser's defaultProfile to
non-off-the-record.
Task-number: QTBUG-69416
Change-Id: I5ca2230e12e457dd9963e386708f2394ae4c8cea
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Chromium has XSS auditing enabled by default - it seems like a bad idea for
QtWebEngine to disable a Chromium security feature by default.
The Google design document for the XSS auditor also claims[1]:
Processing costs are essentially zero unless the URL or POST body includes
any of the four characters " > < '. When those characters are found, we only
invoke heavy processing on those attributes that might be dangerous. This
stands in contrast to the XSS Filter in Internet Explorer, which runs costly
regular expressions.
Thus, worries about performance issues are likely unsubstantiated.
This can be tested using the first link on [2], where Chromium blocks loading by
default but QtWebEngine did not.
[1] https://www.chromium.org/developers/design-documents/xss-auditor
[2] http://webdbg.com/test/xss/auditor.aspx
[ChangeLog][QtWebEngine] The XSSAuditingEnabled setting is now enabled by
default, to mirror Chromium's behavior.
Change-Id: Id57177ec8183335bc94cc8c0c440ac1e10d2c74f
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Reviewed-by: Konstantin Tokarev <annulen@yandex.ru>
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|