| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Suppose having a QWebEngineView as one of many child widgets in a
layout. Resize some child widget such that the position of the
QWebEngineView changes (without changing the position of the top-level
window). Chromium must be informed of the changed global position,
otherwise popups will be opened at the old position.
Also see commit 7f941a34, which originally introduced the coordinate
propagation for top-level window moves.
Task-number: QTBUG-51244
Change-Id: Ieb372e8d7554700d5e8d1e2148ab778094ea3878
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
| |
This was a rudimentary test case for QWebSecurityOrigin.
This API was replaced in QtWebEngine.
Change-Id: If41a7f48906a6b49d5a83c72c69dbfd125804537
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
|
| |
This test requires the QtQuick test support API. Exclude it from build
if the API isn't available. Blacklist one problematic test case.
Remove the - now superfluous - ENABLE_QML_TESTSUPPORT_API checks.
Change-Id: I0b36de182628969e1185d685fa098d6140ae8d9c
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Previously if a selection was made on a web page, and afterwards a find
operation is executed, the selection in the web page would be cleared,
but the call to selectedText() would still return the old selection.
Make sure selectedText() is always cleared, when starting a find
operation, as well as when stopping one.
Change-Id: If78f0fa1dd836a52184015e749ef5a84b9f784cd
Task-number: QTBUG-54071
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Test was flaky due to a couple of reasons, mainly:
1) Size of event recording widget did not match web view size,
which led to some events not being caught.
2) There were some bogus QHoverEvents being sent by QQuickWindow
whenever a QQuickItem is repositioned (can happen when using
a tiling window manager on a Linux machine).
3) Usual timing issues.
Change-Id: Idd9d4fa6bd65095f98125196025e0ad560025348
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
| |
Switch from timed wait to using tryCompare.
Task-number: QTBUG-53777
Change-Id: I845fb8ee7b84641bd23314794512960634423d54
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
We were reporting any mouse button event where the left mouse button was
held as a left mouse button event. We should instead separate the button
changed and the ones held.
Also adds test of the translation.
Change-Id: I1139d6615d54a036dfe843fdb1d1c90b23b467b5
Reviewed-by: Alexandru Croitor <alexandru.croitor@theqtcompany.com>
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
We have changed qrc resources so that they can be accessed from any
security origin. To have a test of local resources we would need a file
URL.
Task-number: QTBUG-53153
Change-Id: I22b730f07785eacd365e3ec591f5dcbd09b6fc84
Reviewed-by: Alexandru Croitor <alexandru.croitor@theqtcompany.com>
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
|
| |
This test is unreliable in 5.6. We have a working version in 5.7.
Change-Id: I1e4e2abe896e479d8960a3a3b990d80d41b844f1
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
|
|
|
|
|
|
|
| |
We have no equivalent of QWebFrame::javaScriptWindowObjectCleared and
won't need it.
Change-Id: I998e0412256887414c125df9ab994216dee22dc9
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
If a QQuickWebEngineView does not have focus, and activeFocusOnPress is
set to false, a user can still partially interact with the view. For
instance hovering the mouse over a link would change the cursor, a link
can be clicked to go to a different page.
Clicking on a text input field would focus the text field, but entering
characters will not be possible, because the view does not have QtQuick
keyboard focus, and clicking does not give the focus (because
activeFocusOnPress is set to false) and this leads to confusing
behavior.
Thus the fix is to make sure no mouse / keyboard events are forwarded
to Chromium if the view has no focus, and activeFocusOnPress is set
to false, in order to maintain a more user-friendly behavior.
Manually forcing the focus via some user-provided method that calls
forceActiveFocus() would allow further proper interaction.
Change-Id: I72c3ff69438972b9a93ee2d415fa1d4b44b86cd9
Reviewed-by: Michael Brüning <michael.bruning@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently if a QWebEngineView or a QQuickWebEngineView is disabled
using setEnabled(false), after loading a web page, the views are
automatically focused, and a user might see a blinking caret in an html
input for example, even though the user can't interact with it.
Fix consists in not calling the Focus() method whenever a view is
disabled.
Change-Id: I1014fb5898a5ddf01a4e9b14c3eaf5d4006e5131
Task-number: QTBUG-53159
Reviewed-by: Michal Klocek <michal.klocek@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
If a QWebEngineView is disabled, input events are still forwarded to
Chromium, thus allowing the user to interact with a web page.
Fix consists in stopping the forwarding of input events in the generic
event handler, just like QWidget::event() does.
Change-Id: Ie822d1f3d640840569a282223d76749686cf3419
Reviewed-by: Michal Klocek <michal.klocek@theqtcompany.com>
|
|\
| |
| |
| | |
refs/staging/5.6
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
It used to be an expected fail, then it seemed to work in 5.6, and for
some reason it fails again in 5.6.1.
Change-Id: I95161e001c3b8af3ea48fc284e448a2d8853e108
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|/
|
|
|
|
|
|
|
| |
WebKit expects pasted images to have the MIME type image/png (see code
and comment in DataObjectItem::getAsFile()).
Task-number: QTBUG-53409
Change-Id: I2b0c1244d309687ad190db26c5b00718ed0c4258
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
Requesting a local file resource without a baseUrl set, should not be
allowed. This hasn't worked before in Webkit, so the checks were marked
with QEXPECT_FAIL. Remove these, as the issue has been fixed, and now
causes unexpected passes.
Change-Id: I7d38692a9961651d48abe28218d7ee022a85f101
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Switches HttpUserAgentSettingsQt to reading local string that lives on
the IO thread, instead of accessing a pointer that may be deleted.
Also adds a test and fixes updating accept-language of already
initialized WebContents.
Change-Id: Iff4e3bd1ac40482a92bed50c3f703ed8b974b0ad
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
Reviewed-by: Michal Klocek <michal.klocek@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It turns out that using the temporary zoom factor is meant to set
the zoom factor for one certain view, and this is actually what we
want.
Also added auto tests for this.
Task-number: QTBUG-51851
Task-number: QTBUG-51969
Change-Id: I8912cbc25637d3c1681026380a2ab3068a964868
Reviewed-by: Michal Klocek <michal.klocek@theqtcompany.com>
|
|
|
|
|
|
|
|
|
| |
- Add test for localStorageEnabled setting.
- Add test for javascriptEnabled setting.
Change-Id: I7a67c24bdf76409148e88500d9a7a092f7896493
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
Reviewed-by: Michael Brüning <michael.bruning@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
Adds test for user-agent override and fixes the override so that it
also works when loading with content instead of URLs.
Change-Id: I3f61b1d91b7b0908e35216722054168d1c514a87
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
| |
The test passes now.
Change-Id: Ifc96c995036759d554938dbbc80ca05a5680d496
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
This test is failing on 32 bit Windows. Unfortunately, we couldn't
reproduce the failure locally yet. Blacklist the test for now.
Change-Id: I7bdbc52033ccc0ed16ccf934418cc56a7783f88b
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
Subjects each *.png file that matched grep -law "sRGB" to:
pngcrush -ow -brute -rem allb -reduce -force
Change-Id: I0fe8ea5a1c89d5c9c0dce765607e55ae9e9cf9fd
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I877bbd8bc5c710370f135a27bcd7f0f7c95a7292
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
|
|
|
|
|
|
|
| |
This tested semi-public API. History points to a WRT-related feature.
Change-Id: I974feec55c5b243701bb2fa481ca16273088e438
Reviewed-by: Michael Brüning <michael.bruning@theqtcompany.com>
|
|
|
|
|
|
|
| |
This API doesn't exist. Meta data can be read by evaluating JavaScript.
Change-Id: I6089de34b463d34d94d4d05e723e12bc163d288d
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I94aa72517dfedc22d9100b85e4d66dbf67bb3207
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
This test makes only sense in QtWebKit where HTML input elements
are implemented with Qt widgets (e.g. QComboBox).
Change-Id: I0fdd63530db762595df7ca370ff2f9b1d73cc0a8
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
| |
QWebEngine will not provide a wrapper API for the now deprecated
Web SQL Database HTML5 API.
Change-Id: I8864b202ce9200145953ba5273f86eddff9e133a
Task-number: QTBUG-51346
Reviewed-by: Michael Brüning <michael.bruning@theqtcompany.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
Blacklist failing tests.
Stabilize some tests.
Change-Id: I9fba81fb70f5b358108ef6935e363297e2658f79
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
this fixes static builds by ensuring that all dependencies are exported.
Task-number: QTBUG-51071
Change-Id: I1d1b38dd5f78d8a949bab7f9a4643a2430657458
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I7c5f59a48688ca73889605ef5850ebedf43782de
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
|
|
|
|
|
|
| |
Change-Id: I6534d36279feffb66eb6a897ed31d4b013b39691
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: Ia0dcc45a3b63da49b82cd4f018979df43e847273
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
|
|
|
|
|
|
|
| |
Also, update the API list.
Change-Id: Ib21feb823ecf58c071316f796a7e251c9f99bff4
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I330de920c339e31725d123eadf5dcc4227dd500f
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
RenderFrameDeleted was called before DidFinishLoad, which meant
m_loadingErrorFrameList was empty and wouldn't detect the page loaded
was an error page. Instead this patch relies on the chromium error-page
url which we already asserted.
Additonally we delay emitting the loadFinished signal until the error
page is also done loading, since the error-page can be considered part
of the load, and we otherwise have a race condition on toPlainText.
Finally we were not getting error-pages when blocking requests because
we reported them as aborted them instead of blocked.
Change-Id: I945eb838b7f080d4e146f18354e8986e1b88b5cd
Task-number: QTBUG-50752
Reviewed-by: Michael Brüning <michael.bruning@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
| |
When a failed load is replaced by an error-page we currently get two
loadFinished signals. So disable error-pages to make the test of failed
load simpler.
Change-Id: I3baaf0f5a06c5e755695d6540316287a860e0d5f
Task-number: QTBUG-50752
Reviewed-by: Michael Brüning <michael.bruning@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Setting a new default QSurfaceFormat after
QtWebEngineCore::initialize() is called, might lead to a crash.
This happens when the new surface format has a different OpenGL profile,
compared to the profile created by web engine in the
RenderWidgetHostViewQtDelegateWidget constructor. The default
constructed QSurfaceFormat has an OpenGL Compatibility profile.
Inside the Cocoa platform plugin when a new shared OpenGL context is
created, it fails to initialize the new context because of the
difference in profiles, and thus ultimately creates an unshared
context, which leads to a crash.
Fix consists in using the shared context QSurfaceFormat in the
RenderWidgetHostViewQtDelegateWidget constructor, and also printing
a fatal warning to notify the developer only to set the new
QSurfaceFormat before the application instance is declared.
Bottom line, if the QSurfaceFormat OpenGL profile has to be
changed, it should be done before QtWebEngineCore::initialize() is
called. Doing so after initialize() is called, will lead to a crash.
Change-Id: I8a07211b592143d736b001556b944d4759802396
Task-number: QTBUG-50665
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
Reviewed-by: Michal Klocek <michal.klocek@theqtcompany.com>
|
|
|
|
|
|
|
|
|
| |
Work around QTBUG-50752 in URL scheme handler test. Triggering the
failing load makes subsequent calls to toPlainTextSync unreliable.
Change-Id: I370d59b935b186fa4e3be3097ee921802f951720
Task-number: QTBUG-50752
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
Sometimes the emission of loadFinished wasn't caught.
Use a QTRY_VERIFY-like function to synchronously load a page.
Change-Id: Ib260a126c0cff7f254c97b1cd89f49b806010235
Reviewed-by: Michael Brüning <michael.bruning@theqtcompany.com>
|
|
|
|
|
|
|
| |
Avoid the usage of mailto. It might get forwarded to QDesktopServices.
Change-Id: I2e8c889e8b0107b11c22c182a64685bc5ac16aae
Reviewed-by: Michael Brüning <michael.bruning@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Most page-state is set in webcontentsadapter::initialize except user
scripts and webchannel. This patch ensures those are initialized too
when changing to a new adapter during history restore.
Change-Id: I4dca23ddab50480b1a72252a038834ce1802ad77
Task-number: QTBUG-50751
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
Reviewed-by: David Rosca <nowrep@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For view-owned URL scheme handlers the destructor would remove the
handler and then trigger URLRequestContextGetterQt::generateStorage.
This would access the browser context from the IO thread while it
already has been destroyed on the browser thread.
Increment the ref count for the browser context before every call
of generateStorage, and decrement it when generateStorage is
finished.
Task-number: QTBUG-50160
Change-Id: Id8b1505891ec56e93bf9d47f33bb8bc3304eb55a
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
Fix a race condition where loadVisuallyCommitted was emitted before
the QSignalSpy in waitForSignal could be connected.
Change-Id: I7fcb9e3b423430405daf99c588ae8b70645333a1
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
Remove the 200 ms wait and instead try to grab the window contents
repeatedly until a certain condition is met.
Change-Id: I96dd720ef41ced9bdb48772939d25dd957081c33
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
URLRequestCustomJob::notifyFailure calls NotifyStartError(status), which
in turn will result in a call to URLRequestCustomJob::Kill.
We must release the lock of m_mutex before calling NotifyStartError,
otherwise m_mutex.lock() will wait forever in Kill.
Change-Id: I319e45049766c2192dfc46a91b352b92ec677bc6
Task-number: QTBUG-50160
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I372366afce46703b9960ef2a8e5f575eaedb2514
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
| |
OnCanSetCookie and OnCanGetCookies should
both be mapped to API. Since the other
call is missing (filter cookies which should
be send) remove existing one from 5.6
Change-Id: I4f42c4a1fee6add7a5efffaf4c38877a1f35ce61
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
|