| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
Task-number: QTBUG-71128
Task-number: QTBUG-69281
Task-number: QTBUG-71229
Change-Id: If6b839ce12efb3ccd69c270d99b10b3756014bb2
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|\
| |
| |
| | |
Change-Id: Iea03d76799b1e94991e536c056a5b918d63041a4
|
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-62053
Change-Id: I2df300239f4f02bb74f1ec27a74874b3877aabd0
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|/
|
|
|
|
|
|
|
|
|
| |
Somehow it's the case that QtTest 5.9 doesn't know how to QCOMPARE a QString
with a normal string literal:
ld: error: undefined symbol: bool QTest::qCompare<QString, char [31]>(QString
const&, char const (&) [31], char const*, char const*, char const*, int)
Change-Id: I13ce50689d9f1c157378b862fb1ed4a572ac7f79
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Although the test should have been fixed by the change and did succeed
during local tests and CI runs, it still fails on the CI.
This prevents a merge to dev.
Partially reverts commit: d62cb919722f0baad4d0826e566b9a63f7fc7488
Task-number: QTBUG-66527
Change-Id: I1571c918034d84bd969ca0d401d4cfb17a10fd68
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
| |
Not needed and triggers race condition.
Task-number: QTBUG-69007
Change-Id: Id57ba527387e5dbe44a8dd6c5a49e7278403ce64
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
The view did not get all the movement events if the mouse was moved
onto the view after it had loaded the page. Fix this by placing the
mouse onto the view before setting the html content.
Task-number: QTBUG-66527
Change-Id: I6d6847d5215005fe9490250b5b4f0c01ce317c23
Reviewed-by: Alexandru Croitor <alexandru.croitor@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>
|
|
|
|
|
|
|
| |
Avoids it crashing due to newer databases found
Change-Id: I5a3c926e5d13fe79d3af80e94e95516035ee75b8
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
|
|
|
|
|
|
| |
Fix classic race in using QSignalSpy::wait
Change-Id: If922609825c8652f326c2e252f64cc407e06ad9c
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
| |
Files for bic test added.
Change-Id: I1ae9f54d70a9891248ad9d65efbf9ec29f022452
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
| |
Calling triggerAction for a context-menu-specific action (like
DownloadLinkToDisk), when no context menu exists, will crash.
Change-Id: Ife647b7a7d8ab1a0d9dcb7d3566be130516e6e63
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is an improvement of:
bab4abab Keep settings synchronized when RVH swaps
Updating settings during RVH swap is flaky because it happens after
the load start of RenderFrame thus it is not guaranteed the
ViewMsg_UpdateWebPreferences message is received by the render process
before the load finishes. This means that the setting might not be applied
on the current page if it was set during the navigation.
The new RVH is created at the very beginning of the navigation, so access
it via the pending or speculative main frame of the page thereby make it
possible to update its settings before the load of RenderFrame starts.
Task-number: QTBUG-66656
Task-number: QTBUG-68424
Change-Id: I4fbba597579551bb3329936ccd7b357cf8daa4ea
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
|
|
|
|
|
| |
Task-number: QTBUG-68424
Change-Id: Idb72162b8aaa92a619fb7baf578f63e3d0171470
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
|
|
|
|
|
|
|
| |
Use qwebchannel.js shipped with qtwebchannel.
Change-Id: Iceead5131d22a6988bbb4e7a3cca1e9f21908e3e
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
There can be more than one RenderViewHost assigned to a single
WebContents. This is the case when starting a page load results changing
processes. When applying settings during the load it updates only the old
RVH. This fix applies the settings to the new RVH too after the swap.
Task-number: QTBUG-66656
Change-Id: I833415b2a34f58e2a9a18c209a025c82b73aa65a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Turns out JavaScript's JSON.stringify is not guaranteed to produce valid UTF-16
strings. It is possible in JavaScript to produce string objects which contain
invalid code units (unmatched surrogate pairs) and JSON.stringify will simply
copy this data to it's output. However, such a string cannot be losslessly
converted to UTF-8 and this leads to fun errors in WebChannelIPCTransport.
This patch
- Adds a test for the scenario above.
- Changes WebChannelIPCTransport to replace these invalid code units with the
Unicode replacement character U+FFFD.
- Changes WebChannelIPCTransportHost to validate the data it gets from the
renderer. Not validating the data defeats the whole point of Chromium's
fancy multi-process architecture: the renderer is not to be trusted.
- Changes WebChannelIPCTransport to throw JavaScript exceptions for various
errors (missing argument, wrong type, invalid JSON). Seems like the polite
thing to do.
Task-number: QTBUG-61969
Change-Id: I83275a0eaed77109dc458b80e27217108dde9f7b
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
| |
Change-Id: I9468faccfc8d943302bb643a57b9f7d1eed8862a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|\
| |
| |
| | |
Change-Id: Id41b0b07a934994471df40de884e1231ee1c0e8e
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
All pages in a site instance should belong to the same browser context, an
invariant which is violated by our WebContentsDelegateQt::OpenURLFromTab.
Task-number: QTBUG-67900
Change-Id: I754180bb96de07448ee3058210391f7856f4caeb
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
800x600 takes more space than expected when building a layout which
contains QWebEngineView. Now this function does exactly the same as
QWidget::sizeHint().
Add an auto test and update examples which were depending on this
hardcoded size.
Change-Id: I3b34b7d37cc99a8d6a49100d6c4b66b5b6a1c999
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Check that the workaround in src/3rdparty bf10f2f32a behaves as expected.
Task-number: QTBUG-62536
Change-Id: Ieae4ad61df303aae5c65d4f7be487574ce32eae0
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Shared workers work, service workers don't. Dedicated workers over custom
schemes trigger the same serialization error as WebSockets.
Task-number: QTBUG-62536
Change-Id: I97b4b8c267f6cfca74bc0abfcca905b0adeabe3d
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
User scripts with parse errors in their @match patterns were included
on all pages. Don't allow to fallback to "@include *" in this case.
Task-number: QTBUG-67726
Change-Id: I8172184e79fe3e515f391bc6cc8274a624e67a19
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Call RenderWidgetHostImpl::ImeCommitText() instead of
RenderWidgetHostImpl::ImeFinishComposingText() to trigger the necessary
JavaScript events on composing.
This fixes IME composition (eg. for dead keys) on web pages which use
custom JavaScript IME handler like facebook.
Task-number: QTBUG-66046
Change-Id: Ibc177995ba6e85eca42ae333decacfe6e788ce41
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|/
|
|
|
| |
Change-Id: Ib63727794a794c20b1ca372ab6457c2a8a141762
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|
|
|
|
|
|
|
|
|
|
| |
Unlike http/https, the same-origin policy for file, qrc and custom schemes is
not specified in any standard and, for the latter two, is not tested by Chromium
either. Therefore some tests might be useful.
Task-number: QTBUG-62536
Change-Id: I194d06c16c44b0a78bf4dba57415a2d2c6b991cc
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
| |
Open devtools from autotest.
Change-Id: I5e78d6259fdd29df5608b45292ee0a786d179e26
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
Remove 'permission' from class and signal names, so e.g.
QWebEngineQuotaPermissionRequest becomes QWebEngineQuotaRequest and
quotaPermissionRequested becomes quotaRequested.
Rename the internal controller classes to public name + "Controller".
Change-Id: I247714ab0a2880adbf4ed8ee68f1b78838ae7a14
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
| |
Seems making a context menu too early is racy.
Change-Id: I9cf268d506e6fcee348102cc3138c475ac94d55c
Reviewed-by: Valentin Fokin <fokinv@inf.u-szeged.hu>
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|
|
|
|
|
|
|
|
| |
Avoid hitting undefined reference to
'bool TestNamespace::QTest::qCompare<unsigned long, unsigned int>...'
Change-Id: I91588869708899e1055619ca9ba46d2061869417
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Avoid creating a blank WebContents for new QWebEnginePages only to replace it
with a different blank WebContents moments later. This problem is already
solved in QQuickWebEngineView by lazy initialization, therefore the patch
moves this lazy initialization into WebContentsAdapter itself so that it can
be applied to both QQuickWebEngineView and QWebEnginePage.
- Try to delay WebContents creation until the first navigation so that we have
enough information available to pick the right initial SiteInstance when
calling WebContents::Create. This is done by triggering WebContents creation
from the first call to a WebContentsAdapter::load* method.
- Use the SiteInstance from WebContentsDelegate::OpenURLFromTab. This method
gets a SiteInstance pointer which we should give to WebContents::Create.
Task-number: QTBUG-65561
Task-number: QTBUG-67163
Change-Id: Id7b351998adefb810cf27c61a1447b61d7f4c606
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
This ties the indexed-db, DOM storage and filesystem access to
cookie rights as they can all do the same cookies. A previous
patch did the same for workers, but we were missing this
class to complete the logic.
Change-Id: I5ea894b6c631bcf25439759174866d260b91b71a
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|
|
|
|
|
| |
Change-Id: Ibca5362c9d1e388885132259fa95a9f575f45fc8
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The waitForRequest/waitForSignal function used by these tests is broken: it
assumes that calling QEventLoop::exit() will immediately exit from the event
loop. In actuality this is not immediate and the event loop may continue to
execute more signals handlers even after exit() has been called. This means that
in e.g. waitForRequest the 'result' variable may be assigned to twice.
Additionally there is a race condition in downloadTwoLinks, where we sometimes
skip the first download. This is not a bug but simply one pending navigation
being aborted in favor of another.
Changes
- Delete waitForRequest. Define one HTTP request handler per test, using state
variables to communicate with the main body of the test. Ignore unknown
requests (including favicon requests). Same for downloadRequested signals.
- Expand downloadTwoLinks and fix expectations.
- Add logging to HTTPServer.
- Unblacklist.
Task-number: QTBUG-66888
Change-Id: I718cac6c4b32a8cc68400fa8ee5b853686c77fcb
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Replace injectionPoint test with a timeout-less version that uses
- both a single-frame and a multi-frame page
- both main and isolated worlds
- cross-process navigation
- window.open
- profile-wide scripts
Task-number: QTBUG-66338
Change-Id: Ica4acb8ada4acc38aa5e1ca00e7512a2e69b785f
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
|
|
|
| |
Moving mouse cursor to (0, 0) by QTest::mouseMove() doesn't seem to work
and the cursor doesn't leave the QWebEngineView. Change initial coordinates
as a workaround.
Change-Id: Iaad895c770a6edc5ed6e375f68b2b56196d84073
Reviewed-by: Viktor Engelmann <viktor.engelmann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
As of version 63, Chromium creates proxy frames also for the main frame in the
frame tree during cross-process navigations. This leads to a segmentation fault
in WebChannelIPCTransport because we assume that all main frames are local.
See https://crrev.com/27caae83cb530daaf49f9a38793e427cdf493a65 for details.
This patch refactors the renderer-side WebChannelIPCTransport from a
RenderViewObserver into a RenderFrameObserver, which prevents the segmentation
fault since the RenderFrameObserver is not created for proxy frames. Most likely
this would have to be done eventually anyway since the RenderView and
RenderViewObserver classes are deprecated and will likely be removed as part of
the Site Isolation project.
Installation is changed to follow Chromium's RenderFrameImpl in the sense of
performing the installation from RenderFrameObserver::DidClearWindowObject
instead of ContentRendererClient::RunScriptsAtDocumentStart. This has the
benefit of avoiding the ScriptForbiddenScope DCHECK.
Additionally there are the following minor changes:
- The deprecated parameterless version of v8::Value::ToObject() method is
replaced with v8::Value::IsObject() check and v8::Local::Cast.
- The deprecated v8::Handle typedef is replaced with v8::Local.
- The deprecated single-parameter WebContentsObserver::OnMessageReceived is
replaced with the new two-parameter version.
- blink::MainThreadIsolate() is used instead of v8::Isolate::GetCurrent() for
Install/Uninstall since we know we are executing on the main thread.
- WebChannelIPCTransportHost is changed to ignore messages from unexpected
renderers in case something goes wrong with the renderers.
- Logging is added to WebChannelIPCTransportHost for debugging purposes.
Some new unit tests are added, all of which fail with the old version.
Task-number: QTBUG-66333
Change-Id: I936d142fb042d9f936a3f9d08d4328ecba595f1f
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
| |
Remove testing of provisionalLoad, use QTRY_COMPARE(spy.count())
instead of spy.wait() to determine number of actual calls.
Change-Id: I37cbca8f27d4e12d42c7b30b8362249c1f76c79e
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
|
|
|
| |
Allows us to use return value instead of magic properties
Task-number: QTBUG-66825
Change-Id: I55202a7aca2e662bc214b05caf767d2da496f026
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
|
| |
It is more flaky with 65-based, so blacklist it until fixed.
Change-Id: I580dd858068b1ad54203618bd241294d17010a19
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
|
|
| |
Form validations messages has moved entirely to being done by Blink.
Change-Id: I6742c111fc59f0baba75b8b37f5d0ec9ae2fb920
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
| |
Change-Id: I121b14d6a44e80f5eea4b159c58f7010d472926e
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
|
|
|
|
|
| |
Task-number: QTBUG-65997
Change-Id: Icb55326c51f1dfff77e8e862e9ced619be17ead1
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
After stopping the load the active URL should be the url of the previous
successful load. Test active URL properly to make the test independent
from the previously executed test.
Change-Id: I53800fe28a722a5f3554b1ff2d4fc31cb8638dfd
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
In-page navigations does not trigger loadFinished thus
m_lastLoadProgress is not reset to -1. Do not ignore loadStarted signal
if m_lastLoadProgress is set to 100 because it can mean a new navigation
after an in-page one.
Task-number: QTBUG-66869
Change-Id: I2b556fe7e3e097c338bdd7ca47e7384722e7eca9
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
| |
Scheme refers to a part of a URL. Protocol refers to a formal standard for
communication over a network. Therefore scheme is more accurate here, since
registerProtocolHandler only defines simple aliases.
Change-Id: I71a1511aa92e41c3e3aca4c0403f54af6d91c563
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Move QWebEngineQuotaPermissionRequest and
QWebEngineRegisterProtocolHandlerPermissionRequest to QtWebEngineCore.
- Delete QQuickWebEngineQuotaPermissionRequest and
QQuickWebEngineRegisterProtocolHandlerPermissionRequest.
Miscellaneous cleanup:
- Mark QWebEngineQuotaPermissionRequest constructors as internal for QDoc.
- Remove superfluous Q_DECLARE_METATYPEs (implied by Q_GADGET).
- Remove Q_UNREACHABLE from default constructor.
For some reason QML seems to default construct an unused temporary object
before throwing it away and copy constructing the actual object. This triggers
Q_UNREACHABLE.
Change-Id: Icf9f4e34996e4c64aec65b734bcb3bbd22b4dc51
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
The previous test will set the webEngineView.loadProgress to 100 due to
a successful load. Therefore, the test can't expect 0 loadProgress at
the beginning. Count the loadProgress changes instead of checking the
initial value.
Change-Id: I8f4727d5b8c7ebf1b7c640e44a9eb3ca5a7fef44
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
| |
Task-number: QTBUG-66888
Change-Id: I11a1526bf6478fc5ffb51f208b59cff7d9a37220
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|