| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
| |
The function signature changed though the ABI stayed the same.
Change-Id: I86ca361b5e4f0c523e1031910df438c23beee876
Fixes: QTBUG-77037
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
|
|
|
|
|
|
|
|
| |
We use UrlPattern to parse @match rules from the metadata headers
of user scripts.
Task-number: QTBUG-76173
Change-Id: Ibd1b73508e3229bcc5babe26e0048da069da6dec
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Incoming tablet events are forwarded as mouse events to Chromium,
because its pointer event unification is still WIP. The following
change introduced a behavior which sets the pressure property to 0
when no button is pressed:
https://chromium-review.googlesource.com/c/chromium/src/+/626039/
This is acceptable for mouse events, but breaks our tablet support.
Revert this until Chromium has a proper API and support of tablet
events.
Change-Id: Ieeffd22fc9337df342367826e39286352c6ba43b
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
| |
Task-number: QTBUG-71885
Change-Id: I441bd742f3ba6d16bfe58d88574b628d3561ecda
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|
|
|
|
|
|
|
|
| |
Needed for root_layer_size.
Task-number: QTBUG-71885
Change-Id: I1d820c98c103ed5987dcf4d7b7aaea254894df2d
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
| |
Change-Id: Ia3744634a746fe49b34e10a3b97762fa5b67745c
Fixes: QTBUG-76869
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
This patch fixes the iterator debugger assert on Windows.
This bug is introduced by:
https://chromium-review.googlesource.com/c/chromium/src/+/1369780
Change-Id: I9a77dac13d226e09d32b08e394ac1a441672668b
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Device Service needs to clean up parts of its internal state as part
of browser shutdown, notably, BatteryStatusService. However, it also
needs to run on the UI thread, and embedded services that run on the UI
thread are not guaranteed to be destroyed as part of browser shutdown
(tasks to destroy these services are posted from the IO thread by
ServiceManagerConnectionImpl::ShutDownOnIOThread, but the UI thread is
typically shut down before these posted tasks are run).
To solve this issue we discussed adding plumbing wherein embedded
services could inform //content that they wanted to be notified when
shutdown was occurring on the main thread. However, on investigation
this plumbing would be painful to implement: it is only
EmbeddedInstanceManager that has direct information of these service
instances, and that object lives far away from //content's
ServiceManagerContext, the object that knows when shutdown is occurring
on the main thread.
This CL takes an alternative approach of having the BatteryStatusService
object be stashed in a SequenceLocalStorageSlot to ensure that it is
destroyed on browser shutdown. I have verified that its destruction is
triggered on shutdown of content_shell.
This CL is conceptually a reland. The original CL (which had the Device
Service directly observe destruction of the UI thread) was reverted for
the following reason:
"crbug.com/856771 presents a crash that it introduced on ChromeOS: the
Device Service instance is now shut down after the DBusThreadManager
global instance, on which it implicitly depends."
However, crbug.com/794105 has continued to be a problem on Windows,
and there aren't active engineering resources for an alternative solution
to this one. Hence, this CL relands an approach that is similar to the
original fix with the shutdown not being applied on ChromeOS to avoid
triggering crbug.com/856771. As I wrote in the revert, "this is all
sad-making, but c'est la vie."
The usage of SequenceLocalStorageSlot rather than direct observation of
MessageLoop is thanks to a suggestion from rockot@.
Bug: 794105, 856771
Change-Id: I700470b12bc8c4d41406657923c25324e0ec05f4
Commit-Queue: Ken Rockot <rockot@google.com>
Reviewed-by: Ken Rockot <rockot@google.com>
Cr-Commit-Position: refs/heads/master@{#665919}
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
One page proposal: https://docs.google.com/document/d/1I-d_va17XsFDPS3PqgwXtdKnpJNF6eiWJ7GsNrVz9AM/edit?usp=sharing
Change-Id: Iac1b3c85936e6331142b52075678b1f629542854
Commit-Queue: Allen Bauer <kylixrd@chromium.org>
Reviewed-by: François Doray <fdoray@chromium.org>
Reviewed-by: Peter Kasting <pkasting@chromium.org>
Reviewed-by: Trent Apted <tapted@chromium.org>
Cr-Commit-Position: refs/heads/master@{#659699}
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This CL fixes a few issues with SequenceLocalStorage:
- Supports holding T with disabled copy/move constructor
by adding emplace().
- Prevents auto default construction with GetValuePointer.
- Query if SequenceLocalStorageSlot currently holds a value
with operator bool().
- Update doc to use NoDestructor.
This makes it possible to replace in many cases:
SequenceLocalStorageSlot<unique_ptr<T>> with
SequenceLocalStorageSlot<T>
and
auto ptr = sls.Get();
if (!ptr) {
ptr = make_unique<T>();
sls.Set(ptr);
}
return *ptr;
with
return sls.GetOrCreateValue();
TBR=fdoray@chromium.org
Change-Id: I3ebb4201fa6d0df296d4a4eaa9b90a1692f3a772
Commit-Queue: Etienne Pierre-Doray <etiennep@chromium.org>
Reviewed-by: François Doray <fdoray@chromium.org>
Cr-Commit-Position: refs/heads/master@{#664708}
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In C++20, std::memory_order is a scoped enum, and
std::memory_order_xyz are inline constexpr variables of type
std::memory_order provided for SC. The present code, however, depended
on the implicit conversion from the normal enum that std::memory_order
was in C++17 to int, which no longer works with scoped enums.
Fix by adding an explicit cast. This works on all standards.
Change-Id: I2f6aa17b671bc8c56e00558eb4133f0db5e3eb61
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
| |
Fix blink::SecurityOrigin::ToString to match url::Origin::Serialize.
Task-number: QTBUG-75651
Change-Id: I46998fb0b4b9f7d537a3a99e2ee1fc5814bb9eec
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
| |
Add CustomScheme::CorsEnabled flag.
Task-number: QTBUG-75651
Change-Id: Id87f2d59b446541c310229d7d99d03972a375695
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Also remove the DCHECK in debug builds.
Should fix the crashes in
tst_QWebEngineView::inputContextQueryInput().
Note this is just a band-aid fix to prevent a crash, the underlying
issue is still there.
Change-Id: Ie416f3517896a0ad978540ae81d1401af4bfbacc
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
| |
Change-Id: If073a1a427eaa0761796806f64e0767b5370f658
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
|
|
|
|
|
|
|
| |
Only wider in MSVC debug ABI.
Change-Id: Ic733442e88a60c0d64933b7d61190fbae69f1b33
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
| |
Without full_wpo_on_official, object files are not linked with LTCG, so
any data generated by /GL is just wasted.
Change-Id: I1cfc40e3775c4265d7ffb235a49252b1a484a69e
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
|
| |
Gcc 6 seems to misscompile pair's template constructor
in std::map emplace, use converting move constructor instead.
Task-number: QTBUG-75097
Change-Id: I813175ff9bf69276dc5e0340ce833fa56fada736
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
|
| |
Chromium pulls in the dsymutil from a separate package. We want to
use the same that is used for the rest of Qt.
Change-Id: I443fd76fbb4ced600061ed0e4314ef13c3db05d8
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
| |
Change-Id: I7c0a85e1fa266334d13051e37885f982e26c9368
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
|
|
|
|
|
|
|
| |
Rhel machines on ci do not have it.
Change-Id: I4cbb6d9d0826a23f8065a9eebba7ba27856e44d4
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|\
| |
| |
| | |
Change-Id: I9aae49aaae80bac927614506984912b259df8631
|
| |
| |
| |
| |
| | |
Change-Id: Ifcd5227841577e8ce81a1b7a54c56caba4d85e02
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Only showed up on CI
Change-Id: Icf1e81ec7449c5e4b939e1c1c0e05087ad6cab30
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
It was using FILE_EXE which is the application executable not the
subprocess one. Only worked because we returned the subprocess
executable for FILE_EXE.
Change-Id: I0a3c4c9c621ea23df086c13197ff76477097cfb7
|
| |
| |
| |
| |
| | |
Change-Id: I98a57b10a04d05fc47efa6e674d229cc246d9b32
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Working around various issues with gcc, and inconsistent use of
constexpr.
Change-Id: Iec46c4362f57d6aa7a8a37784dfe46c0b6563044
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
One more illegal narrowing
Change-Id: Idbd07c9dc3561b1d664a9c747c1f7ebfab7d5773
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Chromium requires the October 2018 Windows SDK. This change documents
and enforces that.
The previous enforcement code was buggy - a previous change updated
the message without updating the check. That's why the version number
skips ahead three instead of just two.
crrev.com/c/1546204 is the first change that requires this SDK.
Change-Id: Iaa1ee2152a07704fa1c7fcb601e2aa1eab92f264
Reviewed-by: Tommy Steimel <steimel@chromium.org>
Commit-Queue: Bruce Dawson <brucedawson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#654094}
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
For some reason it can't be passed by copy.
Change-Id: I76707f447c98bc4be8e6dfe0cd2c29f04e0b56fe
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I81b6d3dba53e9e58632107bdeda33c0c996506de
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: Ic408e6ab9559f9f76eff62d6695d103848a85ab0
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Cherry-pick to fix MSVC builds
Change-Id: I1ad38579162aecb6867c8f82b5157815d4ad06ee
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I3e7b09d39dcfbf13b603bc64f4e21fdcf685f27d
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Also remove feedback_private
Change-Id: I00bef976a03cd1d5ea949a5968af8dddbedbe30a
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: Iddb00ae4d7362da4b90b7f185acb6914c1b4ff4b
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|
| |
| |
| |
| |
| |
| |
| | |
COMPILER_GCC is not defined here.
Change-Id: I664b7da04499fa04df0ed9f7667cbca63ee6bf39
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
It wouldn't be updated if the custom user-agent didn't change.
Change-Id: Ic31ef485e2cb84903f9b48cb9ad5f9f1a723eb92
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This simplification doesn't work on MSVC due to some MSVC bug, so stick
with the old version for now.
Reverting https://chromium-review.googlesource.com/1118061
Change-Id: I3dff6075f9cc37dcc39c26b4f24de0501c8a362f
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I8bb0b269e7a5c793b1fc5a1d9aa260d5e09cd36e
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I9bad374117fceafdaa07326ad4468cda06184248
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
That's just sad.
Change-Id: I447e4b73c953ea9a90a2256d116a750aca044f6d
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Causes asserts on render process crashes otherwise.
Change-Id: I81a7ed3277752c20ab4fc06bd394faa8331d6933
Reviewed-by: Michael Brüning <michael.bruning@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Correct a few mistakes and add missing cases of illegal narrowing.
Change-Id: Ia25f9243787890daafc4b118f29e23b8de947a56
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|
| |
| |
| |
| |
| |
| |
| | |
Cherry-pick to fix MSVC builds
Change-Id: I43b830595119cc57e592c60bbf351ab2a68cc36a
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I459ffb41a73a244922917045fe0d0b3a3862b70b
Reviewed-by: Michal Klocek <michal.klocek@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We need this change so we can replace the previous user PrefService
with a new one when changing the data path in QWebEngineProfile. If
we don't have this change, the code will assert.
This is required to properly fix QTBUG-75740.
Change-Id: I956f0b0549caf258cc49ac0818f6eca0ddcbafd7
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If the PulseAudio daemon has no sources and/or sinks, then the pa_server_info's
default_source_name and/or default_sink_name fields will be null pointers.
Assigning (const char *)nullptr to a std::string triggers a segfault.
Task-number: QTBUG-75629
Change-Id: I056148af53110943f1ee025a6d94d484c8590c4e
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
(cherry picked from commit 8d400b021591c0b619269317e239d8a750ace140)
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Returns the top level frame URL. Used instead of site_for_cookies()
in case of cross-origin iframe navigations.
Task-number: QTBUG-70790
Change-Id: Icec4beddd70b20120d53e17a35385d83d1276ac2
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
The memory instrumentation implemenatation for Linux requires static
linking, so replace it with the stub implemenation.
Change-Id: I7f3c027b27cc309c1acef26a9d5952f1c1ee02e4
Reviewed-by: Peter Varga <pvarga@inf.u-szeged.hu>
|