electron/patches/chromium/webview_fullscreen.patch
electron-roller[bot] bccb7b572a
chore: bump chromium to 123.0.6272.0 (main) (#41130)
* chore: bump chromium in DEPS to 123.0.6265.0

* chore: bump chromium in DEPS to 123.0.6266.0

* chore: update feat_ensure_mas_builds_of_the_same_application_can_use_safestorage.patch

Xref: https://chromium-review.googlesource.com/c/chromium/src/+/5232280

reverts the revert *and* undoes the rewrites 😵

* chore: update chromium/disable_hidden.patch

Xref: https://chromium-review.googlesource.com/c/chromium/src/+/5148491

minor manual intervention

* chore: update patches

* chore: Rename mouse lock to pointer lock

Xref: https://chromium-review.googlesource.com/c/chromium/src/+/5148491

* refactor: fix use of newly-removed QuitCurrentWhenIdleClosureDeprecated()

Xref: https://chromium-review.googlesource.com/c/chromium/src/+/4710351

Xref: https://github.com/electron/electron/pull/26022

This usage was added in #26022 as part of Wayland support and some
simple fiddles WfM with this patch on Wayland, but I'm unsure what
to be testing for & would like a second opinion on this commit.

* chore: bump chromium in DEPS to 123.0.6268.0

* chore: update patches

* 5235021: WebUI: Migrate accessibility and bluetooth internals to WebUIConfig
https://chromium-review.googlesource.com/c/chromium/src/+/5235021

* address review comment for 5196547

See: https://chromium-review.googlesource.com/c/chromium/src/+/5196547
Review comment: https://github.com/electron/electron/pull/41130#discussion_r1468973786

* 5075962: Introduce URLLoaderFactoryBuilder
https://chromium-review.googlesource.com/c/chromium/src/+/5075962

* chore: bump chromium in DEPS to 123.0.6270.0

* the missing semicolon strikes again

* chore: update chromium patches

* chore: update v8 patches

* 5242326: Some followup cleanup to Fuchia files in chrome | https://chromium-review.googlesource.com/c/chromium/src/+/5242326

* chore: bump chromium in DEPS to 123.0.6272.0

* chore: update patches

* chore: update patch after rebase

* 5247339: [Clipboard] Remove ClipboardContentType enum.

https://chromium-review.googlesource.com/c/chromium/src/+/5247339

* Reland: "Allow content shell to enable the built-in DNS resolver"

https://chromium-review.googlesource.com/c/chromium/src/+/5232354

* 5018206: [accessibility] Migrate to ScopedAccessibilityMode

https://chromium-review.googlesource.com/c/chromium/src/+/5018206

* 5246669: [Default Nav Transition]Add GetBackForwardTransitionAnimationManager on WebContentsView | https://chromium-review.googlesource.com/c/chromium/src/+/5246669

* fix: move nut-js to optional spec deps (#41199)

* chore: bump chromium in DEPS to 123.0.6265.0

* chore: bump chromium in DEPS to 123.0.6266.0

* chore: update feat_ensure_mas_builds_of_the_same_application_can_use_safestorage.patch

Xref: https://chromium-review.googlesource.com/c/chromium/src/+/5232280

reverts the revert *and* undoes the rewrites 😵

* chore: update chromium/disable_hidden.patch

Xref: https://chromium-review.googlesource.com/c/chromium/src/+/5148491

minor manual intervention

* chore: update patches

* chore: Rename mouse lock to pointer lock

Xref: https://chromium-review.googlesource.com/c/chromium/src/+/5148491

* refactor: fix use of newly-removed QuitCurrentWhenIdleClosureDeprecated()

Xref: https://chromium-review.googlesource.com/c/chromium/src/+/4710351

Xref: https://github.com/electron/electron/pull/26022

This usage was added in #26022 as part of Wayland support and some
simple fiddles WfM with this patch on Wayland, but I'm unsure what
to be testing for & would like a second opinion on this commit.

* chore: bump chromium in DEPS to 123.0.6268.0

* chore: update patches

* 5235021: WebUI: Migrate accessibility and bluetooth internals to WebUIConfig
https://chromium-review.googlesource.com/c/chromium/src/+/5235021

* address review comment for 5196547

See: https://chromium-review.googlesource.com/c/chromium/src/+/5196547
Review comment: https://github.com/electron/electron/pull/41130#discussion_r1468973786

* 5075962: Introduce URLLoaderFactoryBuilder
https://chromium-review.googlesource.com/c/chromium/src/+/5075962

* chore: bump chromium in DEPS to 123.0.6270.0

* the missing semicolon strikes again

* chore: update chromium patches

* chore: update v8 patches

* 5242326: Some followup cleanup to Fuchia files in chrome | https://chromium-review.googlesource.com/c/chromium/src/+/5242326

* chore: bump chromium in DEPS to 123.0.6272.0

* chore: update patches

* chore: update patch after rebase

* 5247339: [Clipboard] Remove ClipboardContentType enum.

https://chromium-review.googlesource.com/c/chromium/src/+/5247339

* Reland: "Allow content shell to enable the built-in DNS resolver"

https://chromium-review.googlesource.com/c/chromium/src/+/5232354

* 5018206: [accessibility] Migrate to ScopedAccessibilityMode

https://chromium-review.googlesource.com/c/chromium/src/+/5018206

* 5246669: [Default Nav Transition]Add GetBackForwardTransitionAnimationManager on WebContentsView | https://chromium-review.googlesource.com/c/chromium/src/+/5246669

* fix: typescript error for optional dependency

(cherry picked from commit ea17105c8c01bd1f4c1df5e72190a687d92f93cc)

---------

Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com>
Co-authored-by: Charles Kerr <charles@charleskerr.com>
Co-authored-by: PatchUp <73610968+patchup[bot]@users.noreply.github.com>
Co-authored-by: clavin <clavin@electronjs.org>
Co-authored-by: VerteDinde <vertedinde@electronjs.org>
Co-authored-by: John Kleinschmidt <jkleinsc@electronjs.org>
Co-authored-by: Sam Maddock <smaddock@slack-corp.com>
2024-01-31 17:04:13 -06:00

102 lines
5.2 KiB
Diff

From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Cheng Zhao <zcbenz@gmail.com>
Date: Thu, 4 Oct 2018 14:57:02 -0700
Subject: fix: also propagate fullscreen state for outer frame
When entering fullscreen with Element.requestFullscreen in child frames,
the parent frame should also enter fullscreen mode too. Chromium handles
this for iframes, but not for webviews as they are essentially main
frames instead of child frames.
This patch makes webviews propagate the fullscreen state to embedder.It also handles a
DCHECK preventing guest webcontents from becoming the focused webContents.
Note that we also need to manually update embedder's
`api::WebContents::IsFullscreenForTabOrPending` value.
diff --git a/content/browser/renderer_host/render_frame_host_impl.cc b/content/browser/renderer_host/render_frame_host_impl.cc
index ed356e5a1f97afa8a133ae1ab8b7f6a0fc9a97ac..f217dbdef519759971a949c3f0ffcbd098669537 100644
--- a/content/browser/renderer_host/render_frame_host_impl.cc
+++ b/content/browser/renderer_host/render_frame_host_impl.cc
@@ -7502,6 +7502,17 @@ void RenderFrameHostImpl::EnterFullscreen(
}
}
+ // Entering fullscreen from webview should also notify its outer frame.
+ if (frame_tree_node()->render_manager()->IsMainFrameForInnerDelegate()) {
+ RenderFrameProxyHost* outer_proxy =
+ frame_tree_node()->render_manager()->GetProxyToOuterDelegate();
+ DCHECK(outer_proxy);
+ if (outer_proxy->is_render_frame_proxy_live()) {
+ outer_proxy->GetAssociatedRemoteFrame()->WillEnterFullscreen(
+ options.Clone());
+ }
+ }
+
// Focus the window if another frame may have delegated the capability.
if (had_fullscreen_token && !GetView()->HasFocus())
GetView()->Focus();
diff --git a/content/browser/web_contents/web_contents_impl.cc b/content/browser/web_contents/web_contents_impl.cc
index 4a911829b833a952e4e63e2a7ce4850918eec84b..05d42b8808ac3dbb2c3ccfe3213fd9715a373960 100644
--- a/content/browser/web_contents/web_contents_impl.cc
+++ b/content/browser/web_contents/web_contents_impl.cc
@@ -3712,21 +3712,25 @@ KeyboardEventProcessingResult WebContentsImpl::PreHandleKeyboardEvent(
const NativeWebKeyboardEvent& event) {
OPTIONAL_TRACE_EVENT0(TRACE_DISABLED_BY_DEFAULT("content.verbose"),
"WebContentsImpl::PreHandleKeyboardEvent");
- auto* outermost_contents = GetOutermostWebContents();
- // TODO(wjmaclean): Generalize this to forward all key events to the outermost
- // delegate's handler.
- if (outermost_contents != this && IsFullscreen() &&
- event.windows_key_code == ui::VKEY_ESCAPE) {
- // When an inner WebContents has focus and is fullscreen, redirect <esc>
- // key events to the outermost WebContents so it can be handled by that
- // WebContents' delegate.
- if (outermost_contents->PreHandleKeyboardEvent(event) ==
- KeyboardEventProcessingResult::HANDLED) {
+
+ auto handled = delegate_ ? delegate_->PreHandleKeyboardEvent(this, event)
+ : KeyboardEventProcessingResult::NOT_HANDLED;
+
+ if (IsFullscreen() && event.windows_key_code == ui::VKEY_ESCAPE) {
+ if (handled == KeyboardEventProcessingResult::HANDLED)
return KeyboardEventProcessingResult::HANDLED;
+
+ // When an inner WebContents has focus and is fullscreen, traverse through
+ // containing webcontents to any that may handle the escape key.
+ while (auto* outer_web_contents = GetOuterWebContents()) {
+ auto result = outer_web_contents->PreHandleKeyboardEvent(event);
+ if (result == KeyboardEventProcessingResult::HANDLED) {
+ return KeyboardEventProcessingResult::HANDLED;
+ }
}
}
- return delegate_ ? delegate_->PreHandleKeyboardEvent(this, event)
- : KeyboardEventProcessingResult::NOT_HANDLED;
+
+ return handled;
}
bool WebContentsImpl::HandleMouseEvent(const blink::WebMouseEvent& event) {
@@ -3862,7 +3866,7 @@ void WebContentsImpl::EnterFullscreenMode(
OPTIONAL_TRACE_EVENT0("content", "WebContentsImpl::EnterFullscreenMode");
DCHECK(CanEnterFullscreenMode(requesting_frame, options));
DCHECK(requesting_frame->IsActive());
- DCHECK(ContainsOrIsFocusedWebContents());
+ DCHECK(ContainsOrIsFocusedWebContents() || IsGuest());
// When WebView is the `delegate_` we can end up with VisualProperties changes
// synchronously. Notify the view ahead so it can handle the transition.
diff --git a/third_party/blink/renderer/core/fullscreen/fullscreen.cc b/third_party/blink/renderer/core/fullscreen/fullscreen.cc
index eb1ca15be1502b49b5b864935acbafbd6c75d8aa..1e9c96808cdcbf1131a07f33f36b5666bb9ae1e8 100644
--- a/third_party/blink/renderer/core/fullscreen/fullscreen.cc
+++ b/third_party/blink/renderer/core/fullscreen/fullscreen.cc
@@ -111,7 +111,7 @@ void FullscreenElementChanged(Document& document,
// is the iframe element for the out-of-process frame that contains the
// fullscreen element. Hence, it must match :-webkit-full-screen-ancestor.
if (new_request_type & FullscreenRequestType::kForCrossProcessDescendant) {
- DCHECK(IsA<HTMLIFrameElement>(new_element));
+ // DCHECK(IsA<HTMLIFrameElement>(new_element));
new_element->SetContainsFullScreenElement(true);
}
new_element->SetContainsFullScreenElementOnAncestorsCrossingFrameBoundaries(