Add the generated includes/ directory to CEF's "config" so that source files
included in patched Chromium targets (for example, blink_glue.cc) can find
cef_config.h which will be included via `include/internal/cef_types_linux.h`
on Linux.
Fixes a shutdown crash due to `ExtensionsBrowserClient::Set(nullptr)` being
called too early. Some code that may occasionally be triggered via
`content::ContentMainShutdown()` is expecting the extensions objects to still
be valid.
This new ownership pattern matches the code in chrome/.
URLLoaderNetworkServiceObserver is used for routing certificate and
authentication callbacks from the NetworkService to the associated
StoragePartition instance. With request interception enabled this object was
previously only assigned for the initial request. This change adds assignment
for restarted/redirected requests as well.
With PdfUnseasoned the PDF file will be loaded in a dedicated renderer process.
We identify this process by adding the kPdfRenderer command-line flag (similar
to how kExtensionProcess is used to identify an extension renderer process). We
then avoid calling GetNewBrowserInfo for the PDF renderer process because we
know the request will otherwise time out.
We need to override ChromeMimeHandlerViewGuestDelegate to handle
OnGuestAttached/Detached callbacks in order to account for the guest renderer
process hosting the PDF extension.
Additional work will be required to account for the renderer process hosting the
PDF viewer when using `--enable-features=PdfUnseasoned` (see issue #2969).
Browser windows created via CefBrowserHost::CreateBrowser should be shown at
creation time. These windows receive default Chromium styling, unlike
CefBrowserView windows which may contain other client-specified Views and
where show state can be configured via CefWindowDelegate::GetInitialShowState
or by calling CefWindow::Show.
- Remove CefRequestContextHandler::OnBeforePluginLoad and
CefRequestContext::PurgePluginListCache (fixes issue #3047). These methods
stopped being relevant after the removal of Flash support in January 2021.
The last remaining PPAPI plugin (PDF viewer) will switch to a non-plugin
implementation in the near future (see https://crbug.com/702993#c58) and
functionality related to plugin filtering has already been removed in
https://crrev.com/343ae351c9.
Fixes the following error:
Check failed: !pending_frame_callback_. Got overlapping IssueExternalBeginFrame
To test:
Run `cefclient --off-screen-rendering-enabled --external-begin-frame-enabled`
on Windows without crashing.
The `--disable-request-handling-for-testing` command-line flag would previously
disable handling of all schemes, including custom schemes and internal schemes
such as devtools.
This removes CefSettings.ignore_certificate_errors and
CefBrowserSettings.ignore_certificate_errors. Due to NetworkService
requirements these values must now be configured globally via the
"ignore-certificate-errors" command-line flag.
This removes CefBrowserSettings.universal_access_from_file_urls and
CefBrowserSettings.file_access_from_file_urls. Due to NetworkService
requirements these values must now be configured globally via command-line
flags ("allow-universal-access-from-files" and "allow-file-access-from-files"
respectively).
Also remove the kAllowFileAccessFromFileUrls switch in CEF which duplicates
the existing kAllowFileAccessFromFiles switch in Chromium (see issue #1785).
AppCache is deprecated in favor of Service Workers and support will be
fully removed soon (~M95). See https://web.dev/appcache-removal/.
Also add missing "allow-file-access-from-files" command-line switch for
CefBrowserSettings.file_access_from_file_urls.
This causes a race related to |notification_state_lock_| assignment when
GetMainFrame is called from multiple threads. GetMainFrame doesn't
trigger any notifications so it shouldn't need that lock. Instead, only
use NotificationStateLock on the UI thread.
When BackForwardCache is enabled and the user navigates the main frame
back/forward a new RFH may be created for an existing main frame GlobalId value
and CefFrameHostImpl (e.g. an object that was previously Detach()ed after main
frame navigation called SetMainFrame, but for which RenderFrameDeleted was not
subsequently called due to insertion in the BackForwardCache). In this case we
can re-attach the new RFH to the existing main frame CefFrameHostImpl in
RenderFrameHostStateChanged and resume processing of messages.
Swapping back/forward to an existing (already loaded) renderer does not trigger
new notifications for draggable regions (e.g. RenderFrameObserver::
DraggableRegionsChanged is not called by default). We therefore explicitly
request an update of draggable regions by sending the DidStopLoading message to
the renderer.
A new |reattached| parameter is added to CefFrameHandler::OnFrameAttached to
support identification of BackForwardCache usage by the client.
To test with unit tests:
Run `ceftests --gtest_filter=DraggableRegionsTest.DraggableRegionsCrossOrigin
--enable-features=BackForwardCache`
To test manually:
1. Run `cefclient --enable-features=BackForwardCache --use-views
--url=http://tests/draggable`, note that draggable regions work.
2. Load https://www.google.com via the address bar, note that draggable regions
are removed.
3. Go back to http://tests/draggable, note that draggable regions work.
4. Go forward to https://www.google.com, note that draggable regions are
removed.
BackForwardCache is currently being tested via field trials (see
https://crbug.com/1171298) and can be explicitly disabled using the
`--disable-back-forward-cache` or `--disable-features=BackForwardCache`
command-line flags. The default behavior now matches the Chrome runtime.
To test:
Run `cefclient.exe --use-views --hide-frame --hide-controls`
Add `--enable-chrome-runtime` for the same behavior using the Chrome location
bar instead of a text field.
With the introduction of prerendering in Chromium it is now possible for
RenderFrameHosts (RFH) to move between FrameTrees. As a consequence we can no
longer rely on FrameTreeNode IDs to uniquely identify a RFH over its lifespan.
We must now switch to using GlobalRenderFrameHostId (child_id, frame_routing_id)
instead for that purpose. Additionally, we simplify existing code by using the
GlobalRenderFrameHostId struct in all places that previously used a
(render_process_id, render_frame_id) pair, since these concepts are equivalent.
See https://crbug.com/1179502#c8 for additional background.
Marshall Greenblatt
Nick Mueller
Nicolas Dusart
Alex Maitland
Andy Tzeng
Mathieu Lafon
Shezan Baig