electron/patches/chromium/short-circuit_permissions_checks_in_mediastreamdevicescontroller.patch
electron-roller[bot] 3b421ef77f
chore: bump chromium to 134.0.6998.3 (35-x-y) (#45460)
* chore: bump chromium in DEPS to 134.0.6998.1

* chore: bump chromium in DEPS to 134.0.6998.5

* chore: bump chromium in DEPS to 134.0.6998.3

* chore: bump chromium to 134.0.6988.0 (main) (#45334)

* chore: bump chromium in DEPS to 134.0.6976.0

* chore: update mas_avoid_private_macos_api_usage.patch.patch
https://chromium-review.googlesource.com/c/chromium/src/+/6171046
process_info_mac.cc -> process_info_mac.mm

* chore: update build_do_not_depend_on_packed_resource_integrity.patch
https://chromium-review.googlesource.com/c/chromium/src/+/6196857

* chore: update feat_add_support_for_missing_dialog_features_to_shell_dialogs.patch
https://chromium-review.googlesource.com/c/chromium/src/+/6182296
https://chromium-review.googlesource.com/c/chromium/src/+/6183404
https://chromium-review.googlesource.com/c/chromium/src/+/6187853

A lot changed in the upstream implementation. There's a good chance I got
this wrong as threading has changed and moved some variables into globals.

* chore: remove build_remove_vr_directx_helpers_dependency.patch
https://chromium-review.googlesource.com/c/chromium/src/+/6186102
This landed upstream

* chore: e patches all

* chore: update net::CookieInclusionStatus::ExclusionReason enum
https://chromium-review.googlesource.com/c/chromium/src/+/6183252
https://chromium-review.googlesource.com/c/chromium/src/+/6185544

* chore: update content::WebAuthenticationDelegate import
https://chromium-review.googlesource.com/c/chromium/src/+/6189769

* Revert "chore: disable focus handling test due to win32/ia32 regression"

This reverts commit 1a57ba5d59848d0c841ddda59c9299a4f957452a.

* chore: bump chromium in DEPS to 134.0.6978.0

* chore: bump chromium in DEPS to 134.0.6980.0

* chore: bump chromium in DEPS to 134.0.6982.0

* chore: bump chromium in DEPS to 134.0.6984.0

* 6196281: Allow direct embedder IsPdfInternalPluginAllowedOrigin() interaction
https://chromium-review.googlesource.com/c/chromium/src/+/6196281

* 6196283: Delete PdfInternalPluginDelegate
https://chromium-review.googlesource.com/c/chromium/src/+/6196283

* chore: update patches

* chore: bump chromium in DEPS to 134.0.6986.0

* chore: update patches

* 6205762: Support option to use window.showSaveFilePicker() in PDF attachment code
https://chromium-review.googlesource.com/c/chromium/src/+/6205762

See also:
* https://issues.chromium.org/issues/373852607
* 5939153: [PDF] Add PdfUseShowSaveFilePicker feature flag | https://chromium-review.googlesource.com/c/chromium/src/+/5939153
* 6205761: Delete spurious Ink-specific code in pdf_viewer.ts | https://chromium-review.googlesource.com/c/chromium/src/+/6205761

* 6209609: Remove WebVector: Automatic changes
https://chromium-review.googlesource.com/c/chromium/src/+/6209609

* 6205488: UI: make QT5 optional
https://chromium-review.googlesource.com/c/chromium/src/+/6205488

* 6178281: Rename pak files from branding strings
https://chromium-review.googlesource.com/c/chromium/src/+/6178281

* fixup! 6209609: Remove WebVector: Automatic changes https://chromium-review.googlesource.com/c/chromium/src/+/6209609

* 6193249: Switch from safe_browsing::EventResult to enterprise_connectors:EventResult
https://chromium-review.googlesource.com/c/chromium/src/+/6193249

* 6197457: Remove Pause/ResumeReadingBodyFromNet IPCs
https://chromium-review.googlesource.com/c/chromium/src/+/6197457

* 6191230: Record total time spent on a picture in picture window
https://chromium-review.googlesource.com/c/chromium/src/+/6191230

* chore: bump chromium in DEPS to 134.0.6988.0

* chore: update patches

* 6215440: Remove base/ranges/.
https://chromium-review.googlesource.com/c/chromium/src/+/6215440

* Disable unsafe buffers error

Not sure what changed, but we're now seeing unsafe buffer errors in Chromium code, at least when using reclient. Will update this comment if we find out the cause.

* 6187853: SelectFileDialogLinuxPortal: Use dbus_xdg::Request and DbusType
https://chromium-review.googlesource.com/c/chromium/src/+/6187853

* fix `setDisplayMediaRequestHandler` test

Given how this test is written, I would expect this assertion to be false. It seems the oppositue was true before, but that was also acknowledged to be suprising. Seems that the underlying implementation is now fixed and works as expected.

* fixup! 6187853: SelectFileDialogLinuxPortal: Use dbus_xdg::Request and DbusType https://chromium-review.googlesource.com/c/chromium/src/+/6187853

* chore: udpate patches

* Multiple PRS: https://chromium-review.googlesource.com/c/chromium/src/+/6185544 | https://chromium-review.googlesource.com/c/chromium/src/+/6183252

* fix: cast enum class to numeric type

* fix: add 1 to MAX_EXCLUSION_REASON because enum values are zero-based, and we want the total count of reasons.

* Reapply "chore: disable focus handling test due to win32/ia32 regression"

This reverts commit 760b1a519b5919b483c66bc3096eeefb4d7011f4.

* refactor: use ExclusionReasonBitset::kValueCount for size

---------

Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com>
Co-authored-by: Samuel Maddock <smaddock@slack-corp.com>
Co-authored-by: clavin <clavin@electronjs.org>
Co-authored-by: alice <alice@makenotion.com>
Co-authored-by: John Kleinschmidt <jkleinsc@electronjs.org>
(cherry picked from commit 213165a467b84b3fb979a869d9bf10ad21e2d78e)

---------

Co-authored-by: electron-roller[bot] <84116207+electron-roller[bot]@users.noreply.github.com>
2025-02-07 11:14:12 -05:00

104 lines
4.7 KiB
Diff

From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: Jeremy Rose <japthorp@slack-corp.com>
Date: Tue, 12 Jul 2022 16:51:43 -0700
Subject: short-circuit permissions checks in MediaStreamDevicesController
The //components/permissions architecture is complicated and not that
widely used in Chromium, and mostly oriented around showing permissions
UI and/or remembering per-site permissions, which we're not interested
in.
Since we do a permissions check prior to invoking the
MediaStreamDevicesController, and don't (yet) provide the ability to set
granular permissions (e.g. allow video but not audio), just
short-circuit all the permissions checks in MSDC for now to allow us to
unduplicate this code.
diff --git a/components/webrtc/media_stream_devices_controller.cc b/components/webrtc/media_stream_devices_controller.cc
index 13929e4f6c1997a9071a2d1e3982a8c8feac84e4..a9719f72b9994537665c7f6a433ea8499b42122d 100644
--- a/components/webrtc/media_stream_devices_controller.cc
+++ b/components/webrtc/media_stream_devices_controller.cc
@@ -57,7 +57,8 @@ bool PermissionIsRequested(blink::PermissionType permission,
void MediaStreamDevicesController::RequestPermissions(
const content::MediaStreamRequest& request,
MediaStreamDeviceEnumerator* enumerator,
- ResultCallback callback) {
+ ResultCallback callback,
+ bool previously_approved) {
content::RenderFrameHost* rfh = content::RenderFrameHost::FromID(
request.render_process_id, request.render_frame_id);
// The RFH may have been destroyed by the time the request is processed.
@@ -92,7 +93,7 @@ void MediaStreamDevicesController::RequestPermissions(
std::move(callback)));
std::vector<blink::PermissionType> permission_types;
-
+#if 0
content::PermissionController* permission_controller =
web_contents->GetBrowserContext()->GetPermissionController();
@@ -165,19 +166,26 @@ void MediaStreamDevicesController::RequestPermissions(
requested_audio_capture_device_ids;
permission_request_description.requested_video_capture_device_ids =
requested_video_capture_device_ids;
-
+#endif
// It is OK to ignore `request.security_origin` because it will be calculated
// from `render_frame_host` and we always ignore `requesting_origin` for
// `AUDIO_CAPTURE` and `VIDEO_CAPTURE`.
// `render_frame_host->GetMainFrame()->GetLastCommittedOrigin()` will be used
// instead.
- rfh->GetBrowserContext()
- ->GetPermissionController()
- ->RequestPermissionsFromCurrentDocument(
- rfh, permission_request_description,
- base::BindOnce(
- &MediaStreamDevicesController::PromptAnsweredGroupedRequest,
- std::move(controller)));
+ if (previously_approved) {
+ controller->PromptAnsweredGroupedRequest({blink::mojom::PermissionStatus::GRANTED /*audio*/,
+ blink::mojom::PermissionStatus::GRANTED /*video*/});
+ } else {
+ rfh->GetBrowserContext()
+ ->GetPermissionController()
+ ->RequestPermissionsFromCurrentDocument(
+ rfh,
+ content::PermissionRequestDescription(permission_types,
+ request.user_gesture),
+ base::BindOnce(
+ &MediaStreamDevicesController::PromptAnsweredGroupedRequest,
+ std::move(controller)));
+ }
}
MediaStreamDevicesController::~MediaStreamDevicesController() {
@@ -426,6 +434,7 @@ bool MediaStreamDevicesController::PermissionIsBlockedForReason(
return false;
}
+#if 0
// TODO(raymes): This function wouldn't be needed if
// PermissionManager::RequestPermissions returned a denial reason.
content::PermissionResult result =
@@ -436,6 +445,7 @@ bool MediaStreamDevicesController::PermissionIsBlockedForReason(
DCHECK_EQ(blink::mojom::PermissionStatus::DENIED, result.status);
return true;
}
+#endif
return false;
}
diff --git a/components/webrtc/media_stream_devices_controller.h b/components/webrtc/media_stream_devices_controller.h
index 4d3bf184243a4c1296f41070e934730a3bd2620e..78610ff9fb93cfa1c9aae898d6f8c83520fc860a 100644
--- a/components/webrtc/media_stream_devices_controller.h
+++ b/components/webrtc/media_stream_devices_controller.h
@@ -48,7 +48,8 @@ class MediaStreamDevicesController {
// synchronously or asynchronously returned via |callback|.
static void RequestPermissions(const content::MediaStreamRequest& request,
MediaStreamDeviceEnumerator* enumerator,
- ResultCallback callback);
+ ResultCallback callback,
+ bool previously_approved = false);
~MediaStreamDevicesController();