Reland "Add OnReady callback to the speech recognition client"

Patchset 1 contains the reland without any changes.

This CL adds a callback to the speech recognition client that is called
when all of the mojo pipes are bound, indicating that the speech
recognition service is available.

Bug: 1096632
Change-Id: I0006298c3d3191c84c80d920031abfcbf10e5cf6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2260935Reviewed-by: Dale Curtis <dalecurtis@chromium.org>
Reviewed-by: Scott Violet <sky@chromium.org>
Commit-Queue: Evan Liu <evliu@google.com>
Cr-Commit-Position: refs/heads/master@{#781504}

chrome/renderer/chrome_content_renderer_client.cc

@@ -1439,8 +1439,10 @@ ChromeContentRendererClient::CreateWorkerContentSettingsClient(
 #if !defined(OS_ANDROID)
 std::unique_ptr<media::SpeechRecognitionClient>
 ChromeContentRendererClient::CreateSpeechRecognitionClient(
-    content::RenderFrame* render_frame) {
-  return std::make_unique<ChromeSpeechRecognitionClient>(render_frame);
+    content::RenderFrame* render_frame,
+    media::SpeechRecognitionClient::OnReadyCallback callback) {
+  return std::make_unique<ChromeSpeechRecognitionClient>(render_frame,
+                                                         std::move(callback));
 }
 #endif
 

chrome/renderer/chrome_content_renderer_client.h

@@ -151,7 +151,8 @@ class ChromeContentRendererClient
       content::RenderFrame* render_frame) override;
 #if !defined(OS_ANDROID)
   std::unique_ptr<media::SpeechRecognitionClient> CreateSpeechRecognitionClient(
-      content::RenderFrame* render_frame) override;
+      content::RenderFrame* render_frame,
+      media::SpeechRecognitionClient::OnReadyCallback callback) override;
 #endif
   void AddSupportedKeySystems(
       std::vector<std::unique_ptr<::media::KeySystemProperties>>* key_systems)

chrome/renderer/media/chrome_speech_recognition_client.cc

@@ -26,8 +26,9 @@ std::vector<std::string> GetBlockedURLs() {
 }
 
 ChromeSpeechRecognitionClient::ChromeSpeechRecognitionClient(
-    content::RenderFrame* render_frame)
-    : blocked_urls_(GetBlockedURLs()) {
+    content::RenderFrame* render_frame,
+    media::SpeechRecognitionClient::OnReadyCallback callback)
+    : on_ready_callback_(std::move(callback)), blocked_urls_(GetBlockedURLs()) {
   mojo::PendingReceiver<media::mojom::SpeechRecognitionContext>
       speech_recognition_context_receiver =
           speech_recognition_context_.BindNewPipeAndPassReceiver();
@@ -36,6 +37,7 @@ ChromeSpeechRecognitionClient::ChromeSpeechRecognitionClient(
       speech_recognition_client_receiver_.BindNewPipeAndPassRemote(),
       base::BindOnce(&ChromeSpeechRecognitionClient::OnRecognizerBound,
                      base::Unretained(this)));
+
   render_frame->GetBrowserInterfaceBroker()->GetInterface(
       std::move(speech_recognition_context_receiver));
   render_frame->GetBrowserInterfaceBroker()->GetInterface(
@@ -47,6 +49,10 @@ ChromeSpeechRecognitionClient::ChromeSpeechRecognitionClient(
 void ChromeSpeechRecognitionClient::OnRecognizerBound(
     bool is_multichannel_supported) {
   is_multichannel_supported_ = is_multichannel_supported;
+  is_recognizer_bound_ = true;
+
+  if (on_ready_callback_)
+    std::move(on_ready_callback_).Run();
 }
 
 ChromeSpeechRecognitionClient::~ChromeSpeechRecognitionClient() = default;
@@ -61,9 +67,9 @@ void ChromeSpeechRecognitionClient::AddAudio(
 }
 
 bool ChromeSpeechRecognitionClient::IsSpeechRecognitionAvailable() {
+  // TODO(evliu): Check if SODA is available.
   return !is_website_blocked_ && is_browser_requesting_transcription_ &&
-         speech_recognition_recognizer_.is_bound() &&
-         speech_recognition_recognizer_.is_connected();
+         is_recognizer_bound_ && speech_recognition_recognizer_.is_connected();
 }
 
 void ChromeSpeechRecognitionClient::OnSpeechRecognitionRecognitionEvent(

chrome/renderer/media/chrome_speech_recognition_client.h

@@ -28,7 +28,9 @@ class ChromeSpeechRecognitionClient
     : public media::SpeechRecognitionClient,
       public media::mojom::SpeechRecognitionRecognizerClient {
  public:
-  explicit ChromeSpeechRecognitionClient(content::RenderFrame* render_frame);
+  explicit ChromeSpeechRecognitionClient(
+      content::RenderFrame* render_frame,
+      media::SpeechRecognitionClient::OnReadyCallback callback);
   ChromeSpeechRecognitionClient(const ChromeSpeechRecognitionClient&) = delete;
   ChromeSpeechRecognitionClient& operator=(
       const ChromeSpeechRecognitionClient&) = delete;
@@ -61,6 +63,8 @@ class ChromeSpeechRecognitionClient
   void ResetChannelMixer(const media::AudioBuffer& buffer);
   bool IsUrlBlocked(const std::string& url) const;
 
+  media::SpeechRecognitionClient::OnReadyCallback on_ready_callback_;
+
   mojo::Remote<media::mojom::SpeechRecognitionContext>
       speech_recognition_context_;
   mojo::Remote<media::mojom::SpeechRecognitionRecognizer>
@@ -78,6 +82,9 @@ class ChromeSpeechRecognitionClient
 
   // Whether the browser is still requesting transcriptions.
   bool is_browser_requesting_transcription_ = true;
+
+  bool is_recognizer_bound_ = false;
+
   // The temporary audio bus used to mix multichannel audio into a single
   // channel.
   std::unique_ptr<media::AudioBus> monaural_audio_bus_;

content/public/renderer/content_renderer_client.cc

@@ -189,7 +189,8 @@ ContentRendererClient::CreateWorkerContentSettingsClient(
 #if !defined(OS_ANDROID)
 std::unique_ptr<media::SpeechRecognitionClient>
 ContentRendererClient::CreateSpeechRecognitionClient(
-    RenderFrame* render_frame) {
+    RenderFrame* render_frame,
+    media::SpeechRecognitionClient::OnReadyCallback callback) {
   return nullptr;
 }
 #endif

content/public/renderer/content_renderer_client.h

@@ -305,7 +305,9 @@ class CONTENT_EXPORT ContentRendererClient {
 #if !defined(OS_ANDROID)
   // Creates a speech recognition client used to transcribe audio into captions.
   virtual std::unique_ptr<media::SpeechRecognitionClient>
-  CreateSpeechRecognitionClient(RenderFrame* render_frame);
+  CreateSpeechRecognitionClient(
+      RenderFrame* render_frame,
+      media::SpeechRecognitionClient::OnReadyCallback callback);
 #endif
 
   // Returns true if the page at |url| can use Pepper CameraDevice APIs.

content/renderer/media/media_factory.cc

@@ -181,7 +181,7 @@ std::unique_ptr<media::DefaultRendererFactory> CreateDefaultRendererFactory(
       media_log, decoder_factory,
       base::BindRepeating(&content::RenderThreadImpl::GetGpuFactories,
                           base::Unretained(render_thread)),
-      render_frame->CreateSpeechRecognitionClient());
+      render_frame->CreateSpeechRecognitionClient(base::OnceClosure()));
 #endif
   return default_factory;
 }

content/renderer/render_frame_impl.cc

@@ -3676,10 +3676,12 @@ RenderFrameImpl::CreateWorkerContentSettingsClient() {
 
 #if !defined(OS_ANDROID)
 std::unique_ptr<media::SpeechRecognitionClient>
-RenderFrameImpl::CreateSpeechRecognitionClient() {
+RenderFrameImpl::CreateSpeechRecognitionClient(
+    media::SpeechRecognitionClient::OnReadyCallback callback) {
   if (!frame_ || !frame_->View())
     return nullptr;
-  return GetContentClient()->renderer()->CreateSpeechRecognitionClient(this);
+  return GetContentClient()->renderer()->CreateSpeechRecognitionClient(
+      this, std::move(callback));
 }
 #endif
 

content/renderer/render_frame_impl.h

@@ -57,6 +57,7 @@
 #include "ipc/ipc_message.h"
 #include "ipc/ipc_platform_file.h"
 #include "media/base/routing_token_callback.h"
+#include "media/base/speech_recognition_client.h"
 #include "mojo/public/cpp/bindings/associated_receiver.h"
 #include "mojo/public/cpp/bindings/associated_remote.h"
 #include "mojo/public/cpp/bindings/pending_associated_receiver.h"
@@ -136,9 +137,6 @@ class Range;
 
 namespace media {
 class MediaPermission;
-#if !defined(OS_ANDROID)
-class SpeechRecognitionClient;
-#endif
 }
 
 namespace service_manager {
@@ -615,8 +613,8 @@ class CONTENT_EXPORT RenderFrameImpl
   std::unique_ptr<blink::WebContentSettingsClient>
   CreateWorkerContentSettingsClient() override;
 #if !defined(OS_ANDROID)
-  std::unique_ptr<media::SpeechRecognitionClient>
-  CreateSpeechRecognitionClient();
+  std::unique_ptr<media::SpeechRecognitionClient> CreateSpeechRecognitionClient(
+      media::SpeechRecognitionClient::OnReadyCallback callback);
 #endif
   scoped_refptr<blink::WebWorkerFetchContext> CreateWorkerFetchContext()
       override;

media/base/speech_recognition_client.h

@@ -7,6 +7,7 @@
 
 #include <memory>
 
+#include "base/callback.h"
 #include "media/base/audio_buffer.h"
 #include "media/base/media_export.h"
 
@@ -16,6 +17,8 @@ namespace media {
 // captions.
 class MEDIA_EXPORT SpeechRecognitionClient {
  public:
+  using OnReadyCallback = base::OnceCallback<void()>;
+
   virtual ~SpeechRecognitionClient() = default;
 
   virtual void AddAudio(scoped_refptr<AudioBuffer> buffer) = 0;