Make TabCapture requests use the target render process and render view id's for UI permissions.

content/browser/renderer_host/media/web_contents_capture_util.cc

Index: content/browser/renderer_host/media/web_contents_capture_util.cc
diff --git a/content/browser/renderer_host/media/web_contents_capture_util.cc b/content/browser/renderer_host/media/web_contents_capture_util.cc
index 505364a3aba59ae3a4d0afa66fd4f008a6dcc86c..46b73766e0a7f43adddf7a47e2e61268fc21b3cc 100644
--- a/content/browser/renderer_host/media/web_contents_capture_util.cc
+++ b/content/browser/renderer_host/media/web_contents_capture_util.cc
@@ -38,7 +38,9 @@ bool WebContentsCaptureUtil::ExtractTabCaptureTarget(
     const std::string& device_id_param,
     int* render_process_id,
     int* render_view_id) {
-  DCHECK(IsWebContentsDeviceId(device_id_param));
+  if (!IsWebContentsDeviceId(device_id_param))
+    return false;
+
   const std::string device_id = device_id_param.substr(
       arraysize(kVirtualDeviceScheme) - 1);