Fix assigning NULL to scoped_refptr

Update code where NULL or 0 is assigned to a scoped_refptr to reset it. Just
call reset() instead. If operator=(std::nullptr_t) is added these assignments
become ambiguous.

This CL was uploaded by git cl split.

R=mseaborn@chromium.org

Bug: 1024981
Change-Id: I2d6317c780f849bb47e943d48868fe24283120f6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1918157
Auto-Submit: kylechar <kylechar@chromium.org>
Reviewed-by: Mark Seaborn <mseaborn@chromium.org>
Commit-Queue: Mark Seaborn <mseaborn@chromium.org>
Cr-Commit-Position: refs/heads/master@{#715489}

components/nacl/browser/nacl_process_host.cc

@@ -652,7 +652,7 @@ void NaClProcessHost::SendMessageToRenderer(
 
   NaClHostMsg_LaunchNaCl::WriteReplyParams(reply_msg_, result, error_message);
   nacl_host_message_filter_->Send(reply_msg_);
-  nacl_host_message_filter_ = NULL;
+  nacl_host_message_filter_.reset();
   reply_msg_ = NULL;
 }
 

components/nacl/loader/nacl_ipc_adapter_unittest.cc

@@ -39,7 +39,7 @@ class NaClIPCAdapterTest : public testing::Test {
   }
   void TearDown() override {
     sink_ = NULL;  // This pointer is actually owned by the IPCAdapter.
-    adapter_ = NULL;
+    adapter_.reset();
     // The adapter destructor has to post a task to destroy the Channel on the
     // IO thread. For the purposes of the test, we just need to make sure that
     // task gets run, or it will appear as a leak.