[tracing] Fix single-process detection logic of ChildTraceMessageFilter

Currently the instantiation of ChildTraceMessageFilter in
child_thread_impl.cc is guarded by a check on
CommandLine.HasSwitch(kSingleProcess)
The rationale is that in single-process mode we don't want to have any
message filter as the browser-side tracing will cover all the threads.
However, the aforementioned check is not complete, as it does not
covers the case of tests using the InProcessUtilityThreadHelper [1].
In those tests, the kSingleProcess flag is not present, but the
renderer is in-process anyways.
Replacing the the HasSwitch check in with IsInBrowserProcess(),
which covers also the latter case.

[1] for instance extensions/browser/sandboxed_unpacker_unittest.cc

BUG=474973

Review URL: https://codereview.chromium.org/1069253003

Cr-Commit-Position: refs/heads/master@{#324276}

base/trace_event/memory_dump_manager.cc

@@ -77,10 +77,7 @@ void MemoryDumpManager::Initialize() {
 
 void MemoryDumpManager::SetDelegate(MemoryDumpManagerDelegate* delegate) {
   AutoLock lock(lock_);
-  // TODO(primiano): The DCHECK below is disabled just temporary (and it is
-  // strongly useful). It should be re-enabled soon as crbug.com/474973 is fixed
-  // (ETA: end of April 2015). Commenting out just to de-entangle CL deps.
-  // DCHECK(delegate_ == nullptr);
+  DCHECK_EQ(static_cast<MemoryDumpManagerDelegate*>(nullptr), delegate_);
   delegate_ = delegate;
 }
 

content/child/child_thread_impl.cc

@@ -401,8 +401,7 @@ void ChildThreadImpl::Init(const Options& options) {
   channel_->AddFilter(bluetooth_message_filter_->GetFilter());
   channel_->AddFilter(navigator_connect_dispatcher_->GetFilter());
 
-  if (!base::CommandLine::ForCurrentProcess()->HasSwitch(
-          switches::kSingleProcess)) {
+  if (!IsInBrowserProcess()) {
     // In single process mode, browser-side tracing will cover the whole
     // process including renderers.
     channel_->AddFilter(new tracing::ChildTraceMessageFilter(