Android: Remove some extra CHECKs()

content/browser/child_process_launcher.cc

Index: content/browser/child_process_launcher.cc
diff --git a/content/browser/child_process_launcher.cc b/content/browser/child_process_launcher.cc
index 5f708b87922e918ab1a75b6433e8b12fc6a1106e..e284de561fbf5358b194fcd8e3aaa5dba91ef226 100644
--- a/content/browser/child_process_launcher.cc
+++ b/content/browser/child_process_launcher.cc
@@ -184,12 +184,6 @@ void ChildProcessLauncher::Context::Launch(
   // We currently only support renderer and gpu child processes.
   std::string process_type =
       cmd_line->GetSwitchValueASCII(switches::kProcessType);
-  CHECK_NE(switches::kPluginProcess, process_type);
-  CHECK_NE(switches::kPpapiBrokerProcess, process_type);
-  CHECK_NE(switches::kPpapiPluginProcess, process_type);
-  CHECK_NE(switches::kSandboxIPCProcess, process_type);
-  CHECK_NE(switches::kUtilityProcess, process_type);
-  CHECK_NE(switches::kZygoteProcess, process_type);

[Feng Qian, 2014/12/12 18:29:42]

Should we log the unknown process type on Java-side instead of 'assert false'
there? If that happens, we will know what new process type is.

[no sievers, 2014/12/12 18:50:44]

If we want to log we can do that here as well, i.e. 
  CHECK(process_type == switches::kGpuProcess ||    process_type ==
switches::kRendererProcess) << "Unsupported process type" << process_type;

Also note that the call stack here has the original process launch site.The
latter task that has the java assert() is posted as a separate task.

[Feng Qian, 2014/12/12 19:07:31]

Sounds good to me, can you add the process type to log here?

   CHECK(process_type == switches::kGpuProcess ||
         process_type == switches::kRendererProcess);