Add some suggested performance improvements to geturl.

examples/geturl/geturl_handler.h

Index: examples/geturl/geturl_handler.h
===================================================================
--- examples/geturl/geturl_handler.h	(revision 1387)
+++ examples/geturl/geturl_handler.h	(working copy)
@@ -11,12 +11,20 @@
 #include "ppapi/cpp/url_request_info.h"
 #include "ppapi/cpp/instance.h"
 
-#define READ_BUFFER_SIZE 4096
+#define READ_BUFFER_SIZE 32768
 
 // GetURLHandler is used to download data from |url|. When download is
 // finished or when an error occurs, it posts a message back to the browser
 // with the results encoded in the message as a string and self-destroys.
 //
+// pp::URLLoader.GetDownloadProgress() is used to to allocate the memory
+// required for url_response_body_ before the download starts.  (This is not so
+// much of a performance improvement, but it saves some memory since
+// std::string.insert() typically grows the string's capacity by somewhere
+// between 50% to 100% when it needs more memory, depending on the
+// implementation.)  Other performance improvements made as outlined in this
+// bug: http://code.google.com/p/chromium/issues/detail?id=103947
+//
 // EXAMPLE USAGE:
 // GetURLHandler* handler* = GetURLHandler::Create(instance,url);
 // handler->Start();
@@ -69,7 +77,7 @@
   std::string url_;  // URL to be downloaded.
   pp::URLRequestInfo url_request_;
   pp::URLLoader url_loader_;  // URLLoader provides an API to download URLs.
-  char buffer_[READ_BUFFER_SIZE];  // Temporary buffer for reads.
+  char* buffer_;  // Temporary buffer for reads.
   std::string url_response_body_;  // Contains accumulated downloaded data.
   pp::CompletionCallbackFactory<GetURLHandler> cc_factory_;