To fix a bug that causes retry fail.

Often we get 500 internal server error from S3 and HTTPConnection will retry with a file pointer pointing to EOF, which leads to 400 Bad Request -- "RequestTimeout"

git-svn-id: https://wush.net/svn/rightscale/right_http_connection/trunk@3889 9f0cbaf6-ce18-0410-ad37-d14a22affa91

lib/right_http_connection.rb

@@ -331,6 +331,11 @@ def request(request_params, &block)
             request.content_length = body.respond_to?(:lstat) ? body.lstat.size : body.size 
             request.body_stream = request.body
           end
+
+          # Always make sure that the file pointer is set to point to the beginning of the file
+          request.body.seek(0,IO::SEEK_SET) if request.body
+          request.body_stream.seek(0,IO::SEEK_SET) if request.body_stream
+
           response = @http.request(request, &block)
 
           error_reset