Correct examples using string.getBytes() without a charset

[andreho]

Hi,
there is a problem in your Java examples, that obtain an encoded content of a string as a byte-array via java.lang.String.getBytes():byte[] method. This may lead to a poison message, which makes your whole queue not readable and results in an error (bad_utf8_character_code) looking like this example:

=ERROR REPORT==== 7-May-2015::15:59:37 ===
** Handler sockjs_cowboy_handler terminating in websocket_info/3
   for the reason exit:{ucs,{bad_utf8_character_code}}
** Message was go
** Options were {service,"/stomp",#Fun<rabbit_ws_sockjs.1.120271989>,{},
                         "http://cdn.sockjs.org/sockjs-0.2.js",false,true,
                         5000,25000,131072,#Fun<rabbit_ws_sockjs.0.116551930>,
                         undefined}

.... and so on.

The call of string.getBytes() method uses by default "Charset.defaultCharset()" or "ISO-8859-1" which could produce an incompatible byte-sequence and make it impossible to retrive a queued message.

We would advice to change all calls of string.getBytes() to string.getBytes("UTF-8") in existing examples to force usage of a compatible string encoding.

How to reproduce: publish a message with the following content "Läuft alles?" into a queue using your examples and then try to retrieve this message.

Debian, RabbitMQ 3.5.1, SockJS + Stomp
Greetz, Andre.

[michaelklishin]

@andreho feel free to submit a pull request.

[andreho]

Thx ;)