gunicorn gevent error sending certain HTML strings

Question

Gunicorn behinx an nginx proxy on an Ubuntu 12.04.

I have the following function in a bare bones gunicorn server for sending an html string:

def send_html(start_response, replystr):
    try:
        status = '200 OK'
        response_headers = [('Content-type', 'text/html; charset=UTF-8'),
                            ('Content-Length', str(len(replystr)))]
        start_response(status, response_headers)
        return iter([replystr])
    except:
        log.info(traceback.print_exc())
        send_error(start_response, '501')
        return []

If I execute send_html(start_response, 'Hello world') it works just fine. However when I try to send a sizable string containing a news article's content, I seem to get the following error every time:

Traceback (most recent call last):
  File "build/bdist.linux-x86_64/egg/gevent/pywsgi.py", line 504, in handle_one_response
    self.run_application()
  File "build/bdist.linux-x86_64/egg/gevent/pywsgi.py", line 491, in run_application
    self.process_result()
  File "build/bdist.linux-x86_64/egg/gevent/pywsgi.py", line 482, in process_result
    self.write(data)
  File "build/bdist.linux-x86_64/egg/gevent/pywsgi.py", line 375, in write
    self._write_with_headers(data)
  File "build/bdist.linux-x86_64/egg/gevent/pywsgi.py", line 394, in _write_with_headers
    towrite.extend(data)
TypeError: an integer or string of size 1 is required

The gevent pywsgi source code tells me towrite is a byte array:

    def _write_with_headers(self, data):
        towrite = bytearray()
        self.headers_sent = True
        self.finalize_headers()

        towrite.extend('%s %s\r\n' % (self.request_version, self.status))
        for header in self.response_headers:
            towrite.extend('%s: %s\r\n' % header)

        towrite.extend('\r\n')
        if data:
            if self.response_use_chunked:
                ## Write the chunked encoding
                towrite.extend("%x\r\n%s\r\n" % (len(data), data))
            else:
                towrite.extend(data)
        self._sendall(towrite)

Does the article content not play nice with a byte array or something?

Answer 1

I have modified the send_html function to suit my purposes by turning the replystr into a byte array before sending it. Not sure how sustainable this solution is, but it has been working for a while now.

def send_html(start_response, replystr):
    try:
        status = '200 OK'
        #coerce the string to a bytearray
        replystr = bytearray(replystr, 'UTF-8') 
        response_headers = [('Content-type', 'text/html; charset=UTF-8'),
                            ('Content-Length', str(len(replystr)))]
        start_response(status, response_headers)
        return iter([replystr])
    except:
        log.info(traceback.print_exc())
        send_error(start_response, '501')
        return []

Answer 2

doing towrite.extend(str("%x\r\n%s\r\n" % (len(data), data))) will solve the issue.

Reason being 'extend' will not accept unicode.