Problems getting the Heroku example app for Python run locally with SSL

Question

I started out with having the same problem as in this question. As one of the answers suggests one can avoid that particular problem by making the app run without SSL. But as Facebook is set to enforce https for apps in just a few days (October 1 2011) that seems like a solution that won't last. I first tried to enable ssl in app.run (around line 149 in exampleapp.py. Like so:

app.run(host='0.0.0.0', port=port, ssl_context='adhoc')

At first try it failed at start complaining about the missing OpenSSL module. Found a couple of suggestions on how to fix that out there on the net, and choose to do:

(myapp)$ pip install pyopenssl

Now no complaints at start:

(myapp)$ foreman start
10:35:25 web.1     | started with pid 26934
10:35:26 web.1     |  * Running on https://0.0.0.0:5000/
10:35:26 web.1     |  * Restarting with reloader

But then when trying to access the app:

10:35:31 web.1     | ----------------------------------------
10:35:31 web.1     | Exception happened during processing of request from ('127.0.0.1', 61118)
10:35:31 web.1     | Traceback (most recent call last):
10:35:31 web.1     |   File "/usr/local/Cellar/python/2.7.1/lib/python2.7/SocketServer.py", line 284, in _handle_request_noblock
10:35:31 web.1     |     self.process_request(request, client_address)
10:35:31 web.1     |   File "/usr/local/Cellar/python/2.7.1/lib/python2.7/SocketServer.py", line 310, in process_request
10:35:31 web.1     |     self.finish_request(request, client_address)
10:35:31 web.1     |   File "/usr/local/Cellar/python/2.7.1/lib/python2.7/SocketServer.py", line 323, in finish_request
10:35:31 web.1     |     self.RequestHandlerClass(request, client_address, self)
10:35:31 web.1     |   File "/usr/local/Cellar/python/2.7.1/lib/python2.7/SocketServer.py", line 639, in __init__
10:35:31 web.1     |     self.handle()
10:35:31 web.1     |   File "path_to_myapp/lib/python2.7/site-packages/werkzeug/serving.py", line 189, in handle
10:35:31 web.1     |     return rv
10:35:31 web.1     | UnboundLocalError: local variable 'rv' referenced before assignment
10:35:31 web.1     | ----------------------------------------
10:35:31 web.1     | Unhandled exception in thread started by <function inner at 0x10139e050>
10:35:31 web.1     | Traceback (most recent call last):
10:35:31 web.1     |   File "path_to_myapp/lib/python2.7/site-packages/werkzeug/serving.py", line 599, in inner
10:35:31 web.1     |     passthrough_errors, ssl_context).serve_forever()
10:35:31 web.1     |   File "path_to_myapp/lib/python2.7/site-packages/werkzeug/serving.py", line 355, in serve_forever
10:35:31 web.1     |     HTTPServer.serve_forever(self)
10:35:31 web.1     |   File "/usr/local/Cellar/python/2.7.1/lib/python2.7/SocketServer.py", line 227, in serve_forever
10:35:31 web.1     |     self._handle_request_noblock()
10:35:31 web.1     |   File "/usr/local/Cellar/python/2.7.1/lib/python2.7/SocketServer.py", line 287, in _handle_request_noblock
10:35:31 web.1     |     self.shutdown_request(request)
10:35:31 web.1     |   File "/usr/local/Cellar/python/2.7.1/lib/python2.7/SocketServer.py", line 459, in shutdown_request
10:35:31 web.1     |     request.shutdown(socket.SHUT_WR)
10:35:31 web.1     | TypeError: shutdown() takes exactly 0 arguments (1 given)

What to do about this? Is it the wrong Python version or just me missunderstanding some other fundamental thing?

Answer 1

It is a bug in Werkzeug, that got fixed.

Change the Werkzeug version in the requirements.txt file to at least 0.8.2 and run pip install -r requirements.txt to upgrade.