ValueError: dictionary update sequence element #0 has length 1; 2 is required

[chadwhitacre]

https://app.getsentry.com/gittip/gittip/group/3352951/

Traceback (most recent call last):
  File "aspen/website.py", line 66, in handle_safely
    response = self.handle(request)
  File "aspen/website.py", line 99, in handle
    response = request.resource.respond(request)
  File "aspen/resources/dynamic_resource.py", line 57, in respond
    response = self.get_response(context)
  File "aspen/resources/negotiated_resource.py", line 86, in get_response
    media_type = mimeparse.best_match(self.available_types, accept)
  File "site-packages/mimeparse.py", line 121, in best_match
    parsed_header = [parse_media_range(r) for r in _filter_blank(header.split(","))]
  File "site-packages/mimeparse.py", line 54, in parse_media_range
    (type, subtype, params) = parse_mime_type(range)
  File "site-packages/mimeparse.py", line 33, in parse_mime_type
    for param in parts[1:] ])
ValueError: dictionary update sequence element #0 has length 1; 2 is required

[chadwhitacre]

I can't reproduce this. So far my hypothesis is that this is indicative of Heroku moving the ground out from beneath our feet.

[zbynekwinkler]

>>> import mimeparse
>>> mimeparse.best_match(['text/html'], 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;')

These are the params best_match is called with (taken from the sentry page traceback). I guess this is a bug in the mimeparse module. That function should not raise anything, right?

Googling for the problem reveals:
falconry/python-mimeparse#2
http://code.google.com/p/mimeparse/issues/detail?id=11

Unfortunately it is not fixed yet 😢

[zbynekwinkler]

Referencing #1521 to link them together.

[pjz]

Technically that Accept: header is malformed; the trailing semicolon is disallowed since there're no quality parameters after it.

[zbynekwinkler]

Can we fix this in aspen? Like guarding against exceptions raised by mimeparse.best_match. I agree that the header is malformed but lets follow http://en.wikipedia.org/wiki/Robustness_principle.

[chadwhitacre]

@zwn Let's ticket in Aspen if we haven't already.

[chadwhitacre]

Is this solved?

[pjz]

The aspen side was solved 3 months ago.

[chadwhitacre]

We last saw this 10 days ago, just before we landed the aspen 0.28 upgrade (#1699). Before that we were on 0.26.1, and the commit log on aspen-python shows that AspenWeb/pando.py#260 landed after 0.26.1 was released. The sentry error is already resolved. I'm closing this as fixed.