Response wrongly keeps charset in Content-Type header

[wosc]

>>> response = webob.Response()
>>> response.content_type = 'image/jpeg'
>>> response.headers['Content-Type']
'image/jpeg; charset=UTF-8'

I guess the reason is that the content_type property preserves parameters if the new value does not contain any, so the charset is not removed from the header when setting a content type without parameters later on.

[dobesv]

Note that if you specify the content_type as a constructor parameter, there are some "smarts" in the constructor to set the charset (or not) based on the content_type provided. Setting the content_type at a later time does not, unfortunately, benefit from that same logic, so you have to clear the charset manually.

Perhaps every change to content_type could set a default charset in the same manner as in the constructor? Possibly retaining the current charset if there is one and the new content_type looks like a text content type?

[digitalresistor]

This was fixed in #261. Keeping any Content-Type parameters will be removed in WebOb 1.9.