This object represents a WSGI request with a more friendly interface. This does not expose every detail of the WSGI environment, and attempts to express nothing beyond what is available in the environment dictionary.

The only state maintained in this object is the desired charset, its associated errors handler, and the decode_param_names option.

The incoming parameter values will be automatically coerced to unicode objects of the charset encoding when charset is set. The incoming parameter names are not decoded to unicode unless the decode_param_names option is enabled.

When unicode is expected, charset will overridden by the the value of the Content-Type header's charset parameter if one was specified by the client.

The class variable defaults specifies default values for charset, errors, and langauge. These can be overridden for the current request via the registry.

The language default value is considered the fallback during i18n translations to ensure in odd cases that mixed languages don't occur should the language file contain the string but not another language in the accepted languages list. The language value only applies when getting a list of accepted languages from the HTTP Accept header.

This behavior is duplicated from Aquarium, and may seem strange but is very useful. Normally, everything in the code is in "en-us". However, the "en-us" translation catalog is usually empty. If the user requests ["en-us", "zh-cn"] and a translation isn't found for a string in "en-us", you don't want gettext to fallback to "zh-cn". You want it to just use the string itself. Hence, if a string isn't found in the language catalog, the string in the source code will be used.

All other state is kept in the environment dictionary; this is essential for interoperability.

You are free to subclass this object.