R. David Murray <rdmur...@bitdance.com> added the comment: Yeah, the documentation for the email stuff is in the dev docs. There's a short summary in the changes section of the email intro with links to the classes and methods that are affected. But basically you call BinaryFeedParser and feed it a binary data, and everything else works just like it did before, including the fact that get_payload() with no arguments returns a string. If there is non-ASCII data in that string and no charset was specified the binary data will get trashed though. To get the binary data out you call it with decode=True.
I believe you are right that the io module does not support intermixing calls to the main object and its buffer attribute; that's why detach was introduced, I believe. Antoine is nosy on this issue now; he can correct me if I'm wrong. So unfortunately I think we do need to come at this starting from binary at the beginning and *decoding* as needed (I believe http uses latin-1 when no charset is specified, but I need to double check that). That still leaves the problem of what if anything to do about existing programs that expect every value in a FieldStorage to be a string. Introduce a new method or parameter for getting the binary version of the value, possibly with some flag indicating that parsing detected non-ascii data? ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue4953> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
