Glenn Linderman <v+pyt...@g.nevcal.com> added the comment: Pierre said: Option 1 is impossible, because the CGI script sometimes has no control on the stream : for instance on a shared web host, it will receive sys.stdin as a text stream
I say: It is the user code of the CGI script that calls CGI.FieldStorage. So the user could be required (option 1) to first tweak the stdin to be bytes, one way or another. I don't understand any circumstance where a Python CGI script doesn't have control over the settings of the Python IO Stack that it is using to obtain the data... and the CGI spec is defined as a bytestream, so it must be able to read the bytes. Victor said: It is possible to test the type of the stream. I say: Yes, why just assume (as I have been) that the initial precondition is the defaults that Python imposes. Other code could have interposed something else. The user should be allowed to pass in anything that is a TextIOWrapper, or a BytesIO, and CGI should be able to deal with it. If the user passes some other type, it should be assumed to produce bytes from its read() API, and if it doesn't the user gets what he deserves (an error). Since the default Python sys.stdin is a TextIOWrapper, having CGI detect that, and extract its .buffer to use for obtaining bytes, should work fine. If the user already tweaked sys.stdin to be a BytesIO (.buffer or detach()), CGI should detect and use that. If the user substitutes a different class, it should be bytes, and that should be documented, the three cases that could work. ---------- _______________________________________ 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
