Senthil Kumaran added the comment:
The patch looks good, but I have doubt in the claim that, this new behavior is
actually a right thing to do.
RDM wrote:
Now, that said, it seems to me that while it is not (apparently) RFC compliant,
query strings do have a natural way to support null values: a name without a
value. It would therefore be convenient if urlencode supported null values by
turning something like {'josh': None, 'fred': 'abc'} into:
josh&fred=abc
It is correctly recognized that it is not RFC Compliant. A reference to this
behavior exhibited by the software may be helpful. The
`application/x-www-form-urlencoded` type always looks for key=value kind of
query string only. And changes submitted by patch will break for folks who
think "a=None" is actually what they expected when the sent {a:None}.
I think, some present instances where this behavior is exhibited will be good
to note.
----------
_______________________________________
Python tracker <rep...@bugs.python.org>
<http://bugs.python.org/issue18857>
_______________________________________
_______________________________________________
Python-bugs-list mailing list
Unsubscribe:
https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
