Re: Charset for URL decoding (#19468)

2012-12-18 Thread Aymeric Augustin
2012/12/18 Łukasz Rekucki > > I wonder if UTF-8 with "surrogates escape" error mode makes sense here. > Python 3 uses it for decoding file-system paths, where it's not always > possible to determine the charset. I think it's pretty much the same case. > After all, the %-coded bytes can be some b

Re: Charset for URL decoding (#19468)

2012-12-18 Thread Aymeric Augustin
2012/12/18 Karen Tracey > https://code.djangoproject.com/ticket/5738 > > Comment #10 notes that utf-8 is what Django will use but with the last fix > noted against that ticket it is easier for the request class to be > subclassed to change things for an installation where a different charset > fo

Re: Charset for URL decoding (#19468)

2012-12-18 Thread Karen Tracey
On Tue, Dec 18, 2012 at 3:34 AM, Aymeric Augustin < aymeric.augus...@polytechnique.org> wrote: > > Currently, Django uses utf-8. As far as I can tell, that's more a > side-effect of (ab)using force_str than anything else. It also has the > drawback of making it impossible to serve perfectly legit

Re: Charset for URL decoding (#19468)

2012-12-18 Thread Łukasz Rekucki
On 18 December 2012 09:34, Aymeric Augustin < aymeric.augus...@polytechnique.org> wrote: > (complete version follows) > > Hello, > > I'm looking for some feedback on #19468 before making a decision. It's one > of the tickets that currently block the 1.5 release. > > Here's a summary of the problem

Charset for URL decoding (#19468)

2012-12-18 Thread Aymeric Augustin
(complete version follows) Hello, I'm looking for some feedback on #19468 before making a decision. It's one of the tickets that currently block the 1.5 release. Here's a summary of the problem. Django must decode environ['PATH_INFO'] to obtain request.path, where decoding means : 1 - URL-dec

Charset for URL decoding (#19468)

2012-12-17 Thread Aymeric Augustin
Hello, #19468 sparked an interesting debate; Claude and I would like some feedback before making a decision. Here's a summary of the problem. Django must decode environ['PATH_INFO'] to obtain request.path, where decoding means : 1 - URL-decoding to a bytestring 2 - "charset-decoding" to an unico