- From: Martin Nilsson <nilsson@opera.com>
- Date: Fri, 21 Mar 2014 15:43:32 +0100
- To: ietf-http-wg@w3.org
On Fri, 21 Mar 2014 14:54:24 +0100, Julian Reschke <julian.reschke@gmx.de> wrote: > On 2014-03-21 14:47, Nicolas Mailhot wrote: >> ... >> I'll give you a big secret: nobody writes in percent-escaped manually if >> he can avoid it, just like nobody uses html entities. >> >> The bulk of percent-escaped urls has been produced by automatons >> converting human-written plain text that used the document main >> encoding, >> so yes I do expect both encodings to match if the automaton was coded >> properly. >> ... > > That assumes that the "automaton" that did the URI-escaping actually > knew the document encoding, and that the document the URI appears in > never gets re-encoded. > And this is where it gets really depressing, since the encoding the server expects may be different from the documents encoding. For form data you can at least have a hidden variable with a known unicode sequence, but for URLs, which often are processed in an environment outside of the control of the target server, you have to use ASCII until the world realize that UTF-8 always everywhere is the solution. We can stop percent encoding too at that point. /Martin Nilsson -- Using Opera's revolutionary email client: http://www.opera.com/mail/
Received on Friday, 21 March 2014 14:44:05 UTC