- From: Robin Berjon <robin@knowscape.com>
- Date: Wed, 21 Mar 2001 11:40:49 +0100
- To: Philippe Le Hegaret <plh@w3.org>
- Cc: www-style@w3.org
At 16:51 20/03/2001 -0500, Philippe Le Hegaret wrote: >Robin Berjon wrote: >> Is there any reason that I haven't seen that justifies the absence of a >> charset(charset) callback ? > >startDocument receives an InputSource and >InputSource.getEncoding() > Get the character encoding for a byte stream or URI. > >I guess a charset callback could be useful but keep in mind that it might not >always represent the real encoding of the CSS file, depending on the >charset parameter on the text/css mimetype (actually, the rfc2318 doesn't say >anything about that. might it should). Yes I am aware of the limitations of @charset, and of the fact that it can't be used reliably to know the actual encoding, but my idea was that if it's there in the original source, there should be a way for the application to know about that. Besides, implementing an equivalent of InputSource in Perl would be painful, and probably pretty much useless (right now). Next version will bring us something called PerlIO, at which point it'll be much easier, and will make much more sense. Until then, relying on Perl's dwimity with byte streams is probably the best thing to do. Thanks :) -- robin b. To err is human, to purr feline.
Received on Wednesday, 21 March 2001 05:42:23 UTC