Putting this on the radar. INTERNET-DRAFT: Internationalized Uniform Resource Identifiers (IURI) http://www.ietf.org/internet-drafts/draft-masinter-url-i18n-05.txt See also W3C Internationalization - URIs and other identifiers http://www.w3.org/International/O-URL-and-ident.html See also bug 42898: iDNS support
what deos this bug mean ? Any test cases ?
What happens now if I type a non-ASCII URL in the location bar? What should we do? IE5 has a preference under Advanced tab of the menu Tools|Internet Options...: [ ] Always send URLS as UTF-8
why don't you try it ?
See bug 43852 "Send URLs as UTF-8" not working. Probably need to link these either with depend or dup.
mark it as future for now.
Internationalized Resource Identifiers (IRI), April 17, 2002 http://www.w3.org/International/2002/draft-w3c-i18n-iri-00.txt This draft replaces draft-masinter-url-i18n-08.txt.
Due to a series of misunderstandings, this draft was earlier announced as draft-w3c-i18n-iri-00.txt, and has also been published as draft-duerst-i18n-iri-00.txt. This draft replaces draft-masinter-url-i18n-08.txt. Please send comments to email@example.com. Updated URL: http://www.ietf.org/internet-drafts/draft-duerst-iri-00.txt
It is unclear to me what this bug is requesting.
IRIs are now last call: <http://www1.ietf.org/mail-archive/web/ietf-announce/current/msg00383.html> At may be the case that Mozilla already supports (most of) the functionality defined in that draft, but I haven't tested it thoroughly yet.
(In reply to comment #8) > It is unclear to me what this bug is requesting. Me too, so let's make it into a tracking bug.
Does Firefox transmit the path part of the IRI in UTF-8 as RFC 3987 suggests? It does not seem to. Try http://hapax.qc.ca/Recettes/Noël.html Firefox producest http://hapax.qc.ca/Recettes/No%EBl.html, which codes the path as Latin-1, while my server codes the resource as UTF-8 (http://hapax.qc.ca/Recettes/No%C3%ABl.html),,, If an option exists in Firefox to sent the path in UTF-8, where is it? I could not find it.
Additional info: this problem (sending the filename's path in Latin-1) happens when I key in "Noël.html" in the address field, even if the current page's encoding is UTF-8. It does not happen, when you click on a link like this one http://hapax.qc.ca/Recettes/Noël.html from within a page, it that case the path part is properly encoded in UTF-8.
To exemplify the problem (which occurs also within HTML pages) : I created two HTML pages pointing to http://hapax.qc.ca/Recettes/Noël.html. Both links look similar : <a href="http://hapax.qc.ca/Recettes/Noël.html">Noël is here</a> But as you will see the page written in Latin-1 will fail to reach http://hapax.qc.ca/Recettes/No%C3%ABl.html, but not the one in UTF-8. http://hapax.qc.ca/Recettes/LinktoNoelLatin-1.html http://hapax.qc.ca/Recettes/LinktoNoelUTF-8.html In other words, the current page's encoding determines the way the resource path is encoded, I believe this is very annoying, the server has no way of knowing what encoding was chosen, and this is what RFC 3987 averts (all paths are encoded in UTF-8 as a "broker" encoding). I would like RFC 3987 to be implemented at the very least as an option set by the user (preferably as a default but they may be some resistance here).
(In reply to comment #11) > If an option exists in Firefox to sent the path in UTF-8, where is it? I could > not find it. Set network.standard-url.encode-utf8 to "true" in about:config. See bug 284474 for why this is not the default.
Thank you very much, I was looking precisely for this. I wish there would be a less cryptic of setting this parameter But doesn't this contradict RFC 3987? Also one of the examples given to justify this default works perfectly (since then?) with UTF-8 encoded URLs (as well as the Latin-1). I left a comment to that effect. I will follow answers on bug 284474. Thanks again.
[my apology for the accidental wanted- request: not sure how that happened.]