- From: Martin Duerst <duerst@it.aoyama.ac.jp>
- Date: Wed, 28 Mar 2007 13:50:07 +0900
- To: Najib Tounsi <ntounsi@emi.ac.ma>
- Cc: Richard Ishida <ishida@w3.org>, www-international@w3.org
Hi Najib, At 02:25 07/03/28, Najib Tounsi wrote: >Martin Duerst wrote: >> Does such a mixture include labels that contain both Latin and RTL >> characters, or are the scripts separated by dots? In the former >> case, this would be even more peculiar, because such labels >> (mixing RTL and LTR characters) are illegal in IDN. >> >It is one label that is a mixture of RTL end LTR chars. It is invalid of course. >I've noted that not all browsers answer it is invalid. >Here are the tests (http://www.w3c.org.ma/Tests/IDNs/Issue2.html) Quite interesting. >In fact, in a mail I sent before, it was question about: >- Arabic & Hebrew IDNs that are not displayed as claimed by Firefox. .museum TLD is trusted (Firefox displays IDNs in native) but in punycode for Arabic & Hebrew. (http://www.w3c.org.ma/Tests/IDNs/Issue1.html) Yes, that's weird. >- IDNs with %xx in links (http://www.w3c.org.ma/Tests/IDNs/Issue3.html) >Browsers accept links (href attribute) with IDNs in native and punycode but not in escaped notation. >Are %xx encoding a valide notation for IDNs? Yes. RFC 3986 explicitly allows these, but fixes them to use UTF-8. But not all browsers implement them yet. Regards, Martin. #-#-# Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University #-#-# http://www.sw.it.aoyama.ac.jp mailto:duerst@it.aoyama.ac.jp
Received on Wednesday, 28 March 2007 05:19:45 UTC