- From: Karl Dubost <karl@w3.org>
- Date: Thu, 21 Aug 2008 12:26:47 +0900
- To: Ian Hickson <ian@hixie.ch>
- Cc: Toby A Inkster <tai@g5n.co.uk>, Henri Sivonen <hsivonen@iki.fi>, Richard Ishida <ishida@w3.org>, Anne van Kesteren <annevk@opera.com>, public-html@w3.org
Hi, Reading the Editor's Draft 21 August 2008 about xml:lang="" and lang="", I was trying to figure out different test cases. http://dev.w3.org/html5/spec/ The most relevant part for these tests (in text/html) is: If both the xml:lang attribute and the lang attribute are set on an element, user agents must use the xml:lang attribute, and the lang attribute must be ignored for the purposes of determining the element's language. This is an improvement on the previous version which was giving different results depending on the mime-type. I put the results for the previous WD so that it is clearer for people [10 June 2008 result] Comment: Ian there are three mentions of RFC 3066 in the editors draft for meta "Content Language State", hreflang and lang. Could you fix it to RFC 4646? XML processing is defined by http://www.w3.org/TR/2006/REC-xml-20060816/#sec-lang-tag which refers RFC 3066 This would need to be tested in browsers. Case 1 - <html xml:lang="" and lang=""> 1.a text/html -> language unknown [10 June 2008 unknown] 1.b application/xhtml+xml -> language unknown (see note #2) Case 2 - <html xml:lang="fr" and lang="ja"> 2.a text/html -> language fr [10 June 2008 ja] 2.b application/xhtml+xml -> language fr (see note #1) Case 3 - <html xml:lang="fr" and lang=""> 3.a text/html -> language fr [10 June 2008 unknown] 3.b application/xhtml+xml -> language fr (see note #1) Case 4 - <html xml:lang="" and lang="ja"> 4.a text/html -> language unknown [10 June 2008 ja] 4.b application/xhtml+xml -> language unknown (see note #2) Case 5 - <html xml:lang="fr" and lang="fr-CA"> 5.a text/html -> language fr [10 June 2008 fr-CA] 5.b application/xhtml+xml -> language fr (see note #1) Case 6 - <html xml:lang="fr,de" and lang="fr,de"> 6.a text/html -> language fr (see note #3) [10 June 2008 fr] 6.b application/xhtml+xml -> undefined. I don't know what the implementations should do (see note #4) notes: #1 The local value overrides transport protocol information. interesting, that seems to violate HTTP precedence principles. There would be interesting tests to do with http declared in meta or directly configured on the server. #2 except if the value is specified through external transport protocol (ex: http or mime) #3 collect a sequence of characters http://dev.w3.org/html5/spec/#collect (Editor's draft) #4 The last RFC for indentiying languages is RFC 4646 http://www.ietf.org/rfc/rfc4646.txt There are interesting requirements on * buffer sizes for the values of languages tags and what should be done in case of errors. * language tags case -> they are case-insensitive fr-CA and fr-CA and fR-Ca are strictly equivalent (could be another test case) * Classes of Conformance for processors of the values well-formed and validating BUT unfortunately it doesn't say what it should do when the data is non conformant, aka rejecting with an unknown value? warning? stop processing? etc. -- Karl Dubost - W3C http://www.w3.org/QA/ Be Strict To Be Cool
Received on Thursday, 21 August 2008 03:27:28 UTC