- From: Jo Rabin <jrabin@mtld.mobi>
- Date: Mon, 11 Feb 2008 11:15:03 -0000
- To: <public-mobileok-checker@w3.org>
One could argue that the HTTP handling is not clearly enough explained in mobileOK Basic, couldn't one? :-( I think it needs to be spelled out more carefully what happens on external resources and that it is not the intention to count redirects etc. and I think as suggested that it should read 4xx not just 404 failures. If this is the case we should raise an ISSUE ... Jo > -----Original Message----- > From: public-mobileok-checker-request@w3.org [mailto:public-mobileok- > checker-request@w3.org] On Behalf Of Miguel Garcia > Sent: 08 February 2008 08:44 > To: Dominique Hazael-Massieux; public-mobileok-checker@w3.org > Subject: RE: Failure on unrecognized format for LINK_TARGET_FORMAT > > > >>Hi, > >> > >>With the now implemented handling of HTTP errors reporting, the > checker > >>will now fail when making a link to a non-recognized content-type. > >> > >>Typical scenario: > >> * the checker checks a link to an MP3 file > >> * it thus makes a request with its known Accept header > >> * the server doesn't recognize this as matching the allowed types for > >>MP3 (audio/mpeg) > >> * it thus returns a 406 Not Acceptable > >> * which triggers "If the HTTP status represents failure (4xx), other > >>than 404 or a request for authentication (e.g. 401), FAIL" > >> > >>This obviously is not the intended effect, since LINK_TARGET_FORMAT > has > >>a test specifically to check that linked resources are in an > acceptable > >>format, which raises at most a WARN. > >> > >>I think the FAILs in the HTTP Response handling should only be > triggered > >>when analysing the resource itself or one of its included resources > >>(distinction we already make for basic auth). > >> > >>What do you think? > >> > >>Dom > >> > > You are right. > > Furthermore reviewing mobileOk Basic Test 1.0 I have noticed there is a > note in 3.10 LINK_TARGET_FORMAT stating "404 and 5xx HTTP status do not > result in failure when conducting this test". We could expand this note > to cover all 4xx and 5xx HTTP status. > > Changes in checker code should be really easy just adding a condition so > linked resources don't report FAIL in these cases. > > Miguel and Abel
Received on Monday, 11 February 2008 11:15:23 UTC