This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
Tantek Çelik wrote: >>Message-ID: <3A92DD77.7C9676BC@w3.org> >>Date: Tue, 20 Feb 2001 16:11:19 -0500 >>From: Philippe Le Hegaret <plh@w3.org> >>To: Martijn Pieters <mj@digicool.com> >>Cc: www-dom@w3.org >>Subject: Re: DOMImplementation.hasFeature and 'Core', 'XML' and '1.0'. >> >>Martijn Pieters wrote: >> >>>So, should a DOM level 2 implementation return true on hasFeature('Core', >>>'1.0) or hasFeature('XML', '1.0')? I *think* it should, but would like to >>>hear confirmation. >> >>The definition of version '1.0' is driven by the DOM Level 1 specification. >>The DOM Level 2 specification doesn't say anything about >>hasFeature('Core', '1.0') and I don't think it has to, given that you can't >>expect a DOM Level 1 implementation to return true. >>hasFeature('XML', '1.0') must definitively return true for a DOM Level 2 >>implementation since DOM Level 2 is build on top of DOM Level 1. >> >>Philippe > > >Redirecting a reply to www-dom-ts since I am raising a DOM 2 Core Test Suite >issue (in response to a message over two years old!). > >I tried searching the archives, and this is the most recent relevant >reference I could find to the question of whether it is correct for the test >suite to test hasFeature("Core","1.0"). > >I don't think the DOM 2 Core Test Suite should be performing that check, >since no DOM 1 Core implementation could be required to return true for that >since "Core" wasn't added as a feature flag until DOM 2. > >The following tests in the DOM 2 Core Test Suite appear to make that >assumption however, and thus should be fixed to not expect >hasFeature("Core","1.0") to return a meaningful value. > > isSupported03.html > isSupported08.html > isSupported12.html > >Thanks, > >Tantek > >--------------------------------------------------------------------------- >Tantek Çelik tantek@cs.stanford.edu >Tasman Development Lead, Microsoft Corporation tantekc@microsoft.com >Representative to W3C CSS and HTML working groups >--------------------------------------------------------------------------- > > > http://www.w3.org/TR/DOM-Level-2-Core/Core.html#ID-5CED94D7 has: version of type DOMString This is the version number of the feature to test. In Level 2, the string can be either "2.0" or "1.0". If the version is not specified, supporting any version of the feature causes the method to return true. This was apparently interpreted by NIST and the Xerces-J team (maybe in response to a NIST test) that "2.0" and "1.0" should be equivalent. I do not know of any definitive WG statement on this issue.
Created attachment 68 [details] Remove isSupported03, isSupported08 and modified isSupported12
Per DOM WG teleconference 26 Nov 2003, tests that assert the return value for isSupported("Core", "1.0") have been removed.
domimplementationhasfeature01_Core_2 also makes the same assumption, about hasFeature rather than isSupported.
Created attachment 80 [details] Removes assertion of hasFeature("Core", "1.0") == true
Oops... and nodeissupported01, nodeissupported02 too!
Created attachment 82 [details] Removed isSupported("Core", "1.0") from nodeissupported01, 02
Created attachment 89 [details] Fix for level3/core/domimplementationgetfeature01 Another one for you in the new L3 Core tests, domimplementationgetfeature01.
Created attachment 90 [details] Fix for level3/core/domimplementationgetfeature02 Also the L3 test domimplementationgetfeature02: this wants (according to the description text) to check for Core empty-string, but actually passes good old "1.0" in again.
Created attachment 102 [details] Fixes for L3 Core domimplementationgetfeature01 and 02