- From: James Graham <jgraham@opera.com>
- Date: Tue, 22 Mar 2011 14:56:25 +0100
- To: Philippe Le Hegaret <plh@w3.org>
- CC: public-test-infra <public-test-infra@w3.org>, "public-html-testsuite@w3.org" <public-html-testsuite@w3.org>, "public-web-perf@w3.org" <public-web-perf@w3.org>
(adding some known users of testharness.js) On 03/21/2011 04:51 PM, Philippe Le Hegaret wrote: > While I was at it, I wrote a test involving all the existing assertions. > Depending on which browser you use, the results differ and a few > assertions are failing improperly as well (except for assert_unreached). > > I attached the file for you to look at it. > > I'm thinking we'll need some tests for the testharness.js itself. Agreed. The apisample.htm document is supposed to do this to some extent, but it is not perfect. One thing that is clear is that assert_exists and assert_inherits are confusing. The general ideal is that the assertions should test what WebIDL requires. For properties I think it requires that they exist on the prototype of the interface object and so are inherited into instances of that interface via the prototype chain i.e. document.hasOwnProperty("createElement") should be false but "createElement" in document would be true. In light of this, I wonder if assert_exists should be renamed assert_own_property or something, and a higher level function like assert_attribute_exists (name up for bikeshedding) added for checking the attribute is implemented according to the WebIDL requirements. This seems like quite a safe change in the HTML repository. I don't know the situation for other users of testharness.js.
Received on Tuesday, 22 March 2011 13:59:05 UTC