Re: choice of javascript DOM methods suitable for use in tests

On 29/11/2010 18:14, Ian Hickson wrote:
> Which is to say, for the purposes of the test suite I would just treat the
> "expected to" terminology in the rendering section as having the strength
> of RFC2119 MUST statements


yes sure, but I'd still think it useful that the eventual organisation 
of the testsuite means it's easy for (say) an html5 parser 
implementation to easily distinguish which tests apply to such non 
rendering application,

and more particularly for this thread, what's the feeling about mathml 
and svg tests? The most agressively minimalist reading of the html5 spec 
is that they should be parsed but don't need to be rendered.
which is fine, but not particularly useful to an end user.

Correspondingly for the test suite, I'd like to see the test suite have 
mathml parsing tests and mathml rendering tests in two separate 
directories (or whatever way of flagging the status of tests is chosen)
Formally the mathml rendering tests could be off your CR critical path, 
although in fact I don't think any of the tests we had in mind for that 
section wouldn't already be passed by two of the main current browsers.

david

Received on Monday, 29 November 2010 19:58:48 UTC