- From: Linss, Peter <peter.linss@hp.com>
- Date: Fri, 17 Feb 2012 01:20:49 +0000
- To: Chris Lilley <chris@w3.org>
- CC: Boris Zbarsky <bzbarsky@MIT.EDU>, Øyvind Stenhaug <oyvinds@opera.com>, CSS-testsuite <public-css-testsuite@w3.org>
On Feb 16, 2012, at 12:05 PM, Chris Lilley wrote: > On Thursday, February 16, 2012, 7:57:17 PM, Boris wrote: > > BZ> On 2/16/12 1:52 PM, Chris Lilley wrote: >>> I think we should also allow HTML5 (in the XML serialisation, so existing toolchain can consume it). > > BZ> Does the existing toolchain consume reftest references? > > No. But I didn't get the impression that Øyvind was talking about references only. Actually, yes. The toolchain does consume and process references as well. > >>> ØS> Are #help<link>s always necessary for building? > >>> If these are omitted them my understanding is that testing by chapter and reporting by chapter won't work. > > BZ> Is this true specifically for reftest references, which is what Aryeh > BZ> was asking about? > > I wasn't responding to Areyeh in that part. > >>> Similarly, without the assert of what is being tested, it is hard to check the tests, and its hard to auto-annotate the spec to link to the relevant tests. > > BZ> It makes sense to put that information in the test. I see no reason to > BZ> put it in the reference. > > In my later response to Aryeh, I had missed that he was referring to reftest references only. The assertion should not be in the reference.
Received on Friday, 17 February 2012 01:22:07 UTC