W3C home > Mailing lists > Public > public-css-testsuite@w3.org > September 2011

[CSS3-mediaqueries]: Invalid test cases in test suite

From: Arron Eicholz <Arron.Eicholz@microsoft.com>
Date: Mon, 26 Sep 2011 20:52:41 +0000
To: "L. David Baron (dbaron@dbaron.org)" <dbaron@dbaron.org>, "public-css-testsuite@w3.org" <public-css-testsuite@w3.org>
Message-ID: <07349ECFC3608F48BC3B10459913E70B148D2C33@TK5EX14MBXC294.redmond.corp.microsoft.com>
I have diligently been going over the Media Queries test suite and I think I have discovered a few more issues with the test.

  query_should_not_be_parseable("not (orientation)");
  query_should_not_be_parseable("only (orientation)");
  query_should_not_be_parseable("all and color :");
  query_should_not_be_parseable("all and color : 1");

The tests for bad parsing are incorrect in some cases. The cases are currently set up to test as not being parseable. The first 2 cases according to Section 3.1 of the media queries spec, a case where the IDENT production is not matched should fall into the malformed media query case. The last 2 cases are specifically malformed media queries. Looking at the malformed media query case, media queries with unexpected tokens should be represented as 'not all'. This then means that the cases I listed above should all be parseable and the .mediaText will return "screen, not all" for all 4 cases.

features = [ "resolution", "min-resolution", "max-resolution" ];
expression_should_not_be_parseable(feature + ": 0dpi");

The test for resolution should be parseable. According to the spec (section 5) the only information about resolution is that it takes a positive number. Unfortunately we defined 0 to be both positive and negative. Thus this case should be parseable. Granted I do think that 0dpi is silly it should be parseable or we should maybe change the spec to say that resolution must a value of 1 or more.

content_div.style.font = "-moz-initial";

There is also a Mozilla prefixed value setting within the test case. All submitted test cases should not include prefixed values or properties.

--
Thanks,
Arron Eicholz
Received on Monday, 26 September 2011 20:53:12 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 26 September 2011 20:53:22 GMT