W3C home > Mailing lists > Public > public-css-testsuite@w3.org > October 2010

incorrect pass conditions (and missing 'should' flag) in numerous line-height tests

From: L. David Baron <dbaron@dbaron.org>
Date: Fri, 15 Oct 2010 16:54:59 -0700
To: public-css-testsuite@w3.org
Message-ID: <20101015235459.GA9796@pickering.dbaron.org>
The following tests:
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-001.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-001.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-012.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-012.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-023.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-023.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-034.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-034.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-045.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-045.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-056.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-056.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-067.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-067.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-078.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-078.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-089.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-089.xht
http://test.csswg.org/suites/css2.1/20101001/html4/line-height-100.htm
http://test.csswg.org/suites/css2.1/20101001/xhtml1/line-height-100.xht
have an incorrect pass condition.  They're all testing a value of
line-height that is a parse error (since negative values are not
allowed). 

Therefore, what they end up testing is the implementation of
'line-height: normal'.  The spec says (section 10.8):

  We recommend a used value for 'normal' between 1.0 to 1.2.

However, this test asserts that the value is either exactly 1.0 or
exactly 1.2.  (It is also, therefore, testing a "recommends" without
the "should" flag, but that can be fixed in a better way than just
adding the "should" flag.)

I recommend making the following changes:
 * in a declaration before the declaration with the negative value,
   set line-height to a large value (say, 3em)
 * compare the result to a box that has height 3em (so there are
   only 2 boxes total).
 * add the "invalid" flag to the tests.

-David

-- 
L. David Baron                                 http://dbaron.org/
Mozilla Corporation                       http://www.mozilla.com/
Received on Friday, 15 October 2010 23:55:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 15 October 2010 23:55:49 GMT