Re: properties-value-002 : vertical-align only applies to inlines and cells

Hey Gérard,

1) I've added (and committed) the invalid flags as suggested. I'm wondering if those test should be factored out into their own files, separating valid and invalid notations.
2) See answers to vertical-align inline


Am 31.01.2013 um 03:30 schrieb Gérard Talbot:
> http://test.csswg.org/source/contributors/rodneyrehm/submitted/css3-transitions/properties-value-002.html

properties-value-002.html tests property value types that are *not* covered by the specification. I've added this test for properties that accept more than what the spec defined. The spec allows transitioning for <length>, <percentage> but says nothing about <vertical-keywords>. These tests are expected to fail.


> Can you check the code with regards to vertical-align values. It seems
> to me that the failure message is not making sense.

I'm not all that happy with testharness' presentation either. I'd like some meta-property "fail-message" to explain to a human being what the failure of a particular assertion actually means. I'm going to write up my experience with CSSWG testing and will cover this issue in greater detail.


> Fail	vertical-align vertical(keyword) / values	assert_not_equals: must
> not be target value after start got disallowed value "bottom"

The assertion verifies that the computed value of vertical-align is *not* the transition-end-value. If that were the case, no transition would be performed, as we have an immediate change of state. A human readable message for this assertion would be »Property change does not transition«


> Fail	vertical-align vertical(keyword) / events	assert_equals: Expected
> TransitionEnd events triggered on .transition expected
> "vertical-align:0.5s" but got ""

A human readable message would be: »No TransitionEnd triggered«


> but 'vertical-align: top' and 'vertical-align: bottom' do not and can
> not apply to a block-level element like div.transition.

making the div.transition {display: inline;} doesn't change a thing (for me)



Best Regards,
Rod