Sizzle-test | nth-child selectors issue

[~flipper]

First: Terrific project!

In the ":nth-child Selectors" section of the test, specifically, the 3
rules associated with #table-2 tr:nth child ( etc ...):

The five rows are the same #ccc background-color in Safari 4, Safari
5, the latest Webkit nightlies, and OmniWeb 5.10.1 (in OS X), IE8
(Win7 Ultimate), and (yes, sorry ...) Epiphany (Ubuntu 10.04).
(OmniWeb also failed the ":nth-of-type" paragraph text-color tests).
The last two versions of Camino ( OS X) also failed the :nth-child
Selectors, #table-2 tests, but showed all backgrounds as 'orange.'

The above 'outcomes' were not notated as being "unsupported". Camino,
in a large sense, and OmniWeb, had a number of 'failures' that were
noted, as per the design of the tests, as being "unsupported." The
tests that failed, yet were not showing the "unsupported" blurbs,
might be of more concern.

Not quibbling; I find the success ratios, and breadth (in terms of
browsers and platforms), nothing short of astonishing. All the sizzle-
engine tests worked in Firefox and Opera on the OS X install, as well
as in Firefox in Win7, and Galeon and Firefox in Ubuntu. That's
amazing! Great work!

[Aaron Gustafson]

On Aug 3, 11:35 am, "~flipper" <stegner.br...@gmail.com> wrote:

> First: Terrific project!

Thanks.

There have been some changes to the underlying CSS handlers in Webkit
that are probably the root cause of most of this. Others may just need
a nudge or something. I'm revisiting the library to correct Chrome's
mishandling of attribute selectors, so I'll take a look @ the other
stuff then as well.