[Webkit-unassigned] [Bug 18027] CSS3 Selector Test: combination of hover and multiple chained sibling selector fails in Webkit

Wed Nov 25 22:22:54 PST 2009

https://bugs.webkit.org/show_bug.cgi?id=18027

Eric Seidel <eric at webkit.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #41669|review?                     |review-
               Flag|                            |

--- Comment #15 from Eric Seidel <eric at webkit.org>  2009-11-25 22:22:53 PST ---
(From update of attachment 41669)
Sorry.  Bugs tend to get lost in the queue. :(

OK, tests are better, but still not as easy to read as they should be.

+
+The green square should turn red when the textarea is focused.
+
+Test passed.

You could easily hide the "green square" stuff when in DumpRenderTree, you can
look for layoutTestController.  it's OK as is, but could be better.

Also:

PASS -- explanation why it passed

is better than putting "pass" latter in the sentence as it's easier for people
to scan.  See the script-test examples in the tree, their output is PASS :
reason and FAIL : reason.

Likewise, this is also confusing when reading as text:
+The third square should be green and turn red when the pointer hovers over the
first black square.
+
+Test for https://bugs.webkit.org/show_bug.cgi?id=18027 - CSS3 Selector Test:
combination of hover and multiple chained sibling selector fails in Webkit.
+
+The hover effect works!

Extra junk not needed in tests:
+<!DOCTYPE html>
+<html>
+<head>
+<title></title>

Overall this looks like a good change.  I really need to write a document about
writing great layout tests.  Your tests should be as short as possible and
super-easy to read/understand.  The problem is that most people won't have
nearly the level of understanding that you do about the issue, so you need to
make your tests super-easy to read/understand so that they can make quick
judgements about what is passing/failing or what might be wrong.  This is why
we try to encourage things like script-tests as they're generally easy to read.
:)

Why do the tests need to be run from onload?  It's OK for them to be run from
onload, but moving the <script> tag to the end of the body instead would have
the same effect (or very similar).

I think we should go one more round on improving the tests here.  Please just
email me when you update it and I'll be happy to review it right away!  Again,
my apologies for the horrible delay!

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.