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

Re: [CSS 2.1] Invalid test cases

From: Gérard Talbot <css21testsuite@gtalbot.org>
Date: Wed, 21 Jul 2010 15:31:38 -0700
Message-ID: <7dee25e41738e0932cbebabaf204ba01.squirrel@cp3.shieldhost.com>
To: "Arron Eicholz" <Arron.Eicholz@microsoft.com>
Cc: "public-css-testsuite@w3.org" <public-css-testsuite@w3.org>

> I am listing the currently invalid test cases as I have found them in my
> review. These cases can probably be easily updated but I wanted to get a
> list out there so we had something to go through and try and fix.
>

Hello Arron,

Thank you for your list: such list can be, will be very useful.

> http://test.csswg.org/suites/css2.1/20100701/html4/active-selector-000.htm
> CSS does not define what elements can be in the :active state. Since the
> test has a *:active selector this causes an issue and this case is
> currently invalid.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/active-selector-004.htm
> CSS does not define what elements can be in the :active state. Since the
> test has a *:active selector this causes an issue and this case is
> currently invalid.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/at-charset-071.htm
> http://test.csswg.org/suites/css2.1/20100701/html4/character-encoding-031.htm
> é is missing from the p element. I have a feeling something in the build
> process converted it to &eacute; and added that before the p tag.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/attribute-value-selector-010.htm
> http://test.csswg.org/suites/css2.1/20100701/html4/lang-pseudoclass-001.htm
> The build process seems to be stripping out the xml: namespace in
> elements when build as HTML. This shouldn't happen in this case since
> the test is testing the presence of the namespace.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/background-attachment-006.htm
> The swatch-teal.png is missing
>
> http://test.csswg.org/suites/css2.1/20100701/html4/before-after-display-types-001.htm
> flexbox is used in this test case along with -moz- prefixes.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/before-after-dynamic-attr-001.htm
> double colon before and after is only valid in CSS3
>
> http://test.csswg.org/suites/css2.1/20100701/html4/c5504-mrgn-l-002.htm
> Negative margin doesn't seem to define weather it should create
> scrolling on the viewport.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/c61-phys-len-000.htm
> This requires screens to be properly calibrated to a specific size. Also
> not all credit cards are the same size.
>

Arron,

I agree with each and all of your findings listed above.


> http://test.csswg.org/suites/css2.1/20100701/html4/cursor-023.htm
> Text is incorrect for the behavior that is specified.

The text of expected result was updated:
http://test.csswg.org/source/contributors/hixie/submitted/css2.1/ui/cursor-023.xht

> Plus the CSS spec
> does not say what image types are supported for cursors.

I entirely agree with you that it would be best/ideal if the list of
image types for custom cursors was specified.


> This case
> cannot require a png as a valid cursor type since it is not specified
> what user agents are required to support for cursor images.


If png is a W3C recommendation
http://www.w3.org/TR/PNG/
then I think png image format can/should be expected for such testcase.
As submitted, I did not like the testcase for other reasons
http://lists.w3.org/Archives/Public/public-css-testsuite/2010Jul/0009.html
and I have proposed this testcase:

http://www.gtalbot.org/BrowserBugsSection/css21testsuite/cursor-025.html

as a complementary testcase; such new testcase would not require a scale
down of dimensions and the testcase would apply on a <div>.


> http://test.csswg.org/suites/css2.1/20100701/html4/cursor-024.htm
> The CSS spec does not say what image types are supported for cursors.
> This case cannot require a gif as a valid cursor type since it is not
> specified what user agents are required to support for cursor images.

Microsoft IE7 and IE8 support .ani cursor. But only IE supports .ani
cursor so far. Years ago, I created 3 testcases with .ani cursor and I
could submit one.

Again, I entirely agree with you that it would be best/ideal for
everyone involved if the spec could explicitly indicate the image types
and/or cursor types for animated stuff.

> http://test.csswg.org/suites/css2.1/20100701/html4/first-letter-characters-001.htm
> Build parse error.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/first-line-floats-002.htm
> Double colon CSS3 definition for first-line is not valid for CSS 2.1.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/first-line-pseudo-006.htm
> Spec is unclear if an empty block is ignored for first-line.
>
> http://test.csswg.org/suites/css2.1/20100701/html4/floats-107.htm
> Painting of the positioned element comes after the painting of the
> floats thus the red should be visible. This case needs a good rewrite.

Agreed.

(...)

> http://test.csswg.org/suites/css2.1/20100701/html4/table-caption-002.htm
> Multiple caption handling is not clearly defined in CSS.


I believe multiple captions must all be rendered, in document order and
I believe this is clearly defined.

"All elements with 'display: table-caption' must be rendered, as
described in section 17.4."
http://www.w3.org/TR/CSS21/tables.html#table-display

"the table box generates an anonymous box that contains the table box
itself and any caption boxes (in document order). The caption boxes are
block-level boxes that retain their own content, padding, margin, and
border areas, and are rendered as normal blocks inside the anonymous
box. Whether the caption boxes are placed before or after the table box
is decided by the 'caption-side' property, as described below. "
http://www.w3.org/TR/CSS21/tables.html#model
Note the plural form.

And note that if only the first of multiple captions would be rendered,
then I am sure the spec would have said it just like it said so for
table-header-group and table-footer-group elements.

I believe Opera 10.60 passes this test:
http://test.csswg.org/suites/css2.1/20100701/html4/table-caption-002.htm


> http://test.csswg.org/suites/css2.1/20100701/html4/z-index-020.htm
> test uses CSS3 outline-offset. And Outlines can be drawn at 2 different
> stacking contexts so outline in this tests is flaky and can cause false
> failures.

Agreed. Use of outline makes the testcase as coded not entirely reliable.

regards, Gérard
-- 
Contributions to the CSS 2.1 test suite:
http://www.gtalbot.org/BrowserBugsSection/css21testsuite/

CSS 2.1 test suite (beta 1; July 1st 2010):
http://test.csswg.org/suites/css2.1/20100701/html4/toc.html

CSS 2.1 test suite contributors:
http://test.csswg.org/source/contributors/
Received on Wednesday, 21 July 2010 22:32:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 20 September 2010 17:52:02 GMT