W3C home > Mailing lists > Public > public-css-testsuite@w3.org > February 2009

Suggestions and corrections to the MS-supplied CSS 2.1 tests

From: Johannes Rössel <johannes.roessel@uni-rostock.de>
Date: Mon, 9 Feb 2009 14:37:53 +0100
To: public-css-testsuite@w3.org
Message-ID: <499031B1.6070408@uni-rostock.de>

Hello,

I recently ran two browsers just for fun through the 7005 tests 
published by Microsoft [1] and I noted a few inconsistencies, typos and 
other things with some of the tests which I noted during clicking 
through them.

As their site suggests to submit corrections and additions to this list, 
I am doing so now.

===

Many tests are too unspecific to gauge rendering issues without looking 
at the source and understanding how exactly it should render. This 
includes many tests that just ask for a line to be shown, where the 
exact difference to a "box" would be subjective. Misrendering can, in 
some cases cause a thicker line to be rendered which would not actually 
be correct (some min-height tests have this problem in IE 7).

I think in this case it would be best to either provide a reference 
rendering (i. e. "Test passes if both lines below have the same width") 
or use other kinds of easily verifyable tests (like "Test passes if no 
red is visible").

Things like "twice the height/width" or "height which is two times its 
width" are hard to judge accurately in some cases, especially when you 
need scrolling to view the boxes completely.

Some tests following right after another have the same test criteria but 
different phrasings for them (i. e. "Test passes if it has a green 
square with no red on the page." vs. "Test passes if it has a green 
square and there is no red visible on the page." [4024 and 4025]). There 
are more instances of this, though. I'd suggest to use consistent and 
identical phrasing for identical test criteria.

Apart from this, there are some test descriptions that do not accurately 
match the test they're describing or have typos or other errors in them. 
Those I found are listed below in the form <number> <relative URL>: <issue>..

3468 ./css/chapter_10/properties/line-height/line-height-111.htm: "Test 
passes if the two boxes below are the same height." -> "Test passes if 
two of the boxes below are the same height."

3597 ./css/chapter_10/properties/max-width/max-width-094.htm: "Test 
passes if the two boxes below are the same max-width." -> "Test passes 
if the two boxes below are the same width." (you can't resize the window 
small enough that you can test whether the maximum width is the same, so 
I believe what's meant is just the width)

3598 ./css/chapter_10/properties/max-width/max-width-095.htm: "Test 
passes if the two boxes below are the same max-width." -> "Test passes 
if the two boxes below are the same width." (you can't resize the window 
small enough that you can test whether the maximum width is the same, so 
I believe what's meant is just the width)

3975 ./css/chapter_10/rules/inline-non-replaced-007.htm: "Test passes if 
there is red is visible on the page." -> "Test passes if there is red 
visible on the page.". Also maybe reverse the colors that red signals a 
failed test, consistent with other tests.

3976 ./css/chapter_10/rules/inline-non-replaced-height-002.htm: 
"contest" -> "content"

4014 ./css/chapter_10/rules/Max-Width-007.htm: I usually don't care 
about whether a box has content or not (for questions that require just 
a "box", but since the green box has green text on a green background, 
why has the blue box orange text?

4054 ./css/chapter_10/rules/MinMax-Resolved-002.htm: "Test passes it has 
a green square." -> "Test passes if it has a green square." (like the 
immediately following test)

4056 ./css/chapter_10/rules/MinMax-Resolved-004.htm: "Test passes if it 
has a a green square." -> "Test passes if it has a green square." (like 
the immediately preceding test)

4699 ./css/chapter_14/properties/background/background-129.htm: 
Superfluous "ed." at the end of the text.

5361 ./css/chapter_15/properties/font/font-004.htm: "Test passes the 
black box is the same height as the orange square and is the same width 
and the blue square." -> "Test passes the black box is the same height 
as the orange square and is the same width as the blue square."

5370 ./css/chapter_15/properties/font/font-020.htm: "Test passes the 
black box is the same height as the orange square and is the same width 
and the blue square." -> "Test passes the black box is the same height 
as the orange square and is the same width as the blue square."

5474 ./css/chapter_15/properties/font-variant/font-variant-003.htm: 
<meta name="assert" content="The 'font-variant' property set to 
'inherit' inherits is value from the patent element's value." /> (should 
be "parent's", instead of "patent's).

5894 ./css/chapter_16/Rules/text-decoration-033.htm: "and the blue text 
below has an unerline." -> "and the blue text below has an underline."

Furthermore there are some tests where the page title and description 
(meta name="assert") seem to have copy/paste errors. I didn't note the 
number of the test when I found that, though. So may be hard to find 
now. Also this meta tag is something you don't look at most of the time 
when just looking at which tests fail and which pass.

===

All relative URLs can be accessed by prefixing 
"http://samples.msdn.microsoft.com/ietestcenter/" or whereever they may 
appear in future. For the moment this works, though.

Regards,
Johannes Rössel

[1] <http://samples.msdn.microsoft.com/ietestcenter/css.htm>

P.S.: It'd be nice to send followups to me as well, as I am not 
subscribed to the list or monitor it closely.
Received on Monday, 9 February 2009 13:44:18 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 20 September 2010 17:51:57 GMT