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

RE: attribute-value-selector-004.xht not well formed

From: Arron Eicholz <Arron.Eicholz@microsoft.com>
Date: Tue, 10 Mar 2009 09:35:32 -0700
To: Robert Stam <robert@tallcomponents.com>, Anne van Kesteren <annevk@opera.com>, Sylvain Galineau <sylvaing@microsoft.com>, "L. David Baron" <dbaron@dbaron.org>
CC: "public-css-testsuite@w3.org" <public-css-testsuite@w3.org>
Message-ID: <66D5CB5D6AB0694592FAF5487C50368B29B58CDB8A@NA-EXMSG-C111.redmond.corp.microsoft.com>

I totally get all of your arguments but we need to test the rule and the rule is not about error handling it's about numbers at the beginning of IDENTS. Error handling cases are different cases and have different tests. We are testing that attributes and attribute selectors cannot start with numbers. If that is what the case is specifically testing then we need to have invalid syntax in the page to validate that the selector is not actually matching to an invalid attribute.

There is no way around this in this situation you have to have invalid XML in order for this test to ever test the specific rule in the spec. If we do not have a test for that rule then the rule should be removed from the spec which would be a bad idea because that would leave this undefined and then we have even more issues.

I know this is a problem for the build process. And I know this might be an issue for browsers but the rule is the rule and we need a test for it. The only valid way to visually test this is to prove that the attribute selector is not matching the attribute, and in doing that we need and invalid IDENT for the attribute selector and the attribute itself.

--
Thanks,
Arron Eicholz


-----Original Message-----
From: public-css-testsuite-request@w3.org [mailto:public-css-testsuite-request@w3.org] On Behalf Of Robert Stam
Sent: Tuesday, March 10, 2009 12:44 AM
To: Anne van Kesteren; Sylvain Galineau; L. David Baron
Cc: public-css-testsuite@w3.org
Subject: RE: attribute-value-selector-004.xht not well formed

>-----Original Message-----
>From: public-css-testsuite-request@w3.org [mailto:public-css-testsuite-request@w3.org] On Behalf Of Anne van Kesteren
>Sent: Tuesday, March 10, 2009 8:29 AM
>To: Sylvain Galineau; L. David Baron
>Cc: public-css-testsuite@w3.org
>Subject: Re: attribute-value-selector-004.xht not well formed

>On Tue, 10 Mar 2009 02:43:44 -0000, Sylvain Galineau  
><sylvaing@microsoft.com> wrote:
>> Point well taken on fix-ups. However, we are talking here about an  
>> attribute. I agree with your main point i.e. we want to test that the  
>> entire rule is dropped. The best way to do that would be to use :
>>
>> <p 1badAttr="">This should be green</p>
>>
>> At which point we know the entire rule was dropped. Your proposal  
>> assumes that if p is not red then the other selector in the statement  
>> was also dropped. In other words, that the user agent's error handling  
>> is valid. But isn't that what we're trying to test ?

>If the selector is dropped, how could it ever match in the first place? In  
>any case, if you really want to test it like this maybe a better way would  
>be to use
>   <p title="1lala">
>and
>   [title=1lala]
>as that should have the same effect and work better in XML.

I fully agree with Anne, now the test file will be well formed. 
And based on the face that there are no single or double quotes around the 1lala part in the selector, it should be an IDENT.

Regards,
Robert Stam

TallComponents


Received on Tuesday, 10 March 2009 16:36:29 GMT

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