W3C home > Mailing lists > Public > www-style@w3.org > April 2011

[CSS2.1] at-rules, parentheses and error recovery and incorrect link

From: Alan Gresley <alan@css-class.com>
Date: Tue, 05 Apr 2011 03:15:17 +1000
Message-ID: <4D99FCA5.4010000@css-class.com>
To: CSS 3 W3C Group <www-style@w3.org>
Hello all,

I'm trying to find in the spec what is meant to happen when a 
property-value pairing or value alone does not appear in a string with 
parentheses '(....)'.

@media () {
      body p { background: red }

CSS2.1 has this 7.2.1.


Nothing is mentioned about what should happen if no properties values 
are declared. The spec does not even mention properties or values.

The link to the word 'statement' in 7.2.1 is incorrect. It currently 
points here,


instead of here,


but this is incorrect anyway because properties and values are one type 
of statement. Perhaps it should point here instead.


This part of the spec also does not mention properties or values but are 
bundled up as being part of 'everything' in the below spec.

   # An at-rule consists of *everything* up to and including the
   # next semicolon (;) or the next block, whichever comes first.

CSS3 has this (point 3).


   # Properties always require a value to form a declaration.

In this test case.


In Firefox 3.6.16, the line is red.

      p { background: red; color: white }

     @media () { body p { background: orange } }

     p { background: green }

If the parentheses had '(foo)' thought, Firefox 3.6.16 drops it. FF 4b 
does not have the bug.

In this test case.


In IE9 and WebKit, the line is red.

     @media ( { body p { background: orange } }

     p { background: red }

According the nesting of blocks, remainder of the style-sheet should be 
thrown out due to a malformed statement resulting in an unexpected end 
of style sheet. I should mention that IE8 shows a green line.

A string within parentheses in a media query must have either have a 
property and value pairing or value to form a declaration. If no such 
value exist, the media query should be ignored.

In a reply to me [3] in a different thread 'Re: [CSS2.1] blocks with 
missing end bracket and parsing error', Bjoern writes:

   > Some time ago the CSS Working Group decided to change slash
   > clarify the specification to require {[()]} matching when
   > parsing selectors.

Doing some digging I find this is based on issue-106 [1] and these CSS 
WG minutes [2]. Currently browsers are parsing garbage syntax and this 
results in many possible hacks for very recent versions of browsers.

[1] http://wiki.csswg.org/spec/css2.1#issue-106
[2] http://lists.w3.org/Archives/Public/www-style/2008Jul/0589.html
[3] http://lists.w3.org/Archives/Public/www-style/2011Mar/0579.html

Alan http://css-class.com/

Armies Cannot Stop An Idea Whose Time Has Come. - Victor Hugo
Received on Monday, 4 April 2011 17:15:55 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:38:45 UTC