- From: Simon Pieters <simonp@opera.com>
- Date: Thu, 13 Jun 2013 11:53:45 +0200
- To: "L. David Baron" <dbaron@dbaron.org>
- CC: www-style@w3.org
On 6/8/13 9:14 AM, L. David Baron wrote: > Glenn Adams pointed out to me at today's Test the Web Forward event > in Tokyo that the rules in > http://dev.w3.org/csswg/cssom/#serializing-selectors for serializing > :nth-child(an+b) forms drop signs. This changes the meaning of > selectors during serialization, which is incorrect. > > In particular, this rule: > > # 4. If a is one or minus one and b is zero let the value be "n" > # (U+006E). > > omits the sign on a. There is a semantic difference between > :nth-child(n) (which matches all elements) and :nth-child(-n) (which > matches none) that is lost by this rule. > > Likewise: > > # 5. If a is one or minus one let the value be "n" (U+006E), > # followed by "+" (U+002B) if b is positive, followed by b > # serialized as <integer>. > > also omits the sign on a. This loses the semantic difference > between :nth-child(-n+2) which matches only the first or second > child, and :nth-child(n+2) which matches all children except the > first. > > -David > https://dvcs.w3.org/hg/csswg/rev/f15e37a210a8 http://dev.w3.org/csswg/css-syntax/#serializing-anb seems to not have the problems mentioned above. (I could put this in CSSOM instead if that's better.) -- Simon Pieters Opera Software
Received on Thursday, 13 June 2013 09:52:53 UTC