- From: Daniel Tan <lists@novalistic.com>
- Date: Tue, 23 Jun 2015 02:29:53 +0800
- To: "Tab Atkins Jr." <jackalmage@gmail.com>
- CC: "<www-style@w3.org>" <www-style@w3.org>
On 6/23/2015 2:13 AM, Tab Atkins Jr. wrote: > On Mon, Jun 22, 2015 at 10:07 AM, Daniel Tan <lists@novalistic.com> wrote: >> On the surface, it seems fairly trivial to extend pseudo-classes that take >> An+B expressions such that they allow multiple expressions in a single >> pseudo-class. This could be very handy for targeting many specific children >> in a way that cannot be easily written in a single, or few, expressions: >> >> td:nth-of-type(1, 3, 7, 10) >> >> Admittedly, this feature would be used almost exclusively with integers >> rather than entire An+B expressions, but actively restricting the >> microsyntax in lists seems pointless. >> >> There is :matches(), but it requires having to repeat the pseudo-class which >> is not ideal: >> >> td:matches( >> :nth-of-type(1), :nth-of-type(3), >> :nth-of-type(7), :nth-of-type(10) >> ) >> >> This isn't nearly as popular as :matches(), :has(), or :nth-child(An+B of >> sel), but it does come up fairly often. That said I'm curious to know what >> it would take to implement (and justify implementing) such a feature. > > There's nothing wrong with that at a parsing level. The more complex > grammar of :nth-child() is a bit more difficult, though - > comma-separated indexes, or whole entries? > > ~TJ > Good point - what I initially had in mind was :nth-child(An+B, Cn+D of sel1, sel2), but I can see how that could be interpreted as :nth-child(An+B of sel1, Cn+D of sel2) instead, or how someone might want to write it that way. I hadn't quite thought that through. It seems the complexity of this would be too great taking the selector-list argument into account :/ -- Daniel Tan NOVALISTIC <http://NOVALISTIC.com>
Received on Monday, 22 June 2015 18:30:24 UTC