- From: Andrew Fedoniouk <news@terrainformatica.com>
- Date: Sun, 13 Jul 2008 15:47:30 -0700
- To: Lachlan Hunt <lachlan.hunt@lachy.id.au>
- CC: fantasai <fantasai.lists@inkedblade.net>, www-style <www-style@w3.org>
Lachlan Hunt wrote:
>
> fantasai wrote:
>> Have you considered handling these queries by passing an option to
>> querySelector that scopes the selector to the context node? (Or in the
>> case of HTML5 scoped style, defining a syntax for the style element
>> that does the same.) What are the advantages of doing this with a
>> pseudo-class instead of that approach?
>
> I did a full analysis of that and other possible solutions previously on
> public-webapi.
>
> http://lists.w3.org/Archives/Public/public-webapi/2008May/0057.html
> http://lists.w3.org/Archives/Public/public-webapi/2008May/0058.html
>
> The solution discussed in that second email is being considered for
> Selectors API 2, where a new method would be introduced that performs a
> pre-parse on the group of selectors, appends :context to the beginning
> of each selector and then parses as normal. This will provide implicit
> scoping and basically standardise the defacto standard of many JS
> libraries, but the explicit :context gives a little more flexibility. So
> both would be useful.
Am I correct in my understanding that:
localRoot.querySelector("div")
will mean precisely:
localRoot.querySelector(":context div")
?
If "yes" then lookup happens *always* from the localRoot element
and that element is exactly the :root in such selectors.
That is why I do not understand why you need special :context pseudo
if we have :root already.
>
> This is the feature request bug I have for tracking that proposal.
>
> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5860
>
--
Andrew Fedoniouk.
http://terrainformatica.com
Received on Sunday, 13 July 2008 22:48:12 UTC