Non-element Selectors Module Level 1

Group: csswg
Shortname: selectors-nonelement
Level: 1
Status: ED
Work Status: Exploring
ED: http://dev.w3.org/csswg/selectors-nonelement/
TR: http://www.w3.org/TR/selectors-nonelement-1/
Previous version: http://www.w3.org/TR/2014/WD-selectors-nonelement-1-20140603/
Editor: Jirka Kosek, Invited Expert, jirka@kosek.cz
Editor: Tab Atkins Jr., Google, http://xanthir.com/contact/
Abstract: Non-element Selectors extends [[!SELECTORS4]] and allow selecting other kinds of document nodes than elements. This is useful when selectors are used as a general document query language.
Abstract: Non-element Selectors are not intended to be used in CSS, but only as a separate query language in other host environments.
Link Defaults: css-namespaces (dfn) namespace prefix, selectors (dfn) originating element/pseudo-element

Introduction

Selectors are a very popular mechanism for selecting things in HTML and XML content. They are not used only in CSS [[CSS3SYN]] but also as a standalone query language in libraries like jQuery, in newer standardized browser APIs like [[SELECTORS-API]] and in other Web standards like ITS 2.0. The [[SELECTORS4]] specification only defines selectors for selecting element nodes from the document, but some uses of Selectors would like to select other types of nodes as well. This specification extends [[SELECTORS4]] with additional selectors that can be used for selecting non-element nodes in a document tree. Note: Currently the draft only defines means for selecting and matching attribute nodes, but other kinds of nodes, such as comments or processing instructions, might be supported in the future.

Non-element Selectors

Attribute node selector

An attribute node selector represents an attribute node in a document tree. Its syntax is:
		::attr() = ::attr( <>? )
		<namespace-attr> = [ <>? '|' ]? <>
		<na-prefix> = <> | '*'
		<na-name> = <> | '*'
	
No whitespace is allowed between the tokens of <>. <> is divided into two halves: an optional prefix preceding a '|' character, and an attribute name following it. If the <> is provided as an <>, it must match a declared namespace prefix, in which case the selector only matches attributes in that namespace; if it doesn't match a declared namespace prefix, the selector matches nothing. If the <> is provided as a '*' character, the selector matches attributes in any namespace. If the <> is omitted, the selector only matches attributes in no namespace. If the <> is an <>, the selector matches attributes with that name. If the <> is a '*' character, the selector matches attributes with any name. If the <> is omitted entirely, the selector matches any attribute in any namespace. The selector matches an attribute node with the given namespace and name on the originating element, if such an attribute exists. The selector uses pseudo-element syntax.
The following ITS rules use an attribute node selector to switch off translatability of title attribute on abbr elements.
			<rules xmlns="http://www.w3.org/2005/11/its"
			       version="2.0"
			       queryLanguage="css">
			  <translateRule selector="abbr::attr(title)" translate="no"/>
			</rules>
		
Although entirely valid in Selectors used in the scope of CSS, attribute node selectors never generate boxes.