parserxml.php

来自「php 开发的内容管理系统」· PHP 代码 · 共 644 行 · 第 1/2 页

<?php
/**
 *
 * @package MediaWiki
 * @subpackage Experimental
 */

/** */
require_once ('Parser.php');

/**
 * This should one day become the XML->(X)HTML parser
 * Based on work by Jan Hidders and Magnus Manske
 * To use, set
 *    $wgUseXMLparser = true ;
 *    $wgEnableParserCache = false ;
 *    $wgWiki2xml to the path and executable of the command line version (cli)
 * in LocalSettings.php
 * @package MediaWiki
 * @subpackage Experimental
 */

/**
 * the base class for an element
 * @package MediaWiki
 * @subpackage Experimental
 */
class element {
	var $name = '';
	var $attrs = array ();
	var $children = array ();

	/**
	* This finds the ATTRS element and returns the ATTR sub-children as a single string
	* @todo FIXME $parser always empty when calling makeXHTML()
	*/
	function getSourceAttrs() {
		$ret = '';
		foreach ($this->children as $child) {
			if (!is_string($child) AND $child->name == 'ATTRS') {
				$ret = $child->makeXHTML($parser);
			}
		}
		return $ret;
	}

	/**
	 * This collects the ATTR thingies for getSourceAttrs()
	 */
	function getTheseAttrs() {
		$ret = array ();
		foreach ($this->children as $child) {
			if (!is_string($child) AND $child->name == 'ATTR') {
				$ret[] = $child->attrs["NAME"]."='".$child->children[0]."'";
			}
		}
		return implode(' ', $ret);
	}

	function fixLinkTails(& $parser, $key) {
		$k2 = $key +1;
		if (!isset ($this->children[$k2]))
			return;
		if (!is_string($this->children[$k2]))
			return;
		if (is_string($this->children[$key]))
			return;
		if ($this->children[$key]->name != "LINK")
			return;

		$n = $this->children[$k2];
		$s = '';
		while ($n != '' AND (($n[0] >= 'a' AND $n[0] <= 'z') OR $n[0] == '盲' OR $n[0] == '枚' OR $n[0] == '眉' OR $n[0] == '脽')) {
			$s .= $n[0];
			$n = substr($n, 1);
		}
		$this->children[$k2] = $n;

		if (count($this->children[$key]->children) > 1) {
			$kl = array_keys($this->children[$key]->children);
			$kl = array_pop($kl);
			$this->children[$key]->children[$kl]->children[] = $s;
		} else {
			$e = new element;
			$e->name = "LINKOPTION";
			$t = $this->children[$key]->sub_makeXHTML($parser);
			$e->children[] = trim($t).$s;
			$this->children[$key]->children[] = $e;
		}
	}

	/**
	* This function generates the XHTML for the entire subtree
	*/
	function sub_makeXHTML(& $parser, $tag = '', $attr = '') {
		$ret = '';

		$attr2 = $this->getSourceAttrs();
		if ($attr != '' AND $attr2 != '')
			$attr .= ' ';
		$attr .= $attr2;

		if ($tag != '') {
			$ret .= '<'.$tag;
			if ($attr != '')
				$ret .= ' '.$attr;
			$ret .= '>';
		}

		# THIS SHOULD BE DONE IN THE WIKI2XML-PARSER INSTEAD
		#	foreach ( array_keys ( $this->children ) AS $x )
		#	   $this->fixLinkTails ( $parser , $x ) ;

		foreach ($this->children as $child) {
			if (is_string($child)) {
				$ret .= $child;
			} elseif ($child->name != 'ATTRS') {
				$ret .= $child->makeXHTML($parser);
			}
		}
		if ($tag != '')
			$ret .= '</'.$tag.">\n";
		return $ret;
	}

	/**
	* Link functions
	*/
	function createInternalLink(& $parser, $target, $display_title, $options) {
		global $wgUser;
		$skin = $wgUser->getSkin();
		$tp = explode(':', $target); # tp = target parts
		$title = ''; # The plain title
		$language = ''; # The language/meta/etc. part
		$namespace = ''; # The namespace, if any
		$subtarget = ''; # The '#' thingy

		$nt = Title :: newFromText($target);
		$fl = strtoupper($this->attrs['FORCEDLINK']) == 'YES';

		if ($fl || count($tp) == 1) {
			# Plain and simple case
			$title = $target;
		} else {
			# There's stuff missing here...
			if ($nt->getNamespace() == NS_IMAGE) {
				$options[] = $display_title;
				return $parser->makeImage($nt, implode('|', $options));
			} else {
				# Default
				$title = $target;
			}
		}

		if ($language != '') {
			# External link within the WikiMedia project
			return "{language link}";
		} else {
			if ($namespace != '') {
				# Link to another namespace, check for image/media stuff
				return "{namespace link}";
			} else {
				return $skin->makeLink($target, $display_title);
			}
		}
	}

	/** @todo document */
	function makeInternalLink(& $parser) {
		$target = '';
		$option = array ();
		foreach ($this->children as $child) {
			if (is_string($child)) {
				# This shouldn't be the case!
			} else {
				if ($child->name == 'LINKTARGET') {
					$target = trim($child->makeXHTML($parser));
				} else {
					$option[] = trim($child->makeXHTML($parser));
				}
			}
		}

		if (count($option) == 0)
			$option[] = $target; # Create dummy display title
		$display_title = array_pop($option);
		return $this->createInternalLink($parser, $target, $display_title, $option);
	}

	/** @todo document */
	function getTemplateXHTML($title, $parts, & $parser) {
		global $wgLang, $wgUser;
		$skin = $wgUser->getSkin();
		$ot = $title; # Original title
		if (count(explode(':', $title)) == 1)
			$title = $wgLang->getNsText(NS_TEMPLATE).":".$title;
		$nt = Title :: newFromText($title);
		$id = $nt->getArticleID();
		if ($id == 0) {
			# No/non-existing page
			return $skin->makeBrokenLink($title, $ot);
		}

		$a = 0;
		$tv = array (); # Template variables
		foreach ($parts AS $part) {
			$a ++;
			$x = explode('=', $part, 2);
			if (count($x) == 1)
				$key = "{$a}";
			else
				$key = $x[0];
			$value = array_pop($x);
			$tv[$key] = $value;
		}
		$art = new Article($nt);
		$text = $art->getContent(false);
		$parser->plain_parse($text, true, $tv);

		return $text;
	}

	/**
	 * This function actually converts wikiXML into XHTML tags
	 * @todo use switch() !
	 */
	function makeXHTML(& $parser) {
		$ret = '';
		$n = $this->name; # Shortcut

		if ($n == 'EXTENSION') {
			# Fix allowed HTML
			$old_n = $n;
			$ext = strtoupper($this->attrs['NAME']);

			switch($ext) {
				case 'B':
				case 'STRONG':
					$n = 'BOLD';
					break;
				case 'I':
				case 'EM':
					$n = 'ITALICS';
					break;
				case 'U':
					$n = 'UNDERLINED'; # Hey, virtual wiki tag! ;-)
					break;
				case 'S':
					$n = 'STRIKE';
					break;
				case 'P':
					$n = 'PARAGRAPH';
					break;
				case 'TABLE':
					$n = 'TABLE';
					break;
				case 'TR':
					$n = 'TABLEROW';
					break;
				case 'TD':
					$n = 'TABLECELL';
					break;
				case 'TH':
					$n = 'TABLEHEAD';
					break;
				case 'CAPTION':
					$n = 'CAPTION';
					break;
				case 'NOWIKI':
					$n = 'NOWIKI';
					break;
				}
			if ($n != $old_n) {
				unset ($this->attrs['NAME']); # Cleanup
			} elseif ($parser->nowiki > 0) {
				# No 'real' wiki tags allowed in nowiki section
				$n = '';
			}
		} // $n = 'EXTENSION'

		switch($n) {
			case 'ARTICLE':
				$ret .= $this->sub_makeXHTML($parser);
				break;
			case 'HEADING':
				$ret .= $this->sub_makeXHTML($parser, 'h'.$this->attrs['LEVEL']);
				break;
			case 'PARAGRAPH':
				$ret .= $this->sub_makeXHTML($parser, 'p');
				break;
			case 'BOLD':
				$ret .= $this->sub_makeXHTML($parser, 'strong');
				break;
			case 'ITALICS':
				$ret .= $this->sub_makeXHTML($parser, 'em');
				break;

			# These don't exist as wiki markup
			case 'UNDERLINED':
				$ret .= $this->sub_makeXHTML($parser, 'u');
				break;
			case 'STRIKE':
				$ret .= $this->sub_makeXHTML($parser, 'strike');
				break;

			# HTML comment
			case 'COMMENT':
				# Comments are parsed out
				$ret .= '';
				break;


			# Links
			case 'LINK':
				$ret .= $this->makeInternalLink($parser);
				break;
			case 'LINKTARGET':
			case 'LINKOPTION':
				$ret .= $this->sub_makeXHTML($parser);
				break;

			case 'TEMPLATE':