/gforge/plugins/wiki/www/lib/stdlib.php

Large files files are truncated, but you can click here to view the full file

<?php //rcs_id('$Id: stdlib.php,v 1.251 2006/03/19 15:01:00 rurban Exp $');
/*
 Copyright 1999,2000,2001,2002,2004,2005 $ThePhpWikiProgrammingTeam

 This file is part of PhpWiki.

 PhpWiki is free software; you can redistribute it and/or modify
 it under the terms of the GNU General Public License as published by
 the Free Software Foundation; either version 2 of the License, or
 (at your option) any later version.

 PhpWiki is distributed in the hope that it will be useful,
 but WITHOUT ANY WARRANTY; without even the implied warranty of
 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 GNU General Public License for more details.

 You should have received a copy of the GNU General Public License
 along with PhpWiki; if not, write to the Free Software
 Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

/*
  Standard functions for Wiki functionality
    WikiURL ($pagename, $args, $get_abs_url)
    AbsoluteURL ($url)
    IconForLink ($protocol_or_url)
    PossiblyGlueIconToText($proto_or_url, $text)
    IsSafeURL($url)
    LinkURL ($url, $linktext)
    LinkImage ($url, $alt)

    SplitQueryArgs ($query_args)
    LinkPhpwikiURL ($url, $text, $basepage)
    ConvertOldMarkup ($content, $markup_type = "block")
    MangleXmlIdentifier($str)
    UnMangleXmlIdentifier($str)
    
    class Stack { push($item), pop(), cnt(), top() }
    class Alert { show() }
    class WikiPageName {getParent(),isValid(),getWarnings() }

    expand_tabs($str, $tab_width = 8)
    SplitPagename ($page)
    NoSuchRevision ($request, $page, $version)
    TimezoneOffset ($time, $no_colon)
    Iso8601DateTime ($time)
    Rfc2822DateTime ($time)
    ParseRfc1123DateTime ($timestr)
    CTime ($time)
    ByteFormatter ($bytes = 0, $longformat = false)
    __printf ($fmt)
    __sprintf ($fmt)
    __vsprintf ($fmt, $args)

    file_mtime ($filename)
    sort_file_mtime ($a, $b)
    class fileSet {fileSet($directory, $filepattern = false), 
                   getFiles($exclude=false, $sortby=false, $limit=false) }
    class ListRegexExpand { listMatchCallback($item, $key),  
                            expandRegex ($index, &$pages) }

    glob_to_pcre ($glob)
    glob_match ($glob, $against, $case_sensitive = true)
    explodeList ($input, $allnames, $glob_style = true, $case_sensitive = true)
    explodePageList ($input, $perm = false)
    isa ($object, $class)
    can ($object, $method)
    function_usable ($function_name)
    wikihash ($x)
    better_srand ($seed = '')
    count_all ($arg)
    isSubPage ($pagename)
    subPageSlice ($pagename, $pos)

    phpwiki_version ()
    isWikiWord ($word)
    obj2hash ($obj, $exclude = false, $fields = false)
    isUtf8String ($s)
    fixTitleEncoding ($s)
    url_get_contents ($uri)
    GenerateId ($name)
    firstNWordsOfContent ($n, $content)
    extractSection ($section, $content, $page, $quiet = false, $sectionhead = false)
    isExternalReferrer()

  function: LinkInterWikiLink($link, $linktext)
  moved to: lib/interwiki.php
  function: linkExistingWikiWord($wikiword, $linktext, $version)
  moved to: lib/Theme.php
  function: LinkUnknownWikiWord($wikiword, $linktext)
  moved to: lib/Theme.php
  function: UpdateRecentChanges($dbi, $pagename, $isnewpage) 
  gone see: lib/plugin/RecentChanges.php
*/
if (defined('_PHPWIKI_STDLIB_LOADED')) return;
else define('_PHPWIKI_STDLIB_LOADED', true);

define('MAX_PAGENAME_LENGTH', 100);
            
/**
 * Convert string to a valid XML identifier.
 *
 * XML 1.0 identifiers are of the form: [A-Za-z][A-Za-z0-9:_.-]*
 *
 * We would like to have, e.g. named anchors within wiki pages
 * names like "Table of Contents" --- clearly not a valid XML
 * fragment identifier.
 *
 * This function implements a one-to-one map from {any string}
 * to {valid XML identifiers}.
 *
 * It does this by
 * converting all bytes not in [A-Za-z0-9:_-],
 * and any leading byte not in [A-Za-z] to 'xbb.',
 * where 'bb' is the hexadecimal representation of the
 * character.
 *
 * As a special case, the empty string is converted to 'empty.'
 *
 * @param string $str
 * @return string
 */
function MangleXmlIdentifier($str) {
    if (!$str)
        return 'empty.';
    
    return preg_replace('/[^-_:A-Za-z0-9]|(?<=^)[^A-Za-z]/e',
                        "'x' . sprintf('%02x', ord('\\0')) . '.'",
                        $str);
}

function UnMangleXmlIdentifier($str) {
    if ($str == 'empty.')
        return '';
    return preg_replace('/x(\w\w)\./e',
                        "sprintf('%c', hex('\\0'))",
                        $str);
}

/**
 * Returns a name for the WIKI_ID cookie that should be unique on the host.
 * But for it to be unique you must have set a unique WIKI_NAME in your
 * configuration file.
 * @return string The name of the WIKI_ID cookie to use for this wiki.
 */
function GetCookieName() {
    return preg_replace("/[^\d\w]/", "_", WIKI_NAME) . "_WIKI_ID";
}

/**
 * Generates a valid URL for a given Wiki pagename.
 * @param mixed $pagename If a string this will be the name of the Wiki page to link to.
 * 			  If a WikiDB_Page object function will extract the name to link to.
 * 			  If a WikiDB_PageRevision object function will extract the name to link to.
 * @param array $args 
 * @param boolean $get_abs_url Default value is false.
 * @return string The absolute URL to the page passed as $pagename.
 */
function WikiURL($pagename, $args = '', $get_abs_url = false) {
    $anchor = false;
    
    if (is_object($pagename)) {
        if (isa($pagename, 'WikiDB_Page')) {
            $pagename = $pagename->getName();
        }
        elseif (isa($pagename, 'WikiDB_PageRevision')) {
            $page = $pagename->getPage();
            $args['version'] = $pagename->getVersion();
            $pagename = $page->getName();
        }
        elseif (isa($pagename, 'WikiPageName')) {
            $anchor = $pagename->anchor;
            $pagename = $pagename->name;
        } else { // php5
            $anchor = $pagename->anchor;
            $pagename = $pagename->name;
        }
    }
    if (!$get_abs_url and DEBUG and $GLOBALS['request']->getArg('start_debug')) {
    	if (!$args)
    	    $args = 'start_debug=' . $GLOBALS['request']->getArg('start_debug');
    	elseif (is_array($args))
    	    $args['start_debug'] = $GLOBALS['request']->getArg('start_debug');
    	else 
    	    $args .= '&amp;start_debug=' . $GLOBALS['request']->getArg('start_debug');
    }
    if (is_array($args)) {
        $enc_args = array();
        foreach ($args as $key => $val) {
            // avoid default args
            if (USE_PATH_INFO and $key == 'pagename')
            	; 
            elseif ($key == 'action' and $val == 'browse')
        	;
            elseif (!is_array($val)) // ugly hack for getURLtoSelf() which also takes POST vars
              $enc_args[] = urlencode($key) . '=' . urlencode($val);
        }
        $args = join('&', $enc_args);
    }

    if (USE_PATH_INFO or !empty($GLOBALS['WikiTheme']->HTML_DUMP_SUFFIX)) {
        $url = $get_abs_url ? (SERVER_URL . VIRTUAL_PATH . "/") : "";
        $url = $url . preg_replace('/%2f/i', '/', rawurlencode($pagename));
        if (!empty($GLOBALS['WikiTheme']->HTML_DUMP_SUFFIX))
            $url .= $GLOBALS['WikiTheme']->HTML_DUMP_SUFFIX;
        if ($args)
            $url .= "?$args";
    }
    else {
        $url = $get_abs_url ? SERVER_URL . SCRIPT_NAME : basename(SCRIPT_NAME);
        $url .= "?pagename=" . rawurlencode($pagename);
        if ($args)
            $url .= "&$args";
    }
    if ($anchor)
        $url .= "#" . MangleXmlIdentifier($anchor);
    return $url;
}

/** Convert relative URL to absolute URL.
 *
 * This converts a relative URL to one of PhpWiki's support files
 * to an absolute one.
 *
 * @param string $url
 * @return string Absolute URL
 */
function AbsoluteURL ($url) {
    if (preg_match('/^https?:/', $url))
        return $url;
    if ($url[0] != '/') {
        $base = USE_PATH_INFO ? VIRTUAL_PATH : dirname(SCRIPT_NAME);
        while ($base != '/' and substr($url, 0, 3) == "../") {
            $url = substr($url, 3);
            $base = dirname($base);
        }
        if ($base != '/')
            $base .= '/';
        $url = $base . $url;
    }
    return SERVER_URL . $url;
}

function DataURL ($url) {
    if (preg_match('/^https?:/', $url))
        return $url;
    $url = NormalizeWebFileName($url);
    if (DEBUG and $GLOBALS['request']->getArg('start_debug') and substr($url,-4,4) == '.php')
        $url .= "?start_debug=1"; // XMLRPC and SOAP debugging helper.
    return AbsoluteURL($url);
}

/**
 * Generates icon in front of links.
 *
 * @param string $protocol_or_url URL or protocol to determine which icon to use.
 *
 * @return HtmlElement HtmlElement object that contains data to create img link to
 * icon for use with url or protocol passed to the function. False if no img to be
 * displayed.
 */
function IconForLink($protocol_or_url) {
    global $WikiTheme;
    if (0 and $filename_suffix == false) {
        // display apache style icon for file type instead of protocol icon
        // - archive: unix:gz,bz2,tgz,tar,z; mac:dmg,dmgz,bin,img,cpt,sit; pc:zip;
        // - document: html, htm, text, txt, rtf, pdf, doc
        // - non-inlined image: jpg,jpeg,png,gif,tiff,tif,swf,pict,psd,eps,ps
        // - audio: mp3,mp2,aiff,aif,au
        // - multimedia: mpeg,mpg,mov,qt
    } else {
        list ($proto) = explode(':', $protocol_or_url, 2);
        $src = $WikiTheme->getLinkIconURL($proto);
        if ($src)
            return HTML::img(array('src' => $src, 'alt' => "", 'class' => 'linkicon', 'border' => 0));
        else
            return false;
    }
}

/**
 * Glue icon in front of or after text.
 * Pref: 'noLinkIcons'  - ignore icon if set
 * Theme: 'LinkIcons'   - 'yes'   at front
 *                      - 'no'    display no icon
 *                      - 'front' display at left
 *                      - 'after' display at right
 *
 * @param string $protocol_or_url Protocol or URL.  Used to determine the
 * proper icon.
 * @param string $text The text.
 * @return XmlContent.
 */
function PossiblyGlueIconToText($proto_or_url, $text) {
    global $request, $WikiTheme;
    if ($request->getPref('noLinkIcons'))
        return $text;
    $icon = IconForLink($proto_or_url);
    if (!$icon)
        return $text;
    if ($where = $WikiTheme->getLinkIconAttr()) {
        if ($where == 'no') return $text;
        if ($where != 'after') $where = 'front';
    } else {
        $where = 'front';
    }
    if ($where == 'after') {
        // span the icon only to the last word (tie them together), 
        // to let the previous words wrap on line breaks.
        if (!is_object($text)) {
            preg_match('/^(\s*\S*)(\s*)$/', $text, $m);
            list (, $prefix, $last_word) = $m;
        }
        else {
            $last_word = $text;
            $prefix = false;
        }
        $text = HTML::span(array('style' => 'white-space: nowrap'),
                           $last_word, HTML::Raw('&nbsp;'), $icon);
        if ($prefix)
            $text = HTML($prefix, $text);
        return $text;
    }
    // span the icon only to the first word (tie them together), 
    // to let the next words wrap on line breaks
    if (!is_object($text)) {
        preg_match('/^\s*(\S*)(.*?)\s*$/', $text, $m);
        list (, $first_word, $tail) = $m;
    }
    else {
        $first_word = $text;
        $tail = false;
    }
    $text = HTML::span(array('style' => 'white-space: nowrap'),
                       $icon, $first_word);
    if ($tail)
        $text = HTML($text, $tail);
    return $text;
}

/**
 * Determines if the url passed to function is safe, by detecting if the characters
 * '<', '>', or '"' are present.
 * Check against their urlencoded values also.
 *
 * @param string $url URL to check for unsafe characters.
 * @return boolean True if same, false else.
 */
function IsSafeURL($url) {
    return !preg_match('/([<>"])|(%3C)|(%3E)|(%22)/', $url);
}

/**
 * Generates an HtmlElement object to store data for a link.
 *
 * @param string $url URL that the link will point to.
 * @param string $linktext Text to be displayed as link.
 * @return HtmlElement HtmlElement object that contains data to construct an html link.
 */
function LinkURL($url, $linktext = '') {
    // FIXME: Is this needed (or sufficient?)
    if(! IsSafeURL($url)) {
        $link = HTML::strong(HTML::u(array('class' => 'baduri'),
                                     _("BAD URL -- remove all of <, >, \"")));
    }
    else {
        if (!$linktext)
            $linktext = preg_replace("/mailto:/A", "", $url);
        $args = array('href' => $url);
        if ( defined('EXTERNAL_LINK_TARGET') ) // can also be set in the css
            $args['target'] = is_string(EXTERNAL_LINK_TARGET) ? EXTERNAL_LINK_TARGET : "_blank";
        $link = HTML::a($args, PossiblyGlueIconToText($url, $linktext));
    }
    $link->setAttr('class', $linktext ? 'namedurl' : 'rawurl');
    return $link;
}

/**
 * Inline Images
 *
 * Syntax: [image.png size=50% border=n align= hspace= vspace= width= height=]
 * Disallows sizes which are too small. 
 * Spammers may use such (typically invisible) image attributes to higher their GoogleRank.
 *
 * Handle embeddable objects, like svg, class, vrml, swf, svgz, pdf, avi, wmv especially.
 */
function LinkImage($url, $alt = false) {
    $force_img = "png|jpg|gif|jpeg|bmp|pl|cgi";
    // Disallow tags in img src urls. Typical CSS attacks.
    // FIXME: Is this needed (or sufficient?)
    if(! IsSafeURL($url)) {
        $link = HTML::strong(HTML::u(array('class' => 'baduri'),
                                     _("BAD URL -- remove all of <, >, \"")));
    } else {
        // support new syntax: [image.jpg size=50% border=n]
        if (!preg_match("/\.(".$force_img.")/i", $url))
            $ori_url = $url;
        $arr = split(' ',$url);
        if (count($arr) > 1) {
            $url = $arr[0];
        }
        if (empty($alt)) $alt = basename($url);
        $link = HTML::img(array('src' => $url, 'alt' => $alt, 'title' => $alt));
        if (count($arr) > 1) {
            array_shift($arr);
            foreach ($arr as $attr) {
                if (preg_match('/^size=(\d+%)$/',$attr,$m)) {
                    $link->setAttr('width',$m[1]);
                    $link->setAttr('height',$m[1]);
                }
                if (preg_match('/^size=(\d+)x(\d+)$/',$attr,$m)) {
                    $link->setAttr('width',$m[1]);
                    $link->setAttr('height',$m[2]);
                }
                if (preg_match('/^border=(\d+)$/',$attr,$m))
                    $link->setAttr('border',$m[1]);
                if (preg_match('/^align=(\w+)$/',$attr,$m))
                    $link->setAttr('align',$m[1]);
                if (preg_match('/^hspace=(\d+)$/',$attr,$m))
                    $link->setAttr('hspace',$m[1]);
                if (preg_match('/^vspace=(\d+)$/',$attr,$m))
                    $link->setAttr('vspace',$m[1]);
            }
        }
        // Check width and height as spam countermeasure
        if (($width  = $link->getAttr('width')) and ($height = $link->getAttr('height'))) {
            //$width  = (int) $width; // px or % or other suffix
            //$height = (int) $height;
            if (($width < 3 and $height < 10) or 
                ($height < 3 and $width < 20) or 
                ($height < 7 and $width < 7))
            {
                trigger_error(_("Invalid image size"), E_USER_WARNING);
                return '';
            }
        } else {
            // Older php versions crash here with certain png's: 
            // confirmed for 4.1.2, 4.1.3, 4.2.3; 4.3.2 and 4.3.7 are ok
            //   http://phpwiki.sourceforge.net/demo/themes/default/images/http.png
            // See http://bugs.php.net/search.php?cmd=display&search_for=getimagesize
            if (!check_php_version(4,3) and preg_match("/^http.+\.png$/i",$url))
                ; // it's safe to assume that this will fail.
            elseif (!DISABLE_GETIMAGESIZE and ($size = @getimagesize($url))) {
                $width  = $size[0];
                $height = $size[1];
                if (($width < 3 and $height < 10) 
                    or ($height < 3 and $width < 20)
                    or ($height < 7 and $width < 7))
                {
                    trigger_error(_("Invalid image size"), E_USER_WARNING);
                    return '';
                }
            }
        }
    }
    $link->setAttr('class', 'inlineimage');

    /* Check for inlined objects. Everything allowed in INLINE_IMAGES besides
     * png|jpg|gif|jpeg|bmp|pl|cgi
     * Note: Allow cgi's (pl,cgi) returning images.
     */
    if (!preg_match("/\.(".$force_img.")/i", $url)) {
        //HTML::img(array('src' => $url, 'alt' => $alt, 'title' => $alt));
        // => HTML::object(array('src' => $url)) ...;
        return ImgObject($link, $ori_url);
    }
    return $link;
}

/**
 * <object> / <embed> tags instead of <img> for all non-image extensions allowed via INLINE_IMAGES
 * Called by LinkImage(), not directly.
 * Syntax: [image.svg size=50% border=n align= hspace= vspace= width= height=]
 * $alt may be an alternate img
 * TODO: Need to unify with WikiPluginCached::embedObject()
 *
 * Note that Safari 1.0 will crash with <object>, so use only <embed>
 *   http://www.alleged.org.uk/pdc/2002/svg-object.html
 */
function ImgObject($img, $url) {
    // get the url args: data="sample.svgz" type="image/svg+xml" width="400" height="300"
    $args = split(' ', $url);
    if (count($args) >= 1) {
        $url = array_shift($args);
        foreach ($args as $attr) {
            if (preg_match('/^type=(\S+)$/',$attr,$m))
                $img->setAttr('type', $m[1]);
            if (preg_match('/^data=(\S+)$/',$attr,$m))
                $img->setAttr('data', $m[1]);
        }
    }
    $type = $img->getAttr('type');
    if (!$type) {
        // TODO: map extension to mime-types if type is not given and php < 4.3
        if (function_exists('mime_content_type'))
            $type = mime_content_type($url);
    }
    $link = HTML::object(array_merge($img->_attr, array('src' => $url, 'type' => $type)));
    $link->setAttr('class', 'inlineobject');
    if (isBrowserSafari()) {
        return HTML::embed($link->_attr);
    }
    $link->pushContent(HTML::embed($link->_attr));
    return $link;
}


class Stack {

    // var in php5 deprecated
    function Stack() {
        $this->items = array();
        $this->size = 0;
    }
    function push($item) {
        $this->items[$this->size] = $item;
        $this->size++;
        return true;
    }  
    
    function pop() {
        if ($this->size == 0) {
            return false; // stack is empty
        }  
        $this->size--;
        return $this->items[$this->size];
    }  
    
    function cnt() {
        return $this->size;
    }  
    
    function top() {
        if($this->size)
            return $this->items[$this->size - 1];
        else
            return '';
    }
    
}  
// end class definition

function SplitQueryArgs ($query_args = '') 
{
    // FIXME: use the arg-seperator which might not be &
    $split_args = split('&', $query_args);
    $args = array();
    while (list($key, $val) = each($split_args))
        if (preg_match('/^ ([^=]+) =? (.*) /x', $val, $m))
            $args[$m[1]] = $m[2];
    return $args;
}

function LinkPhpwikiURL($url, $text = '', $basepage = false) {
    $args = array();
    
    if (!preg_match('/^ phpwiki: ([^?]*) [?]? (.*) $/x', $url, $m)) {
        return HTML::strong(array('class' => 'rawurl'),
                            HTML::u(array('class' => 'baduri'),
                                    _("BAD phpwiki: URL")));
    }

    if ($m[1])
        $pagename = urldecode($m[1]);
    $qargs = $m[2];
    
    if (empty($pagename) &&
        preg_match('/^(diff|edit|links|info)=([^&]+)$/', $qargs, $m)) {
        // Convert old style links (to not break diff links in
        // RecentChanges).
        $pagename = urldecode($m[2]);
        $args = array("action" => $m[1]);
    }
    else {
        $args = SplitQueryArgs($qargs);
    }

    if (empty($pagename))
        $pagename = $GLOBALS['request']->getArg('pagename');

    if (isset($args['action']) && $args['action'] == 'browse')
        unset($args['action']);
    
    /*FIXME:
      if (empty($args['action']))
      $class = 'wikilink';
      else if (is_safe_action($args['action']))
      $class = 'wikiaction';
    */
    if (empty($args['action']) || is_safe_action($args['action']))
        $class = 'wikiaction';
    else {
        // Don't allow administrative links on unlocked pages.
        $dbi = $GLOBALS['request']->getDbh();
        $page = $dbi->getPage($basepage ? $basepage : $pagename);
        if (!$page->get('locked'))
            return HTML::span(array('class' => 'wikiunsafe'),
                              HTML::u(_("Lock page to enable link")));
        $class = 'wikiadmin';
    }
    
    if (!$text)
        $text = HTML::span(array('class' => 'rawurl'), $url);

    $wikipage = new WikiPageName($pagename);
    if (!$wikipage->isValid()) {
        global $WikiTheme;
        return $WikiTheme->linkBadWikiWord($wikipage, $url);
    }
    
    return HTML::a(array('href'  => WikiURL($pagename, $args),
                         'class' => $class),
                   $text);
}

/**
 * A class to assist in parsing wiki pagenames.
 *
 * Now with subpages and anchors, parsing and passing around
 * pagenames is more complicated.  This should help.
 */
class WikiPageName
{
    /** Short name for page.
     *
     * This is the value of $name passed to the constructor.
     * (For use, e.g. as a default label for links to the page.)
     */
    //var $shortName;

    /** The full page name.
     *
     * This is the full name of the page (without anchor).
     */
    //var $name;
    
    /** The anchor.
     *
     * This is the referenced anchor within the page, or the empty string.
     */
    //var $anchor;
    
    /** Constructor
     *
     * @param mixed $name Page name.
     * WikiDB_Page, WikiDB_PageRevision, or string.
     * This can be a relative subpage name (like '/SubPage'),
     * or can be the empty string to refer to the $basename.
     *
     * @param string $anchor For links to anchors in page.
     *
     * @param mixed $basename Page name from which to interpret
     * relative or other non-fully-specified page names.
     */
    function WikiPageName($name, $basename=false, $anchor=false) {
        if (is_string($name)) {
            $this->shortName = $name;
            if (strstr($name, ':')) {
                list($moniker, $this->shortName) = split (":", $name, 2);
	  	$map = getInterwikiMap(); // allow overrides to custom maps
                if (isset($map->_map[$moniker])) {
                    $url = $map->_map[$moniker];
                    if (strstr($url, '%s'))
                        $url = sprintf($url, $this->shortName);
                    else
                        $url .= $this->shortName;
                    // expand Talk or User, but not to absolute urls!
                    if (strstr($url, '//')) {
                        if ($moniker == 'Talk')
                            $name = $name . SUBPAGE_SEPARATOR . _("Discussion");
                        elseif ($moniker == 'User')
                            $name = $name;
                    } else {
                        $name = $url;
                    }
                    if (strstr($name, '?'))
                        list($name, $dummy) = split("?", $name, 2);
                }
            }
	    // FIXME: We should really fix the cause for "/PageName" in the WikiDB
            if ($name == '' or $name[0] == SUBPAGE_SEPARATOR) {
                if ($basename)
                    $name = $this->_pagename($basename) . $name;
                else {
                    $name = $this->_normalize_bad_pagename($name);
		    $this->shortName = $name;
                }
            }
        }
        else {
            $name = $this->_pagename($name);
            $this->shortName = $name;
        }

        $this->name = $this->_check($name);
        $this->anchor = (string)$anchor;
    }

    function getName() {
    	return $this->name;
    }
    
    function getParent() {
        $name = $this->name;
        if (!($tail = strrchr($name, SUBPAGE_SEPARATOR)))
            return false;
        return substr($name, 0, -strlen($tail));
    }

    function isValid($strict = false) {
        if ($strict)
            return !isset($this->_errors);
        return (is_string($this->name) and $this->name != '');
    }

    function getWarnings() {
        $warnings = array();
        if (isset($this->_warnings))
            $warnings = array_merge($warnings, $this->_warnings);
        if (isset($this->_errors))
            $warnings = array_merge($warnings, $this->_errors);
        if (!$warnings)
            return false;
        
        return sprintf(_("'%s': Bad page name: %s"),
                       $this->shortName, join(', ', $warnings));
    }
    
    function _pagename($page) {
	if (isa($page, 'WikiDB_Page'))
	    return $page->getName();
        elseif (isa($page, 'WikiDB_PageRevision'))
	    return $page->getPageName();
        elseif (isa($page, 'WikiPageName'))
	    return $page->name;
        // '0' or e.g. '1984' should be allowed though
        if (!is_string($page) and !is_integer($page)) {
            trigger_error(sprintf("Non-string pagename '%s' (%s)(%s)",
                                  $page, gettype($page), get_class($page)),
                          E_USER_NOTICE);
        }
	//assert(is_string($page));
	return $page;
    }

    function _normalize_bad_pagename($name) {
        trigger_error("Bad pagename: " . $name, E_USER_WARNING);

        // Punt...  You really shouldn't get here.
        if (empty($name)) {
            global $request;
            return $request->getArg('pagename');
        }
        assert($name[0] == SUBPAGE_SEPARATOR);
        $this->_errors[] = sprintf(_("Leading %s not allowed"), SUBPAGE_SEPARATOR);
        return substr($name, 1);
    }


    function _check($pagename) {
        // Compress internal white-space to single space character.
        $pagename = preg_replace('/[\s\xa0]+/', ' ', $orig = $pagename);
        if ($pagename != $orig)
            $this->_warnings[] = _("White space converted to single space");
    
        // Delete any control characters.
        if (DATABASE_TYPE == 'cvs' or DATABASE_TYPE == 'file') {
            $pagename = preg_replace('/[\x00-\x1f\x7f\x80-\x9f]/', '', $orig = $pagename);
            if ($pagename != $orig)
                $this->_errors[] = _("Control characters not allowed");
        }

        // Strip leading and trailing white-space.
        $pagename = trim($pagename);

        $orig = $pagename;
        while ($pagename and $pagename[0] == SUBPAGE_SEPARATOR)
            $pagename = substr($pagename, 1);
        if ($pagename != $orig)
            $this->_errors[] = sprintf(_("Leading %s not allowed"), SUBPAGE_SEPARATOR);

        // ";" is urlencoded, so safe from php arg-delim problems
        /*if (strstr($pagename, ';')) {
            $this->_warnings[] = _("';' is deprecated");
            $pagename = str_replace(';', '', $pagename);
        }*/
        
        // not only for SQL, also to restrict url length
        if (strlen($pagename) > MAX_PAGENAME_LENGTH) {
            $pagename = substr($pagename, 0, MAX_PAGENAME_LENGTH);
            $this->_errors[] = _("too long");
        }

        // disallow some chars only on file and cvs
        if ((DATABASE_TYPE == 'cvs' or DATABASE_TYPE == 'file') 
            and preg_match('/(:|\.\.)/', $pagename, $m)) {
            $this->_warnings[] = sprintf(_("Illegal chars %s removed"), $m[1]);
            $pagename = str_replace('..', '', $pagename);
            $pagename = str_replace(':', '', $pagename);
        }
        
        return $pagename;
    }
}

/**
 * Convert old page markup to new-style markup.
 *
 * @param string $text Old-style wiki markup.
 *
 * @param string $markup_type
 * One of: <dl>
 * <dt><code>"block"</code>  <dd>Convert all markup.
 * <dt><code>"inline"</code> <dd>Convert only inline markup.
 * <dt><code>"links"</code>  <dd>Convert only link markup.
 * </dl>
 *
 * @return string New-style wiki markup.
 *
 * @bugs Footnotes don't work quite as before (esp if there are
 *   multiple references to the same footnote.  But close enough,
 *   probably for now....
 * @bugs  Apache2 and IIS crash with OldTextFormattingRules or
 *   AnciennesR%E8glesDeFormatage. (at the 2nd attempt to do the anchored block regex)
 *   It only crashes with CreateToc so far, but other pages (not in pgsrc) are 
 *   also known to crash, even with Apache1.
 */
function ConvertOldMarkup ($text, $markup_type = "block") {

    static $subs;
    static $block_re;
    
    // FIXME:
    // Trying to detect why the 2nd paragraph of OldTextFormattingRules or
    // AnciennesR%E8glesDeFormatage crashes. 
    // It only crashes with CreateToc so far, but other pages (not in pgsrc) are 
    // also known to crash, even with Apache1.
    $debug_skip = false;
    // I suspect this only to crash with Apache2 and IIS.
    if (in_array(php_sapi_name(),array('apache2handler','apache2filter','isapi'))
        and preg_match("/plugin CreateToc/", $text)) 
    {
    	trigger_error(_("The CreateTocPlugin is not yet old markup compatible! ")
    	             ._("Please remove the CreateToc line to be able to reformat this page to old markup. ")
    	             ._("Skipped."), E_USER_WARNING);
        $debug_skip = true;
        //if (!DEBUG) return $text;
        return $text;
    }

    if (empty($subs)) {
        /*****************************************************************
         * Conversions for inline markup:
         */

        // escape tilde's
        $orig[] = '/~/';
        $repl[] = '~~';

        // escape escaped brackets
        $orig[] = '/\[\[/';
        $repl[] = '~[';

        // change ! escapes to ~'s.
        global $WikiNameRegexp, $request;
        $bang_esc[] = "(?:" . ALLOWED_PROTOCOLS . "):[^\s<>\[\]\"'()]*[^\s<>\[\]\"'(),.?]";
        // before 4.3.9 pcre had a memory release bug, which might hit us here. so be safe.
        if (check_php_version(4,3,9)) {
          $map = getInterwikiMap();
          if ($map_regex = $map->getRegexp())
            $bang_esc[] = $map_regex . ":[^\\s.,;?()]+"; // FIXME: is this really needed?
        }
        $bang_esc[] = $WikiNameRegexp;
        $orig[] = '/!((?:' . join(')|(', $bang_esc) . '))/';
        $repl[] = '~\\1';

        $subs["links"] = array($orig, $repl);

        // Temporarily URL-encode pairs of underscores in links to hide
        // them from the re for bold markup.
        $orig[] = '/\[[^\[\]]*?__[^\[\]]*?\]/e';
        $repl[] = 'str_replace(\'__\', \'%5F%5F\', \'\\0\')';

        // Escape '<'s
        //$orig[] = '/<(?!\?plugin)|(?<!^)</m';
        //$repl[] = '~<';
        
        // Convert footnote references.
        $orig[] = '/(?<=.)(?<!~)\[\s*(\d+)\s*\]/m';
        $repl[] = '#[|ftnt_ref_\\1]<sup>~[[\\1|#ftnt_\\1]~]</sup>';

        // Convert old style emphases to HTML style emphasis.
        $orig[] = '/__(.*?)__/';
        $repl[] = '<strong>\\1</strong>';
        $orig[] = "/''(.*?)''/";
        $repl[] = '<em>\\1</em>';

        // Escape nestled markup.
        $orig[] = '/^(?<=^|\s)[=_](?=\S)|(?<=\S)[=_*](?=\s|$)/m';
        $repl[] = '~\\0';
        
        // in old markup headings only allowed at beginning of line
        $orig[] = '/!/';
        $repl[] = '~!';

        // Convert URL-encoded pairs of underscores in links back to
        // real underscores after bold markup has been converted.
        $orig = '/\[[^\[\]]*?%5F%5F[^\[\]]*?\]/e';
        $repl = 'str_replace(\'%5F%5F\', \'__\', \'\\0\')';

        $subs["inline"] = array($orig, $repl);

        /*****************************************************************
         * Patterns which match block markup constructs which take
         * special handling...
         */

        // Indented blocks
        $blockpats[] = '[ \t]+\S(?:.*\s*\n[ \t]+\S)*';
        // Tables
        $blockpats[] = '\|(?:.*\n\|)*';

        // List items
        $blockpats[] = '[#*;]*(?:[*#]|;.*?:)';

        // Footnote definitions
        $blockpats[] = '\[\s*(\d+)\s*\]';

        if (!$debug_skip) {
        // Plugins
        $blockpats[] = '<\?plugin(?:-form)?\b.*\?>\s*$';
        }

        // Section Title
        $blockpats[] = '!{1,3}[^!]';
        /*
	removed .|\n in the anchor not to crash on /m because with /m "." already includes \n
	this breaks headings but it doesn't crash anymore (crash on non-cgi, non-cli only)
	*/
        $block_re = ( '/\A((?:.|\n)*?)(^(?:'
                      . join("|", $blockpats)
                      . ').*$)\n?/m' );
        
    }
    
    if ($markup_type != "block") {
        list ($orig, $repl) = $subs[$markup_type];
        return preg_replace($orig, $repl, $text);
    }
    else {
        list ($orig, $repl) = $subs['inline'];
        $out = '';
	//FIXME:
	// php crashes here in the 2nd paragraph of OldTextFormattingRules, 
	// AnciennesR%E8glesDeFormatage and more 
	// See http://www.pcre.org/pcre.txt LIMITATIONS
	 while (preg_match($block_re, $text, $m)) {
            $text = substr($text, strlen($m[0]));
            list (,$leading_text, $block) = $m;
            $suffix = "\n";
            
            if (strchr(" \t", $block[0])) {
                // Indented block
                $prefix = "<pre>\n";
                $suffix = "\n</pre>\n";
            }
            elseif ($block[0] == '|') {
                // Old-style table
                $prefix = "<?plugin OldStyleTable\n";
                $suffix = "\n?>\n";
            }
            elseif (strchr("#*;", $block[0])) {
                // Old-style list item
                preg_match('/^([#*;]*)([*#]|;.*?:) */', $block, $m);
                list (,$ind,$bullet) = $m;
                $block = substr($block, strlen($m[0]));
                
                $indent = str_repeat('     ', strlen($ind));
                if ($bullet[0] == ';') {
                    //$term = ltrim(substr($bullet, 1));
                    //return $indent . $term . "\n" . $indent . '     ';
                    $prefix = $ind . $bullet;
                }
                else
                    $prefix = $indent . $bullet . ' ';
            }
            elseif ($block[0] == '[') {
                // Footnote definition
                preg_match('/^\[\s*(\d+)\s*\]/', $block, $m);
                $footnum = $m[1];
                $block = substr($block, strlen($m[0]));
                $prefix = "#[|ftnt_${footnum}]~[[${footnum}|#ftnt_ref_${footnum}]~] ";
            }
            elseif ($block[0] == '<') {
                // Plugin.
                // HACK: no inline markup...
                $prefix = $block;
                $block = '';
            }
            elseif ($block[0] == '!') {
                // Section heading
                preg_match('/^!{1,3}/', $block, $m);
                $prefix = $m[0];
                $block = substr($block, strlen($m[0]));
            }
            else {
                // AAck!
                assert(0);
            }
            if ($leading_text) $leading_text = preg_replace($orig, $repl, $leading_text);
            if ($block) $block = preg_replace($orig, $repl, $block);
            $out .= $leading_text;
            $out .= $prefix;
            $out .= $block;
            $out .= $suffix;
        }
        return $out . preg_replace($orig, $repl, $text);
    }
}


/**
 * Expand tabs in string.
 *
 * Converts all tabs to (the appropriate number of) spaces.
 *
 * @param string $str
 * @param integer $tab_width
 * @return string
 */
function expand_tabs($str, $tab_width = 8) {
    $split = split("\t", $str);
    $tail = array_pop($split);
    $expanded = "\n";
    foreach ($split as $hunk) {
        $expanded .= $hunk;
        $pos = strlen(strrchr($expanded, "\n")) - 1;
        $expanded .= str_repeat(" ", ($tab_width - $pos % $tab_width));
    }
    return substr($expanded, 1) . $tail;
}

/**
 * Split WikiWords in page names.
 *
 * It has been deemed useful to split WikiWords (into "Wiki Words") in
 * places like page titles. This is rumored to help search engines
 * quite a bit.
 *
 * @param $page string The page name.
 *
 * @return string The split name.
 */
function SplitPagename ($page) {
    
    if (preg_match("/\s/", $page))
        return $page;           // Already split --- don't split any more.
    
    // This algorithm is specialized for several languages.
    // (Thanks to Pierrick MEIGNEN)
    // Improvements for other languages welcome.
    static $RE;
    if (!isset($RE)) {
        // This mess splits between a lower-case letter followed by
        // either an upper-case or a numeral; except that it wont
        // split the prefixes 'Mc', 'De', or 'Di' off of their tails.
        switch ($GLOBALS['LANG']) {
        case 'en':
        case 'it':
        case 'es': 
        case 'de':
            $RE[] = '/([[:lower:]])((?<!Mc|De|Di)[[:upper:]]|\d)/';
            break;
        case 'fr': 
            $RE[] = '/([[:lower:]])((?<!Mc|Di)[[:upper:]]|\d)/';
            break;
        }
	$sep = preg_quote(SUBPAGE_SEPARATOR, '/');
        // This the single-letter words 'I' and 'A' from any following
        // capitalized words.
        switch ($GLOBALS['LANG']) {
        case 'en': 
            $RE[] = "/(?<= |${sep}|^)([AI])([[:upper:]][[:lower:]])/";
            break;
        case 'fr': 
            $RE[] = "/(?<= |${sep}|^)([À])([[:upper:]][[:lower:]])/";
            break;
        }
        // Split numerals from following letters.
        $RE[] = '/(\d)([[:alpha:]])/';
        // Split at subpage seperators. TBD in Theme.php
        $RE[] = "/([^${sep}]+)(${sep})/";
        
        foreach ($RE as $key)
            $RE[$key] = pcre_fix_posix_classes($key);
    }

    foreach ($RE as $regexp) {
	$page = preg_replace($regexp, '\\1 \\2', $page);
    }
    return $page;
}

function NoSuchRevision (&$request, $page, $version) {
    $html = HTML(HTML::h2(_("Revision Not Found")),
                 HTML::p(fmt("I'm sorry.  Version %d of %s is not in the database.",
                             $version, WikiLink($page, 'auto'))));
    include_once('lib/Template.php');
    GeneratePage($html, _("Bad Version"), $page->getCurrentRevision());
    $request->finish();
}


/**
 * Get time offset for local time zone.
 *
 * @param $time time_t Get offset for this time. Default: now.
 * @param $no_colon boolean Don't put colon between hours and minutes.
 * @return string Offset as a string in the format +HH:MM.
 */
function TimezoneOffset ($time = false, $no_colon = false) {
    if ($time === false)
        $time = time();
    $secs = date('Z', $time);

    if ($secs < 0) {
        $sign = '-';
        $secs = -$secs;
    }
    else {
        $sign = '+';
    }
    $colon = $no_colon ? '' : ':';
    $mins = intval(($secs + 30) / 60);
    return sprintf("%s%02d%s%02d",
                   $sign, $mins / 60, $colon, $mins % 60);
}


/**
 * Format time in ISO-8601 format.
 *
 * @param $time time_t Time.  Default: now.
 * @return string Date and time in ISO-8601 format.
 */
function Iso8601DateTime ($time = false) {
    if ($time === false)
        $time = time();
    $tzoff = TimezoneOffset($time);
    $date  = date('Y-m-d', $time);
    $time  = date('H:i:s', $time);
    return $date . 'T' . $time . $tzoff;
}

/**
 * Format time in RFC-2822 format.
 *
 * @param $time time_t Time.  Default: now.
 * @return string Date and time in RFC-2822 format.
 */
function Rfc2822DateTime ($time = false) {
    if ($time === false)
        $time = time();
    return date('D, j M Y H:i:s ', $time) . TimezoneOffset($time, 'no colon');
}

/**
 * Format time in RFC-1123 format.
 *
 * @param $time time_t Time.  Default: now.
 * @return string Date and time in RFC-1123 format.
 */
function Rfc1123DateTime ($time = false) {
    if ($time === false)
        $time = time();
    return gmdate('D, d M Y H:i:s \G\M\T', $time);
}

/** Parse date in RFC-1123 format.
 *
 * According to RFC 1123 we must accept dates in the following
 * formats:
 *
 *   Sun, 06 Nov 1994 08:49:37 GMT  ; RFC 822, updated by RFC 1123
 *   Sunday, 06-Nov-94 08:49:37 GMT ; RFC 850, obsoleted by RFC 1036
 *   Sun Nov  6 08:49:37 1994       ; ANSI C's asctime() format
 *
 * (Though we're only allowed to generate dates in the first format.)
 */
function ParseRfc1123DateTime ($timestr) {
    $timestr = trim($timestr);
    if (preg_match('/^ \w{3},\s* (\d{1,2}) \s* (\w{3}) \s* (\d{4}) \s*'
                   .'(\d\d):(\d\d):(\d\d) \s* GMT $/ix',
                   $timestr, $m)) {
        list(, $mday, $mon, $year, $hh, $mm, $ss) = $m;
    }
    elseif (preg_match('/^ \w+,\s* (\d{1,2})-(\w{3})-(\d{2}|\d{4}) \s*'
                       .'(\d\d):(\d\d):(\d\d) \s* GMT $/ix',
                       $timestr, $m)) {
        list(, $mday, $mon, $year, $hh, $mm, $ss) = $m;
        if ($year < 70) $year += 2000;
        elseif ($year < 100) $year += 1900;
    }
    elseif (preg_match('/^\w+\s* (\w{3}) \s* (\d{1,2}) \s*'
                       .'(\d\d):(\d\d):(\d\d) \s* (\d{4})$/ix',
                       $timestr, $m)) {
        list(, $mon, $mday, $hh, $mm, $ss, $year) = $m;
    }
    else {
        // Parse failed.
        return false;
    }

    $time = strtotime("$mday $mon $year ${hh}:${mm}:${ss} GMT");
    if ($time == -1)
        return false;           // failed
    return $time;
}

/**
 * Format time to standard 'ctime' format.
 *
 * @param $time time_t Time.  Default: now.
 * @return string Date and time.
 */
function CTime ($time = false)
{
    if ($time === false)
        $time = time();
    return date("D M j H:i:s Y", $time);
}


/**
 * Format number as kilobytes or bytes.
 * Short format is used for PageList
 * Long format is used in PageInfo
 *
 * @param $bytes       int.  Default: 0.
 * @param $longformat  bool. Default: false.
 * @return class FormattedText (XmlElement.php).
 */
function ByteFormatter ($bytes = 0, $longformat = false) {
    if ($bytes < 0)
        return fmt("-???");
    if ($bytes < 1024) {
        if (! $longformat)
            $size = fmt("%s b", $bytes);
        else
            $size = fmt("%s bytes", $bytes);
    }
    else {
        $kb = round($bytes / 1024, 1);
        if (! $longformat)
            $size = fmt("%s k", $kb);
        else
            $size = fmt("%s Kb (%s bytes)", $kb, $bytes);
    }
    return $size;
}

/**
 * Internationalized printf.
 *
 * This is essentially the same as PHP's built-in printf
 * with the following exceptions:
 * <ol>
 * <li> It passes the format string through gettext().
 * <li> It supports the argument reordering extensions.
 * </ol>
 *
 * Example:
 *
 * In php code, use:
 * <pre>
 *    __printf("Differences between versions %s and %s of %s",
 *             $new_link, $old_link, $page_link);
 * </pre>
 *
 * Then in locale/po/de.po, one can reorder the printf arguments:
 *
 * <pre>
 *    msgid "Differences between %s and %s of %s."
 *    msgstr "Der Unterschiedsergebnis von %3$s, zwischen %1$s und %2$s."
 * </pre>
 *
 * (Note that while PHP tries to expand $vars within double-quotes,
 * the values in msgstr undergo no such expansion, so the '$'s
 * okay...)
 *
 * One shouldn't use reordered arguments in the default format string.
 * Backslashes in the default string would be necessary to escape the
 * '$'s, and they'll cause all kinds of trouble....
 */ 
function __printf ($fmt) {
    $args = func_get_args();
    array_shift($args);
    echo __vsprintf($fmt, $args);
}

/**
 * Internationalized sprintf.
 *
 * This is essentially the same as PHP's built-in printf with the
 * following exceptions:
 *
 * <ol>
 * <li> It passes the format string through gettext().
 * <li> It supports the argument reordering extensions.
 * </ol>
 *
 * @see __printf
 */ 
function __sprintf ($fmt) {
    $args = func_get_args();
    array_shift($args);
    return __vsprintf($fmt, $args);
}

/**
 * Internationalized vsprintf.
 *
 * This is essentially the same as PHP's built-in printf with the
 * following exceptions:
 *
 * <ol>
 * <li> It passes the format string through gettext().
 * <li> It supports the argument reordering extensions.
 * </ol>
 *
 * @see __printf
 */ 
function __vsprintf ($fmt, $args) {
    $fmt = gettext($fmt);
    // PHP's sprintf doesn't support variable with specifiers,
    // like sprintf("%*s", 10, "x"); --- so we won't either.
    
    if (preg_match_all('/(?<!%)%(\d+)\$/x', $fmt, $m)) {
        // Format string has '%2$s' style argument reordering.
        // PHP doesn't support this.
        if (preg_match('/(?<!%)%[- ]?\d*[^- \d$]/x', $fmt))
            // literal variable name substitution only to keep locale
            // strings uncluttered
            trigger_error(sprintf(_("Can't mix '%s' with '%s' type format strings"),
                                  '%1\$s','%s'), E_USER_WARNING); //php+locale error
        
        $fmt = preg_replace('/(?<!%)%\d+\$/x', '%', $fmt);
        $newargs = array();
        
        // Reorder arguments appropriately.
        foreach($m[1] as $argnum) {
            if ($argnum < 1 || $argnum > count($args))
                trigger_error(sprintf(_("%s: argument index out of range"), 
                                      $argnum), E_USER_WARNING);
            $newargs[] = $args[$argnum - 1];
        }
        $args = $newargs;
    }
    
    // Not all PHP's have vsprintf, so...
    array_unshift($args, $fmt);
    return call_user_func_array('sprintf', $args);
}

function file_mtime ($filename) {
    if ($stat = @stat($filename))
        return $stat[9];
    else 
        return false;
}

function sort_file_mtime ($a, $b) {
    $ma = file_mtime($a);
    $mb = file_mtime($b);
    if (!$ma or !$mb or $ma == $mb) return 0;
    return ($ma > $mb) ? -1 : 1;
}

class fileSet {
    /**
     * Build an array in $this->_fileList of files from $dirname.
     * Subdirectories are not traversed.
     *
     * (This was a function LoadDir in lib/loadsave.php)
     * See also http://www.php.net/manual/en/function.readdir.php
     */
    function getFiles($exclude=false, $sortby=false, $limit=false) {
        $list = $this->_fileList;

        if ($sortby) {
            require_once('lib/PageList.php');
            switch (Pagelist::sortby($sortby, 'db')) {
            case 'pagename ASC': break;
            case 'pagename DESC': 
                $list = array_reverse($list); 
                break;
            case 'mtime ASC': 
                usort($list,'sort_file_mtime'); 
                break;
            case 'mtime DESC': 
                usort($list,'sort_file_mtime');
                $list = array_reverse($list); 
                break;
            }
        }
        if ($limit)
            return array_splice($list, 0, $limit);
        return $list;
    }

    function _filenameSelector($filename) {
        if (! $this->_pattern )
            return true;
        else {
            if (! $this->_pcre_pattern )
                $this->_pcre_pattern = glob_to_pcre($this->_pattern);
            return preg_match('/' . $this->_pcre_pattern . ($this->_case ? '/' : '/i'), 
                              $filename);
        }
    }

    function fileSet($directory, $filepattern = false) {
        $this->_fileList = array();
        $this->_pattern = $filepattern;
        if ($filepattern)
            $this->_pcre_pattern = glob_to_pcre($this->_pattern);
        $this->_case = !isWindows();
        $this->_pathsep = '/';

        if (empty($directory)) {
            trigger_error(sprintf(_("%s is empty."), 'directoryname'),
                          E_USER_NOTICE);
            return; // early return
        }

        @ $dir_handle = opendir($dir=$directory);
        if (empty($dir_handle)) {
            trigger_error(sprintf(_("Unable to open directory '%s' for reading"),
                                  $dir), E_USER_NOTICE);
            return; // early return
        }

        while ($filename = readdir($dir_handle)) {
            if ($filename[0] == '.' || filetype($dir . $this->_pathsep . $filename) != 'file')
                continue;
            if ($this->_filenameSelector($filename)) {
                array_push($this->_fileList, "$filename");
                //trigger_error(sprintf(_("found file %s"), $filename),
                //                      E_USER_NOTICE); //debugging
            }
        }
        closedir($dir_handle);
    }
};

// File globbing

// expands a list containing regex's to its matching entries
class ListRegexExpand {
    //var $match, $list, $index, $case_sensitive;
    function ListRegexExpand (&$list, $match, $case_sensitive = true) {
    	$this->match = $match;
    	$this->list = &$list;
    	$this->case_sensitive = $case_sensitive;	
        //$this->index = false;
    }
    function listMatchCallback ($item, $key) {
        $quoted = str_replace('/','\/',$item);
    	if (preg_match('/' . $this->match . ($this->case_sensitive ? '/' : '/i'), 
                       $quoted)) {
	    unset($this->list[$this->index]);
            $this->list[] = $item;
        }
    }
    function expandRegex ($index, &$pages) {
    	$this->index = $index;
    	array_walk($pages, array($this, 'listMatchCallback'));
        return $this->list;
    }
}

// Convert fileglob to regex style:
// Convert some wildcards to pcre style, escape the rest
// Escape . \\ + * ? [ ^ ] $ ( ) { } = ! < > | : /
// Fixed bug #994994: "/" in $glob.
function glob_to_pcre ($glob) {
    // check simple case: no need to escape
    $escape = '\[](){}=!<>|:/';
    if (strcspn($glob, $escape . ".+*?^$") == strlen($glob))
        r…