Gazelle/classes/text.class.php

<?php
#declare(strict_types=1);

class Text
{
    /**
     * Array of valid tags; tag => max number of attributes
     * @var array $ValidTags
     */
    private static $ValidTags = [
      'b' => 0,
      'u' => 0,
      'i' => 0,
      's' => 0,
      '*' => 0,
      '#' => 0,
      #'ch' => 0,
      #'uch' => 0,
      'artist' => 0,
      'user' => 0,
      'n' => 0,
      'inlineurl' => 0,
      'inlinesize' => 1,
      'headline' => 1,
      'align' => 1,
      'color' => 1,
      'colour' => 1,
      'size' => 1,
      'url' => 1,
      'img' => 1,
      'quote' => 1,
      'pre' => 1,
      'code' => 1,
      'tex' => 0,
      'hide' => 1,
      'spoiler' => 1,
      'plain' => 0,
      'important' => 0,
      'torrent' => 0,
      'rule' => 0,
      'embed' => 0,
    ];


    /**
     * Array of smilies; code => image file in STATIC_SERVER/common/smileys
     * @var array $Smileys
     */
    private static $Smileys = [];


    /**
     * Processed version of the $Smileys array, see {@link smileys}
     * @var array $ProcessedSmileys
     */
    private static $ProcessedSmileys = [];


    /**
     * Whether or not to turn images into URLs (used inside [quote] tags).
     * This is an integer reflecting the number of levels we're doing that
     * transition, i.e. images will only be displayed as images if $NoImg <= 0.
     * By setting this variable to a negative number you can delay the
     * transition to a deeper level of quotes.
     * @var int $NoImg
     */
    private static $NoImg = 0;


    /**
     * Internal counter for the level of recursion in to_html
     * @var int $Levels
     */
    private static $Levels = 0;


    /**
     * The maximum amount of nesting allowed (exclusive)
     * In reality n-1 nests are shown.
     * @var int $MaximumNests
     */
    private static $MaximumNests = 10;


    /**
     * Used to detect and disable parsing (e.g. TOC) within quotes
     * @var int $InQuotes
     */
    private static $InQuotes = 0;


    /**
     * Used to [hide] quote trains starting with the specified depth (inclusive)
     * @var int $NestsBeforeHide
     *
     * This defaulted to 5 but was raised to 10 to effectively "disable" it until
     * an optimal number of nested [quote] tags is chosen. The variable $MaximumNests
     * effectively overrides this variable, if $MaximumNests is less than the value
     * of $NestsBeforeHide.
     */
    private static $NestsBeforeHide = 10;


    /**
     * Array of headlines for Table of Contents (TOC)
     * @var array $HeadLines
     */
    private static $Headlines;


    /**
     * Counter for making headline URLs unique
     * @var int $HeadLines
     */
    private static $HeadlineID = 0;


    /**
     * Depth
     * @var array $HeadlineLevels
     */
    private static $HeadlineLevels = array('1', '2', '3', '4');


    /**
     * TOC enabler
     * @var bool $TOC
     */
    public static $TOC = false;

    
    /**
     * Output BBCode as XHTML
     *
     * This used to suck and now it only blows.
     * If $Str isn't BBcode, call Parsedown in safe mode.
     * Otherwise, continue with the default parser.
     *
     * @param string $Str BBCode text
     * @param bool $OutputTOC Ouput TOC near (above) text
     * @param int $Min See {@link parse_toc}
     * @return string
     */
    public static function full_format($Str, $OutputTOC = true, $Min = 3)
    {
        $ENV = ENV::go();

        # Here's the magic pattern:
        if (!preg_match(
            "/$ENV->BBCODE_REGEX/s",
            $Str
        )) {
            $Parsedown = new ParsedownExtra();
            $Parsedown->setSafeMode(true);

            # Prepare clean escapes
            $Str = html_entity_decode($Str, ENT_QUOTES | ENT_HTML5, 'UTF-8');

            # Parse early and post-process
            $Parsed = $Parsedown->text($Str);
            
            # Replace links to $ENV->SITE_DOMAIN
            $Parsed = preg_replace(
                "/<a href=\"$ENV->RESOURCE_REGEX$ENV->SITE_DOMAIN\//",
                '<a href="/',
                $Parsed
            );
                
            # Replace external links and add Wikipedia-style class
            $Parsed = preg_replace(
                '/<a href="https?:\/\//',
                '<a class="external" rel="noopener" rel="noreferrer" href="https://',
                $Parsed
            );

            $Parsed = preg_replace(
                '/<a href="ftps?:\/\//',
                '<a class="external" rel="noopener" rel="noreferrer" href="ftps://',
                $Parsed
            );

            return $Parsed;

            # Markdown ToC not happening yet
            # Shouldn't parse_toc() output HTML
            /*
            self::$Headlines = [];
            if (self::$TOC) {
                $List = [];

                foreach (count(1, 6) as $x) {
                    array_push(
                        $List,
                        preg_replace(
                            "/(\#{$x})([^#].*)\1/i",
                            "<h$x>$2</h$x>",
                            $Str
                        )
                    );
                }
            }
            */

        /*
        return $P =
            ((self::$TOC && $OutputTOC)
                ? self::parse_toc($Min)
                : null)
            . $Parsedown->text($Str);
        */
        } else {
            global $Debug;
            $Debug->set_flag('BBCode start');

            self::$Headlines = [];
            $Str = display_str($Str);

            # Checkboxes: broken and stupid
            /*
            $Str = preg_replace('/\[\\[(ch|uch)]\]/i', '', $Str);
            $Str = preg_replace('/\[ch\]/i', '[ch][/ch]', $Str);
            $Str = preg_replace('/\[uch\]/i', '[uch][/uch]', $Str);
            */

            // Inline links
            $URLPrefix = '(\[url\]|\[url\=|\[img\=|\[img\])';
            $Str = preg_replace('/'.$URLPrefix.'\s+/i', '$1', $Str);
            $Str = preg_replace('/(?<!'.$URLPrefix.')http(s)?:\/\//i', '$1[inlineurl]http$2://', $Str);
            $Str = preg_replace('/\[embed\]\[inlineurl\]/', '[embed]', $Str);

            // For anonym.to and archive.org links, remove any [inlineurl] in the middle of the link
            /*
            $Str = preg_replace_callback(
                '/(?<=\[inlineurl\]|'.$URLPrefix.')(\S*\[inlineurl\]\S*)/m',
                function ($matches) {
                    return str_replace("[inlineurl]", "", $matches[0]);
                },
                $Str
            );
            */

            if (self::$TOC) {
                $Str = preg_replace('/(\={5})([^=].*)\1/i', '[headline=4]$2[/headline]', $Str);
                $Str = preg_replace('/(\={4})([^=].*)\1/i', '[headline=3]$2[/headline]', $Str);
                $Str = preg_replace('/(\={3})([^=].*)\1/i', '[headline=2]$2[/headline]', $Str);
                $Str = preg_replace('/(\={2})([^=].*)\1/i', '[headline=1]$2[/headline]', $Str);
            } else {
                $Str = preg_replace('/(\={4})([^=].*)\1/i', '[inlinesize=3]$2[/inlinesize]', $Str);
                $Str = preg_replace('/(\={3})([^=].*)\1/i', '[inlinesize=5]$2[/inlinesize]', $Str);
                $Str = preg_replace('/(\={2})([^=].*)\1/i', '[inlinesize=7]$2[/inlinesize]', $Str);
            }

            # nl2br() is necessary even though it results in "squashed" BBcode source
            # (it unintelligently converts multiple new lines to multiple line breaks)
            $HTML = nl2br(self::to_html(self::parse($Str)));

            if (self::$TOC && $OutputTOC) {
                $HTML = self::parse_toc($Min) . $HTML;
            }

            $Debug->set_flag('BBCode end');
            return $HTML;
        }
    }


    /**
     * strip_bbcode()
     */
    public static function strip_bbcode($Str)
    {
        $Str = display_str($Str);

        // Inline links
        $Str = preg_replace(
            '/(?<!(\[url\]|\[url\=|\[img\=|\[img\]))http(s)?:\/\//i',
            '$1[inlineurl]http$2://',
            $Str
        );

        return nl2br(self::raw_text(self::parse($Str)));
    }


    /**
     * valid_url()
     */
    private static function valid_url($Str, $Extension = '', $Inline = false)
    {
        $Regex = '/^';
        $Regex .= '(https?|ftps?|irc):\/\/'; // protocol
        $Regex .= '(\w+(:\w+)?@)?'; // user:pass@
        $Regex .= '(';
        $Regex .= '(([0-9]{1,3}\.){3}[0-9]{1,3})|'; // IP or...
        $Regex .= '(([a-z0-9\-\_]+\.)+\w{2,6})'; // sub.sub.sub.host.com
        $Regex .= ')';
        $Regex .= '(:[0-9]{1,5})?'; // port
        $Regex .= '\/?'; // slash?
        $Regex .= '(\/?[0-9a-z\-_.,\?&=@~%\/:;()+|!#]+)*'; // /file

        if (!empty($Extension)) {
            $Regex.=$Extension;
        }

        // Query string
        if ($Inline) {
            $Regex .= '(\?([0-9a-z\-_.,%\/\@~&=:;()+*\^$!#|?]|\[\d*\])*)?';
        } else {
            $Regex .= '(\?[0-9a-z\-_.,%\/\@[\]~&=:;()+*\^$!#|?]*)?';
        }

        $Regex .= '(#[a-z0-9\-_.,%\/\@[\]~&=:;()+*\^$!]*)?'; // #anchor
        $Regex .= '$/i';

        return preg_match($Regex, $Str, $Matches);
    }


    /**
     * local_url()
     */
    public static function local_url($Str)
    {
        $URLInfo = parse_url($Str);
        if (!$URLInfo) {
            return false;
        }
        $Host = $URLInfo['host'];

        // If for some reason your site does not require subdomains
        // or contains a directory in the SITE_DOMAIN,
        // revert to the line below.
        if ($Host === SITE_DOMAIN || $Host === 'www.'.SITE_DOMAIN) {
            if (empty($URLInfo['port']) && preg_match('/(\S+\.)*'.SITE_DOMAIN.'/', $Host)) {
                $URL = '';

                // Things break if the path starts with '//'
                if (!empty($URLInfo['path'])) {
                    $URL .= ltrim($URLInfo['path'], '/');
                }

                if (!empty($URLInfo['query'])) {
                    $URL .= "?$URLInfo[query]";
                }

                if (!empty($URLInfo['fragment'])) {
                    $URL .= "#$URLInfo[fragment]";
                }
                return $URL ? "/$URL" : false;
            } else {
                return false;
            }
        }
    }

    
    /**
     * How parsing works
     *
     * Parsing takes $Str, breaks it into blocks, and builds it into $Array.
     * Blocks start at the beginning of $Str, when the parser encounters a [, and after a tag has been closed.
     * This is all done in a loop.
     *
     *
     * EXPLANATION OF PARSER LOGIC
     *
     * 1) Find the next tag (regex).
     *   1a) If there aren't any tags left, write everything remaining to a block and return (done parsing).
     *   1b) If the next tag isn't where the pointer is, write everything up to there to a text block.
     *
     * 2) See if it's a [[wiki-link]] or an ordinary tag, and get the tag name.
     *
     * 3) If it's not a wiki link:
     *   3a) Check it against the self::$ValidTags array to see if it's actually a tag and not [bullshit].
     *       If it's [not a tag], just leave it as plaintext and move on.
     *   3b) Get the attribute, if it exists [name=attribute].
     *
     * 4) Move the pointer past the end of the tag.
     *
     * 5) Find out where the tag closes (beginning of [/tag]).
     *   5a) Different for different types of tag. Some tags don't close, others are weird like [*].
     *   5b) If it's a normal tag, it may have versions of itself nested inside, e.g.,
     *       [quote=bob]*
     *         [quote=joe]I am a redneck!**[/quote]
     *         Me too!
     *       ***[/quote]
     *
     *       If we're at the position *, the first [/quote] tag is denoted by **.
     *       However, our quote tag doesn't actually close there.
     *       We must perform a loop which checks the number of opening [quote] tags,
     *       and make sure they are all closed before we find our final [/quote] tag (***).
     *
     *   5c) Get the contents between [open] and [/close] and call it the block.
     *       In many cases, this will be parsed itself later on, in a new parse() call.
     *   5d) Move the pointer past the end of the [/close] tag.
     *
     * 6) Depending on what type of tag we're dealing with, create an array with the attribute and block.
     *    In many cases, the block may be parsed here itself. Stick them in the $Array.
     *
     * 7) Increment array pointer, start again (past the end of the [/close] tag).
     */
    private static function parse($Str)
    {
        $i = 0; // Pointer to keep track of where we are in $Str
        $Len = strlen($Str);
        $Array = [];
        $ArrayPos = 0;
        $StrLC = strtolower($Str);

        while ($i < $Len) {
            $Block = '';

            // 1) Find the next tag (regex)
            // [name(=attribute)?]|[[wiki-link]]
            $IsTag = preg_match(
                "/((\[[a-zA-Z*#]+)(=(?:[^\n'\"\[\]]|\[\d*\])+)?\])|(\[\[[^\n\"'\[\]]+\]\])/",
                $Str,
                $Tag,
                PREG_OFFSET_CAPTURE,
                $i
            );

            // 1a) If there aren't any tags left, write everything remaining to a block
            if (!$IsTag) {
                // No more tags
                $Array[$ArrayPos] = substr($Str, $i);
                break;
            }

            // 1b) If the next tag isn't where the pointer is, write everything up to there to a text block.
            $TagPos = $Tag[0][1];
            if ($TagPos > $i) {
                $Array[$ArrayPos] = substr($Str, $i, $TagPos - $i);
                ++$ArrayPos;
                $i = $TagPos;
            }

            // 2) See if it's a [[wiki-link]] or an ordinary tag, and get the tag name
            if (!empty($Tag[4][0])) { // Wiki-link
                $WikiLink = true;
                $TagName = substr($Tag[4][0], 2, -2);
                $Attrib = '';
            } else { // 3) If it's not a wiki link:
                $WikiLink = false;
                $TagName = strtolower(substr($Tag[2][0], 1));

                // 3a) check it against the self::$ValidTags array to see if it's actually a tag and not [bullshit]
                if (!isset(self::$ValidTags[$TagName])) {
                    $Array[$ArrayPos] = substr($Str, $i, ($TagPos - $i) + strlen($Tag[0][0]));
                    $i = $TagPos + strlen($Tag[0][0]);
                    ++$ArrayPos;
                    continue;
                }

                $MaxAttribs = self::$ValidTags[$TagName];

                // 3b) Get the attribute, if it exists [name=attribute]
                if (!empty($Tag[3][0])) {
                    $Attrib = substr($Tag[3][0], 1);
                } else {
                    $Attrib = '';
                }
            }

            // 4) Move the pointer past the end of the tag
            $i = $TagPos + strlen($Tag[0][0]);

            // 5) Find out where the tag closes (beginning of [/tag])

            // Unfortunately, BBCode doesn't have nice standards like XHTML
            // [*], [img=...], and http:// follow different formats
            // Thus, we have to handle these before we handle the majority of tags

            // 5a) Different for different types of tag. Some tags don't close, others are weird like [*]
            if ($TagName === 'img' && !empty($Tag[3][0])) { //[img=...]
                $Block = ''; // Nothing inside this tag
                // Don't need to touch $i
            } elseif ($TagName === 'inlineurl') { // We did a big replace early on to turn http:// into [inlineurl]http://

                // Let's say the block can stop at a newline or a space
                $CloseTag = strcspn($Str, " \n\r", $i);
                if ($CloseTag === false) { // Block finishes with URL
                    $CloseTag = $Len;
                }

                if (preg_match('/[!,.?:]+$/', substr($Str, $i, $CloseTag), $Match)) {
                    $CloseTag -= strlen($Match[0]);
                }

                $URL = substr($Str, $i, $CloseTag);
                if (substr($URL, -1) === ')' && substr_count($URL, '(') < substr_count($URL, ')')) {
                    $CloseTag--;
                    $URL = substr($URL, 0, -1);
                }
                $Block = $URL; // Get the URL

                // strcspn returns the number of characters after the offset $i, not after the beginning of the string
                // Therefore, we use += instead of the = everywhere else
                $i += $CloseTag; // 5d) Move the pointer past the end of the [/close] tag.
            } elseif ($WikiLink === true || $TagName === 'n') {
                // Don't need to do anything - empty tag with no closing
            } elseif ($TagName === '*' || $TagName === '#') {
                // We're in a list. Find where it ends
                $NewLine = $i;

                do { // Look for \n[*]
                    $NewLine = strpos($Str, "\n", $NewLine + 1);
                } while ($NewLine !== false && substr($Str, $NewLine + 1, 3) === "[$TagName]");

                $CloseTag = $NewLine;
                if ($CloseTag === false) { // Block finishes with list
                    $CloseTag = $Len;
                }

                $Block = substr($Str, $i, $CloseTag - $i); // Get the list
                $i = $CloseTag; // 5d) Move the pointer past the end of the [/close] tag.
            } else {
                // 5b) If it's a normal tag, it may have versions of itself nested inside
                $CloseTag = $i - 1;
                $InTagPos = $i - 1;
                $NumInOpens = 0;
                $NumInCloses = -1;

                $InOpenRegex = '/\[('.$TagName.')';
                if ($MaxAttribs > 0) {
                    $InOpenRegex .= "(=[^\n'\"\[\]]+)?";
                }
                $InOpenRegex .= '\]/i';

                // Every time we find an internal open tag of the same type, search for the next close tag
                // (as the first close tag won't do - it's been opened again)
                do {
                    $CloseTag = strpos($StrLC, "[/$TagName]", $CloseTag + 1);
                    if ($CloseTag === false) {
                        $CloseTag = $Len;
                        break;
                    } else {
                        $NumInCloses++; // Majority of cases
                    }

                    // Is there another open tag inside this one?
                    $OpenTag = preg_match($InOpenRegex, $Str, $InTag, PREG_OFFSET_CAPTURE, $InTagPos + 1);
                    if (!$OpenTag || $InTag[0][1] > $CloseTag) {
                        break;
                    } else {
                        $InTagPos = $InTag[0][1];
                        $NumInOpens++;
                    }
                } while ($NumInOpens > $NumInCloses);

                // Find the internal block inside the tag
                $Block = substr($Str, $i, $CloseTag - $i); // 5c) Get the contents between [open] and [/close] and call it the block
                $i = $CloseTag + strlen($TagName) + 3; // 5d) Move the pointer past the end of the [/close] tag
            }

            // 6) Depending on what type of tag we're dealing with, create an array with the attribute and block
            switch ($TagName) {
            case 'inlineurl':
              $Array[$ArrayPos] = array('Type' => 'inlineurl', 'Attr' => $Block, 'Val' => '');
              break;


            case 'url':
              $Array[$ArrayPos] = array('Type' => 'img', 'Attr' => $Attrib, 'Val' => $Block);
              if (empty($Attrib)) { // [url]http://...[/url] - always set URL to attribute
                  $Array[$ArrayPos] = array('Type' => 'url', 'Attr' => $Block, 'Val' => '');
              } else {
                  $Array[$ArrayPos] = array('Type' => 'url', 'Attr' => $Attrib, 'Val' => self::parse($Block));
              }
              break;


            case 'quote':
              $Array[$ArrayPos] = array('Type' => 'quote', 'Attr' => self::parse($Attrib), 'Val' => self::parse($Block));
              break;


            case 'img':
            case 'image':
              if (empty($Block)) {
                  $Block = $Attrib;
              }

              $Array[$ArrayPos] = array('Type' => 'img', 'Val' => $Block);
              break;


            case 'aud':
            case 'mp3':
            case 'audio':
              if (empty($Block)) {
                  $Block = $Attrib;
              }

              $Array[$ArrayPos] = array('Type' => 'aud', 'Val' => $Block);
              break;


            case 'user':
              $Array[$ArrayPos] = array('Type' => 'user', 'Val' => $Block);
              break;


            case 'artist':
              $Array[$ArrayPos] = array('Type' => 'artist', 'Val' => $Block);
              break;


            case 'torrent':
              $Array[$ArrayPos] = array('Type' => 'torrent', 'Val' => $Block);
              break;


            case 'tex':
              $Array[$ArrayPos] = array('Type' => 'tex', 'Val' => $Block);
              break;


            case 'rule':
              $Array[$ArrayPos] = array('Type' => 'rule', 'Val' => $Block);
              break;


            case 'pre':
            case 'code':
            case 'plain':
              $Block = strtr($Block, array('[inlineurl]' => ''));

              $Callback = function ($matches) {
                  $n = $matches[2];
                  $text = '';

                  if ($n < 5 && $n > 0) {
                      $e = str_repeat('=', $matches[2] + 1);
                      $text = $e . $matches[3] . $e;
                  }
                  return $text;
              };

              $Block = preg_replace_callback('/\[(headline)\=(\d)\](.*?)\[\/\1\]/i', $Callback, $Block);
              $Block = preg_replace('/\[inlinesize\=3\](.*?)\[\/inlinesize\]/i', '====$1====', $Block);
              $Block = preg_replace('/\[inlinesize\=5\](.*?)\[\/inlinesize\]/i', '===$1===', $Block);
              $Block = preg_replace('/\[inlinesize\=7\](.*?)\[\/inlinesize\]/i', '==$1==', $Block);
              $Array[$ArrayPos] = array('Type' => $TagName, 'Val' => $Block);
              break;
          

            case 'spoiler':
            case 'hide':
              $Array[$ArrayPos] = array('Type' => 'hide', 'Attr' => $Attrib, 'Val' => self::parse($Block));
              break;


            case 'embed':
              $Array[$ArrayPos] = array('Type' => 'embed', 'Val' => $Block);
              break;


            case '#':
            case '*':
              $Array[$ArrayPos] = array('Type' => 'list');
              $Array[$ArrayPos]['Val'] = explode("[$TagName]", $Block);
              $Array[$ArrayPos]['ListType'] = $TagName === '*' ? 'ul' : 'ol';
              $Array[$ArrayPos]['Tag'] = $TagName;

              foreach ($Array[$ArrayPos]['Val'] as $Key => $Val) {
                  $Array[$ArrayPos]['Val'][$Key] = self::parse(trim($Val));
              }
              break;


            case 'n':
              $ArrayPos--;
              break; // n serves only to disrupt bbcode (backwards compatibility - use [pre])


            default:
              if ($WikiLink === true) {
                  $Array[$ArrayPos] = array('Type' => 'wiki','Val' => $TagName);
              } else {
                  // Basic tags, like [b] or [size=5]
                  $Array[$ArrayPos] = array('Type' => $TagName, 'Val' => self::parse($Block));
                  if (!empty($Attrib) && $MaxAttribs > 0) {
                      $Array[$ArrayPos]['Attr'] = strtolower($Attrib);
                  }
              }
            }

            $ArrayPos++; // 7) Increment array pointer, start again (past the end of the [/close] tag)
        }

        return $Array;
    }


    /**
     * Generates a navigation list for TOC
     * @param int $Min Minimum number of headlines required for a TOC list
     */
    public static function parse_toc($Min = 3)
    {
        if (count(self::$Headlines) > $Min) {
            $list = '<ol class="navigation_list">';
            $i = 0;
            $level = 0;
            $off = 0;

            foreach (self::$Headlines as $t) {
                $n = (int) $t[0];
                if ($i === 0 && $n > 1) {
                    $off = $n - $level;
                }

                self::headline_level($n, $level, $list, $i, $off);
                $list .= sprintf('<li><a href="#%2$s">%1$s</a>', $t[1], $t[2]);
                $level = $t[0];
                $off = 0;
                $i++;
            }

            $list .= str_repeat('</li></ol>', $level);
            $list .= "\n\n";
            return $list;
        }
    }


    /**
     * Generates the list items and proper depth
     *
     * First check if the item should be higher than the current level
     *  - Close the list and previous lists
     *
     * Then check if the item should go lower than the current level
     *  - If the list doesn't open on level one, use the Offset
     *  - Open appropriate sub lists
     *
     * Otherwise the item is on the same as level as the previous item
     *
     * @param int $ItemLevel Current item level
     * @param int $Level Current list level
     * @param str $List reference to an XHTML string
     * @param int $i Iterator digit
     * @param int $Offset If the list doesn't start at level 1
     */
    private static function headline_level(&$ItemLevel, &$Level, &$List, $i, &$Offset)
    {
        if ($ItemLevel < $Level) {
            $diff = $Level - $ItemLevel;
            $List .= '</li>' . str_repeat('</ol></li>', $diff);
        } elseif ($ItemLevel > $Level) {
            $diff = $ItemLevel - $Level;
            if ($Offset > 0) {
                $List .= str_repeat('<li><ol>', $Offset - 2);
            }

            if ($ItemLevel > 1) {
                $List .= $i === 0 ? '<li>' : '';
                $List .= "\n<ol>\n";
            }
        } else {
            $List .= $i > 0 ? '</li>' : '<li>';
        }
    }


    /*
     * Hax prevention
     *
     * That's the original comment on this.
     * Most likely this was implemented to avoid anyone nesting enough elements
     * to reach PHP's memory limit as nested elements are solved recursively.
     * Original value of 10, it is now replaced in favor of $MaximumNests.
     *
     * If this line is ever executed then something is, in fact, being haxed.
     * The if before the block type switch for different tags should always be limiting ahead of this line.
     */
    private static function to_html($Array)
    {
        self::$Levels++;

        // Hax prevention, breaks upon exceeding nests
        if (self::$Levels > self::$MaximumNests) {
            return $Block['Val'];
        }

        $Str = '';
        foreach ($Array as $Block) {
            if (is_string($Block)) {
                $Str .= self::smileys($Block);
                continue;
            }

            if (self::$Levels < self::$MaximumNests) {
                switch ($Block['Type']) {
                case 'b':
                  $Str .= '<strong>'.self::to_html($Block['Val']).'</strong>';
                  break;


                case 'u':
                  $Str .= '<span style="text-decoration: underline;">'.self::to_html($Block['Val']).'</span>';
                  break;


                case 'i':
                  $Str .= '<span style="font-style: italic;">'.self::to_html($Block['Val'])."</span>";
                  break;


                case 's':
                  $Str .= '<span style="text-decoration: line-through;">'.self::to_html($Block['Val']).'</span>';
                  break;


                case 'important':
                  $Str .= '<strong class="important_text">'.self::to_html($Block['Val']).'</strong>';
                  break;


                case 'user':
                  $Str .= '<a href="user.php?action=search&amp;search='.urlencode($Block['Val']).'">'.$Block['Val'].'</a>';
                  break;


                case 'artist':
                  $Str .= '<a href="artist.php?artistname='.urlencode(Format::undisplay_str($Block['Val'])).'">'.$Block['Val'].'</a>';
                  break;


                case 'rule':
                  $Rule = trim(strtolower($Block['Val']));
                  if ($Rule[0] !== 'r' && $Rule[0] !== 'h') {
                      $Rule = 'r'.$Rule;
                  }

                  $Str .= '<a href="rules.php?p=upload#'.urlencode(Format::undisplay_str($Rule)).'">'.preg_replace('/[aA-zZ]/', '', $Block['Val']).'</a>';
                  break;

                case 'torrent':
                  $Pattern = '/('.SITE_DOMAIN.'\/torrents\.php.*[\?&]id=)?(\d+)($|&|\#).*/i';
                  $Matches = [];

                  if (preg_match($Pattern, $Block['Val'], $Matches)) {
                      if (isset($Matches[2])) {
                          $GroupID = $Matches[2];
                          $Groups = Torrents::get_groups(array($GroupID), true, true, false);

                          if ($Groups[$GroupID]) {
                              $Group = $Groups[$GroupID];
                              $Str .= Artists::display_artists($Group['Artists']).'<a href="torrents.php?id='.$GroupID;

                              if (preg_match('/torrentid=(\d+)/i', $Block['Val'], $Matches)) {
                                  $Str .= '&torrentid='.$Matches[1];
                              }

                              $Str .= '"';
                              if (!isset($LoggedUser['CoverArt']) || $LoggedUser['CoverArt']) {
                                  $Str .= ' data-cover="'.ImageTools::process($Group['WikiImage'], 'thumb').'"';
                              }

                              $Name = empty($Group['Name']) ? (empty($Group['Title2']) ? $Group['NameJP'] : $Group['Title2']) : $Group['Name'];
                              $Str .= '>'.$Name.'</a>';
                          } else {
                              $Str .= '[torrent]'.str_replace('[inlineurl]', '', $Block['Val']).'[/torrent]';
                          }
                      }
                  } else {
                      $Str .= '[torrent]'.str_replace('[inlineurl]', '', $Block['Val']).'[/torrent]';
                  }
                  break;


                case 'wiki':
                  $Str .= '<a href="wiki.php?action=article&amp;name='.urlencode($Block['Val']).'">'.$Block['Val'].'</a>';
                  break;


                case 'tex':
                  $Str .= '<img class="tex_img" style="vertical-align: middle;" src="'.STATIC_SERVER.'blank.gif" onload="if (this.src.substr(this.src.length - 9, this.src.length) == \'blank.gif\') { this.src = \''.ImageTools::process('https://chart.googleapis.com/chart?cht=tx&chf=bg,s,FFFFFF00&chl='.urlencode(mb_convert_encoding($Block['Val'], 'UTF-8', 'HTML-ENTITIES'))).'\'; }" alt="'.$Block['Val'].'" />';
                  break;


                case 'plain':
                  $Str .= $Block['Val'];
                  break;


                case 'pre':
                  $Str .= '<pre>'.$Block['Val'].'</pre>';
                  break;


                case 'code':
                  $Str .= '<code>'.$Block['Val'].'</code>';
                  break;


                /*
                case 'ch':
                  $Str .= '<input type="checkbox" checked="checked" disabled="disabled">';
                  break;
                */


                /*
                case 'uch':
                  $Str .= '<input type="checkbox" disabled="disabled">';
                  break;
                */


                case 'list':
                  $Str .= "<$Block[ListType] class=\"postlist\">";
                  foreach ($Block['Val'] as $Line) {
                      $Str .= '<li>'.self::to_html($Line).'</li>';
                  }
                      $Str .= '</'.$Block['ListType'].'>';
                  break;


                case 'align':
                  $ValidAttribs = array('left', 'center', 'right');
                  if (!in_array($Block['Attr'], $ValidAttribs)) {
                      $Str .= '[align='.$Block['Attr'].']'.self::to_html($Block['Val']).'[/align]';
                  } else {
                      $Str .= '<div style="text-align: '.$Block['Attr'].';">'.self::to_html($Block['Val']).'</div>';
                  }
                  break;


                case 'color':
                case 'colour':
                  $ValidAttribs = array('aqua', 'black', 'blue', 'fuchsia', 'green', 'grey', 'lime', 'maroon', 'navy', 'olive', 'purple', 'red', 'silver', 'teal', 'white', 'yellow');
                  if (!in_array($Block['Attr'], $ValidAttribs) && !preg_match('/^#[0-9a-f]{6}$/', $Block['Attr'])) {
                      $Str .= '[color='.$Block['Attr'].']'.self::to_html($Block['Val']).'[/color]';
                  } else {
                      $Str .= '<span style="color: '.$Block['Attr'].';">'.self::to_html($Block['Val']).'</span>';
                  }
                  break;


                case 'headline':
                  $text = self::to_html($Block['Val']);
                  $raw = self::raw_text($Block['Val']);

                  if (!in_array($Block['Attr'], self::$HeadlineLevels)) {
                      $Str .= sprintf('%1$s%2$s%1$s', str_repeat('=', $Block['Attr'] + 1), $text);
                  } else {
                      $id = '_' . crc32($raw . self::$HeadlineID);

                      if (self::$InQuotes === 0) {
                          self::$Headlines[] = array($Block['Attr'], $raw, $id);
                      }
                          
                      $Str .= sprintf('<h%1$d id="%3$s">%2$s</h%1$d>', ($Block['Attr'] + 2), $text, $id);
                      self::$HeadlineID++;
                  }
                  break;


                case 'inlinesize':
                case 'size':
                  $ValidAttribs = array('1', '2', '3', '4', '5', '6', '7', '8', '9', '10');
                  if (!in_array($Block['Attr'], $ValidAttribs)) {
                      $Str .= '[size='.$Block['Attr'].']'.self::to_html($Block['Val']).'[/size]';
                  } else {
                      $Str .= '<span class="size'.$Block['Attr'].'">'.self::to_html($Block['Val']).'</span>';
                  }
                  break;


                case 'quote':
                  self::$NoImg++; // No images inside quote tags
                  self::$InQuotes++;

                  if (self::$InQuotes === self::$NestsBeforeHide) { // Put quotes that are nested beyond the specified limit in [hide] tags.
                      $Str .= '<strong>Older quotes</strong>: <a class="spoilerButton">Show</a>';
                      $Str .= '<blockquote class="hidden spoiler">';
                  }

                  if (!empty($Block['Attr'])) {
                      $Exploded = explode('|', self::to_html($Block['Attr']));
                      if (isset($Exploded[1]) && (is_numeric($Exploded[1]) || (in_array($Exploded[1][0], array('a', 't', 'c', 'r')) && is_numeric(substr($Exploded[1], 1))))) {
                          // The part after | is either a number or starts with a, t, c or r, followed by a number (forum post, artist comment, torrent comment, collage comment or request comment, respectively)
                          $PostID = trim($Exploded[1]);
                          $Str .= '<a data-quote-jump="'.$PostID.'"><strong class="quoteheader">'.$Exploded[0].'</strong> wrote: </a>';
                      } else {
                          $Str .= '<strong class="quoteheader">'.$Exploded[0].'</strong> wrote: ';
                      }
                  }
                      
                  $Str .= '<blockquote>'.self::to_html($Block['Val']).'</blockquote>';
                  if (self::$InQuotes === self::$NestsBeforeHide) { //Close quote the deeply nested quote [hide].
                        $Str .= '</blockquote><br />'; // Ensure new line after quote train hiding
                  }

                  self::$NoImg--;
                  self::$InQuotes--;
                  break;


                case 'hide':
                  $Str .= '<strong>'.(($Block['Attr']) ? $Block['Attr'] : 'Hidden text').'</strong>: <a class="spoilerButton">Show</a>';
                  $Str .= '<blockquote class="hidden spoiler">'.self::to_html($Block['Val']).'</blockquote>';
                  break;


                case 'img':
                  if (self::$NoImg > 0 && self::valid_url($Block['Val'])) {
                      $Str .= '<a rel="noreferrer" target="_blank" href="'.$Block['Val'].'">'.$Block['Val'].'</a> (image)';
                      break;
                  }

                  if (!self::valid_url($Block['Val'], '\.(jpe?g|gif|png|bmp|tiff)')) {
                      $Str .= '[img]'.$Block['Val'].'[/img]';
                  } else {
                      $LocalURL = self::local_url($Block['Val']);
                      if ($LocalURL) {
                          $Str .= '<img class="lightbox-init" alt="'.$Block['Val'].'" src="'.$LocalURL.'" />';
                      } else {
                          $Str .= '<img class="lightbox-init" alt="'.$Block['Val'].'" src="'.ImageTools::process($Block['Val']).'" />';
                      }
                  }
                  break;


                case 'aud':
                  if (self::$NoImg > 0 && self::valid_url($Block['Val'])) {
                      $Str .= '<a rel="noreferrer" target="_blank" href="'.$Block['Val'].'">'.$Block['Val'].'</a> (audio)';
                      break;
                  }

                  if (!self::valid_url($Block['Val'], '\.(mp3|ogg|wav)')) {
                      $Str .= '[aud]'.$Block['Val'].'[/aud]';
                  } else {
                      // todo: Proxy this for staff?
                      $Str .= '<audio controls="controls" src="'.$Block['Val'].'"><a rel="noreferrer" target="_blank" href="'.$Block['Val'].'">'.$Block['Val'].'</a></audio>';
                  }
                  break;


                case 'url':
                  // Make sure the URL has a label
                  if (empty($Block['Val'])) {
                      $Block['Val'] = $Block['Attr'];
                      $NoName = true; // If there isn't a Val for this
                  } else {
                      $Block['Val'] = self::to_html($Block['Val']);
                      $NoName = false;
                  }

                  if (!self::valid_url($Block['Attr'])) {
                      $Str .= '[url='.$Block['Attr'].']'.$Block['Val'].'[/url]';
                  } else {
                      $LocalURL = self::local_url($Block['Attr']);
                      if ($LocalURL) {
                          if ($NoName) {
                              $Block['Val'] = substr($LocalURL, 1);
                          }
                          $Str .= '<a href="'.$LocalURL.'">'.$Block['Val'].'</a>';
                      } else {
                          $Str .= '<a rel="noreferrer" target="_blank" href="'.$Block['Attr'].'">'.$Block['Val'].'</a>';
                      }
                  }
                  break;


                case 'inlineurl':
                  if (!self::valid_url($Block['Attr'], '', true)) {
                      $Array = self::parse($Block['Attr']);
                      $Block['Attr'] = $Array;
                      $Str .= self::to_html($Block['Attr']);
                  } else {
                      $LocalURL = self::local_url($Block['Attr']);
                      if ($LocalURL) {
                          $Str .= '<a href="'.$LocalURL.'">'.substr($LocalURL, 1).'</a>';
                      } else {
                          $Str .= '<a rel="noreferrer" target="_blank" href="'.$Block['Attr'].'">'.$Block['Attr'].'</a>';
                      }
                  }
                  break;


                case 'embed':
                  $Val = str_replace(' ', '', $Block['Val']);
                  if (self::valid_url($Val) && substr($Val, -4) === 'webm') {
                      $Str .= '<video class="webm" preload controls><source src="'.ImageTools::process($Val).'" /></video>';
                  }
                  break;
                }
            }
        }

        self::$Levels--;
        return $Str;
    }


    /**
     * raw_text()
     */
    private static function raw_text($Array)
    {
        $Str = '';
        foreach ($Array as $Block) {
            if (is_string($Block)) {
                $Str .= $Block;
                continue;
            }

            switch ($Block['Type']) {
            case 'headline':
              break;


            case 'b':
            case 'u':
            case 'i':
            case 's':
            case 'color':
            case 'size':
            case 'quote':
            case 'align':
              $Str .= self::raw_text($Block['Val']);
              break;


            case 'tex': // Since this will never strip cleanly, just remove it
              break;


            case 'artist':
            case 'user':
            case 'wiki':
            case 'pre':
            case 'code':
            case 'aud':
            case 'img':
              $Str .= $Block['Val'];
              break;


            case 'list':
              foreach ($Block['Val'] as $Line) {
                  $Str .= $Block['Tag'].self::raw_text($Line);
              }
              break;


            case 'url':
              // Make sure the URL has a label
              if (empty($Block['Val'])) {
                  $Block['Val'] = $Block['Attr'];
              } else {
                  $Block['Val'] = self::raw_text($Block['Val']);
              }

              $Str .= $Block['Val'];
              break;


            case 'inlineurl':
              if (!self::valid_url($Block['Attr'], '', true)) {
                  $Array = self::parse($Block['Attr']);
                  $Block['Attr'] = $Array;
                  $Str .= self::raw_text($Block['Attr']);
              } else {
                  $Str .= $Block['Attr'];
              }
              break;
            }
        }

        return $Str;
    }


    /**
     * smileys()
     */
    private static function smileys($Str)
    {
        return $Str;

        /*
        if (!empty(G::$LoggedUser['DisableSmileys'])) {
            return $Str;
        }

        if (count(self::$ProcessedSmileys) === 0 && count(self::$Smileys) > 0) {
            foreach (self::$Smileys as $Key => $Val) {
                self::$ProcessedSmileys[$Key] = '<img src="'.STATIC_SERVER.'common/smileys/'.$Val.'" alt="" />';
            }
            reset(self::$ProcessedSmileys);
        }

        $Str = strtr($Str, self::$ProcessedSmileys);
        return $Str;
        */
    }
}


/*
// Uncomment this part to test the class via command line:
function display_str($Str) {
  return $Str;
}
function check_perms($Perm) {
  return true;
}
$Str = "hello
[pre]http://anonym.to/?http://whatshirts.portmerch.com/
====hi====
===hi===
==hi==[/pre]
====hi====
hi";
echo Text::full_format($Str);
echo "\n"
*/