<?

/*
 * Simple HTML 4.01 parser class for PHP
 *
 * Copyright (c) 2007, Matthew Mondor
 * ALL RIGHTS RESERVED.
 */

class HTMLParser
{

        private $str = '';
        private $str_len = 0;
        private $str_i = 0;

        private $callback = NULL;
        private $udata = NULL;

        private $state = NULL;

        private $tag = NULL;
        private $tag_len = 0;
        private $attribute = NULL;
        private $attributes = array();
        private $last_free = NULL;

        public function __construct($str, $function, &$udata)
        {
                $this->str = $str;
                $this->str_len = strlen($str);
                $this->callback = $function;
                $this->udata = &$udata;
        }

        public function parse()
        {
                $this->state = 'state_0';
                for ($this->str_i = 0; $this->str_i < $this->str_len; ) {
                        if (($x = $this->{$this->state}()) !== NULL)
                                return $x;
                }
                /* Send any remaining data */
                if ($this->last_free < $this->str_i) {
                        $s = trim(substr($this->str, $this->last_free,
                            $this->str_i - $this->last_free));
                        if (strlen($s) > 0) {
                                $o = array(
                                    'type' => 'content',
                                    'content' => $s);
                                $f = &$this->callback;
                                if (($r = $f($o, $this->udata)) != NULL)
                                        return $r;
                        }
                }

                /* NOTREACHED */
                return NULL;
        }

        private function state_0()
        {
                $this->last_free = $this->str_i;

                /* Free-run until '<', switch to state 1 if found */
                for (; $this->str_i < $this->str_len; $this->str_i++) {
                        if ($this->str{$this->str_i} == '<') {

                                if ($this->last_free < $this->str_i) {
                                        $s = trim(substr($this->str,
                                            $this->last_free, $this->str_i -
                                            $this->last_free));
                                        if (strlen($s) > 0) {
                                                $o = array(
                                                    'type' => 'content',
                                                    'content' => $s);
                                                $f = &$this->callback;
                                                if (($r = $f($o, $this->udata))
                                                    != NULL)
                                                        return $r;
                                        }
                                }

                                $this->state = 'state_1';
                                return NULL;
                        }
                }
                return NULL;
        }

        private function state_1()
        {
                /*
                 * Skip comment if any;
                 * error if reaching EOF before comment ends
                 */
                if ($this->str_i <= $this->str_len - 4 &&
                    ($i = strpos($this->str, '<!--', $this->str_i)) !== FALSE
                    && $i == $this->str_i) {
                        /* Was a comment, switch back to state 0 */
                        $this->str_i += 4;
                        if (($i = strpos($this->str, '-->', $this->str_i))
                            === FALSE)
                                return "Expected '-->' but reached EOF";
                        $this->str_i = $i + 3;
                        $this->state = 'state_0';
                        return NULL;
                }

                /* Not a comment, switch to state 2 */
                $this->state = 'state_2';
                return NULL;
        }

        private function state_2()
        {
                /* Tag must start with '<' */
                if ($this->str{$this->str_i} != '<')
                        return "Expected '<' at position {$this->str_i}";

                /* Tag start, switch to state 3 */
                $this->str_i++;
                $this->state = 'state_3';
                return NULL;
        }

        private function state_3()
        {
                /* Load tag name */
                for ($this->tag = '', $this->tag_len = 0;
                    $this->str_i < $this->str_len &&
                    !ctype_space($this->str{$this->str_i}) &&
                    $this->str{$this->str_i} != '>';
                    $this->str_i++, $this->tag_len++)
                        $this->tag .= $this->str{$this->str_i};
                if ($this->tag_len == 0)
                        return "Expected tag name at position {$this->str_i}";

                /* Non-empty tag name, switch to state 4 */
                $this->attributes = array();
                $this->state = 'state_4';
                return NULL;
        }

        private function state_4()
        {
                /* Skip whitespaces */
                for (; $this->str_i < $this->str_len &&
                     ctype_space($this->str{$this->str_i}); $this->str_i++) ;

                /*
                 * Upon end of tag, call the user function with the tag
                 * and switch to state 0.
                 */
                if ($this->str{$this->str_i} == '>') {
                        $this->str_i++;
                        $o = array(
                            'type' => 'tag',
                            'tag' => array(
                                'name' => strtolower($this->tag),
                                'attributes' => $this->attributes));
                        $f = &$this->callback;
                        if (($r = $f($o, $this->udata)) != NULL)
                                return $r;
                        $this->state = 'state_0';
                        return NULL;
                }

                /* Load attribute name */
                for ($this->attribute = '';
                    $this->str_i < $this->str_len &&
                    !ctype_space($this->str{$this->str_i}) &&
                    $this->str{$this->str_i} != '>';
                    $this->str_i++)
                        $this->attribute .= $this->str{$this->str_i};
                if (strlen($this->attribute) == 0)
                        return 'Expected attribute name at position ' .
                         "{$this->str_i}";

                /* Skip whitespaces */
                for (; $this->str_i < $this->str_len &&
                     ctype_space($this->str{$this->str_i}); $this->str_i++) ;

                /*
                 * Nonempty attribute name, is it assigned a value?
                 * Switch to state 4 if empty, or 5 to assign it.
                 */
                if ($this->str{$this->str_i} != '=') {
                        $this->attributes[strtolower($this->attribute)] = NULL;
                        $this->state = 'state_4';
                        return NULL;
                }
                $this->str_i++;
                $this->state = 'state_5';
                return NULL;
        }

        private function state_5()
        {
                /* Skip whitespaces */
                for (; $this->str_i < $this->str_len &&
                     ctype_space($this->str{$this->str_i}); $this->str_i++) ;

                /* Assignment must start with '"' */
                if ($this->str{$this->str_i} != '"')
                        return "Expected '\"' at position {$this->str_i}";
                $this->str_i++;

                /*
                 * Find ending '"', error if EOF;  Assign attribute and switch
                 * back to state 4.
                 */
                if (($i = strpos($this->str, '"', $this->str_i)) === FALSE) {
                        --$this->str_i;
                        return "Expected closing '\"' to match position " .
                            "{$this->str_i} but reached EOF";
                }
                if ($i > $this->str_i) {
                        $this->attributes[strtolower($this->attribute)] =
                            substr($this->str, $this->str_i,
                            ($i - $this->str_i));
                        $this->str_i = $i + 1;
                } else {
                        $this->attributes[strtolower($this->attribute)] = NULL;
                        $this->str_i++;
                }
                $this->state = 'state_4';
                return NULL;
        }

}



/*
 * Test the parser
 */

function po_callback(&$o, &$udata)
{
        print_r($o);

        /*
        if ($o['type'] == 'content')
                print($o['content']);
        else {
                print("<{$o['tag']['name']}");
                $a = $o['tag']['attributes'];
                foreach($a as $k => $v) {
                        if ($v != NULL)
                                print(" {$k}=\"{$v}\"");
                        else
                                print(" {$k}");
                }
                print('>');
                $udata['tags']++;
        }
        */

        $udata['tags']++;

        return NULL;
}

$p_udata = array();
$p_udata['tags'] = 0;

$po = new HTMLParser(file_get_contents('bookmarks.html'), 'po_callback',
    $p_udata);
print($po->parse() . "\n");

print("Tags: {$p_udata['tags']}\n");

?>