src/htmLawed/htmLawed.php ➔ hl_bal()

Complexity

Conditions	100
Paths	> 20000

Size

Total Lines

263

Duplication

Lines	84
Ratio	31.94 %

Code Coverage

Tests	124
CRAP Score	544.3709

Importance

Changes

0

<?php
function hl_bal($t, $do = 1, $in = 'div') {
    // balance tags
// by content
    $cB = array('blockquote' => 1, 'form' => 1, 'map' => 1, 'noscript' => 1); // Block
    $cE = array('area' => 1, 'br' => 1, 'col' => 1, 'command' => 1, 'embed' => 1, 'hr' => 1, 'img' => 1, 'input' => 1, 'isindex' => 1, 'keygen' => 1, 'link' => 1, 'meta' => 1, 'param' => 1, 'source' => 1, 'track' => 1, 'wbr' => 1); // Empty
    $cF = array('a' => 1, 'article' => 1, 'aside' => 1, 'audio' => 1, 'button' => 1, 'canvas' => 1, 'del' => 1, 'details' => 1, 'div' => 1, 'dd' => 1, 'fieldset' => 1, 'figure' => 1, 'footer' => 1, 'header' => 1, 'iframe' => 1, 'ins' => 1, 'li' => 1, 'main' => 1, 'menu' => 1, 'nav' => 1, 'noscript' => 1, 'object' => 1, 'section' => 1, 'style' => 1, 'td' => 1, 'th' => 1, 'video' => 1); // Flow; later context-wise dynamic move of ins & del to $cI
    $cI = array('abbr' => 1, 'acronym' => 1, 'address' => 1, 'b' => 1, 'bdi' => 1, 'bdo' => 1, 'big' => 1, 'caption' => 1, 'cite' => 1, 'code' => 1, 'data' => 1, 'datalist' => 1, 'dfn' => 1, 'dt' => 1, 'em' => 1, 'figcaption' => 1, 'font' => 1, 'h1' => 1, 'h2' => 1, 'h3' => 1, 'h4' => 1, 'h5' => 1, 'h6' => 1, 'hgroup' => 1, 'i' => 1, 'kbd' => 1, 'label' => 1, 'legend' => 1, 'mark' => 1, 'meter' => 1, 'output' => 1, 'p' => 1, 'pre' => 1, 'progress' => 1, 'q' => 1, 'rb' => 1, 'rt' => 1, 's' => 1, 'samp' => 1, 'small' => 1, 'span' => 1, 'strike' => 1, 'strong' => 1, 'sub' => 1, 'summary' => 1, 'sup' => 1, 'time' => 1, 'tt' => 1, 'u' => 1, 'var' => 1); // Inline
    $cN = array('a' => array('a' => 1, 'address' => 1, 'button' => 1, 'details' => 1, 'embed' => 1, 'keygen' => 1, 'label' => 1, 'select' => 1, 'textarea' => 1), 'address' => array('address' => 1, 'article' => 1, 'aside' => 1, 'header' => 1, 'keygen' => 1, 'footer' => 1, 'nav' => 1, 'section' => 1), 'button' => array('a' => 1, 'address' => 1, 'button' => 1, 'details' => 1, 'embed' => 1, 'fieldset' => 1, 'form' => 1, 'iframe' => 1, 'input' => 1, 'keygen' => 1, 'label' => 1, 'select' => 1, 'textarea' => 1), 'fieldset' => array('fieldset' => 1), 'footer' => array('header' => 1, 'footer' => 1), 'form' => array('form' => 1), 'header' => array('header' => 1, 'footer' => 1), 'label' => array('label' => 1), 'main' => array('main' => 1), 'meter' => array('meter' => 1), 'noscript' => array('script' => 1), 'pre' => array('big' => 1, 'font' => 1, 'img' => 1, 'object' => 1, 'script' => 1, 'small' => 1, 'sub' => 1, 'sup' => 1), 'progress' => array('progress' => 1), 'rb' => array('ruby' => 1), 'rt' => array('ruby' => 1), 'time' => array('time' => 1)); // Illegal
    $cN2 = array_keys($cN);
    $cS = array('colgroup' => array('col' => 1), 'datalist' => array('option' => 1), 'dir' => array('li' => 1), 'dl' => array('dd' => 1, 'dt' => 1), 'hgroup' => array('h1' => 1, 'h2' => 1, 'h3' => 1, 'h4' => 1, 'h5' => 1, 'h6' => 1), 'menu' => array('li' => 1), 'ol' => array('li' => 1), 'optgroup' => array('option' => 1), 'option' => array('#pcdata' => 1), 'rbc' => array('rb' => 1), 'rp' => array('#pcdata' => 1), 'rtc' => array('rt' => 1), 'ruby' => array('rb' => 1, 'rbc' => 1, 'rp' => 1, 'rt' => 1, 'rtc' => 1), 'select' => array('optgroup' => 1, 'option' => 1), 'script' => array('#pcdata' => 1), 'table' => array('caption' => 1, 'col' => 1, 'colgroup' => 1, 'tfoot' => 1, 'tbody' => 1, 'tr' => 1, 'thead' => 1), 'tbody' => array('tr' => 1), 'tfoot' => array('tr' => 1), 'textarea' => array('#pcdata' => 1), 'thead' => array('tr' => 1), 'tr' => array('td' => 1, 'th' => 1), 'ul' => array('li' => 1)); // Specific - immediate parent-child
    if ($GLOBALS['C']['direct_list_nest']) {
        $cS['ol'] = $cS['ul'] = $cS['menu'] += array('menu' => 1, 'ol' => 1, 'ul' => 1);
    }
    $cO = array('address' => array('p' => 1), 'applet' => array('param' => 1), 'audio' => array('source' => 1, 'track' => 1), 'blockquote' => array('script' => 1), 'details' => array('summary' => 1), 'fieldset' => array('legend' => 1, '#pcdata' => 1),  'figure' => array('figcaption' => 1), 'form' => array('script' => 1), 'map' => array('area' => 1), 'object' => array('param' => 1, 'embed' => 1), 'video' => array('source' => 1, 'track' => 1)); // Other
    $cT = array('colgroup' => 1, 'dd' => 1, 'dt' => 1, 'li' => 1, 'option' => 1, 'p' => 1, 'td' => 1, 'tfoot' => 1, 'th' => 1, 'thead' => 1, 'tr' => 1); // Omitable closing
// block/inline type; a/ins/del both type; #pcdata: text
    $eB = array('a' => 1, 'address' => 1, 'article' => 1, 'aside' => 1, 'blockquote' => 1, 'center' => 1, 'del' => 1, 'details' => 1, 'dir' => 1, 'dl' => 1, 'div' => 1, 'fieldset' => 1, 'figure' => 1, 'footer' => 1, 'form' => 1, 'ins' => 1, 'h1' => 1, 'h2' => 1, 'h3' => 1, 'h4' => 1, 'h5' => 1, 'h6' => 1, 'header' => 1, 'hr' => 1, 'isindex' => 1, 'main' => 1, 'menu' => 1, 'nav' => 1, 'noscript' => 1, 'ol' => 1, 'p' => 1, 'pre' => 1, 'section' => 1, 'style' => 1, 'table' => 1, 'ul' => 1);
    $eI = array('#pcdata' => 1, 'a' => 1, 'abbr' => 1, 'acronym' => 1, 'applet' => 1, 'audio' => 1, 'b' => 1, 'bdi' => 1, 'bdo' => 1, 'big' => 1, 'br' => 1, 'button' => 1, 'canvas' => 1, 'cite' => 1, 'code' => 1, 'command' => 1, 'data' => 1, 'datalist' => 1, 'del' => 1, 'dfn' => 1, 'em' => 1, 'embed' => 1, 'figcaption' => 1, 'font' => 1, 'i' => 1, 'iframe' => 1, 'img' => 1, 'input' => 1, 'ins' => 1, 'kbd' => 1, 'label' => 1, 'link' => 1, 'map' => 1, 'mark' => 1, 'meta' => 1, 'meter' => 1, 'object' => 1, 'output' => 1, 'progress' => 1, 'q' => 1, 'ruby' => 1, 's' => 1, 'samp' => 1, 'select' => 1, 'script' => 1, 'small' => 1, 'span' => 1, 'strike' => 1, 'strong' => 1, 'sub' => 1, 'summary' => 1, 'sup' => 1, 'textarea' => 1, 'time' => 1, 'tt' => 1, 'u' => 1, 'var' => 1, 'video' => 1, 'wbr' => 1);
    $eN = array('a' => 1, 'address' => 1, 'article' => 1, 'aside' => 1, 'big' => 1, 'button' => 1, 'details' => 1, 'embed' => 1, 'fieldset' => 1, 'font' => 1, 'footer' => 1, 'form' => 1, 'header' => 1, 'iframe' => 1, 'img' => 1, 'input' => 1, 'keygen' => 1, 'label' => 1, 'meter' => 1, 'nav' => 1, 'object' => 1, 'progress' => 1, 'ruby' => 1, 'script' => 1, 'select' => 1, 'small' => 1, 'sub' => 1, 'sup' => 1, 'textarea' => 1, 'time' => 1); // Exclude from specific ele; $cN values
    $eO = array('area' => 1, 'caption' => 1, 'col' => 1, 'colgroup' => 1, 'command' => 1, 'dd' => 1, 'dt' => 1, 'hgroup' => 1, 'keygen' => 1, 'legend' => 1, 'li' => 1, 'optgroup' => 1, 'option' => 1, 'param' => 1, 'rb' => 1, 'rbc' => 1, 'rp' => 1, 'rt' => 1, 'rtc' => 1, 'script' => 1, 'source' => 1, 'tbody' => 1, 'td' => 1, 'tfoot' => 1, 'thead' => 1, 'th' => 1, 'tr' => 1, 'track' => 1); // Missing in $eB & $eI
    $eF = $eB + $eI;

    // $in sets allowed child
    $in = ((isset($eF[$in]) && '#pcdata' != $in) or isset($eO[$in])) ? $in : 'div';
    if (isset($cE[$in])) {
        return !$do ? '' : str_replace(array('<', '>'), array('&lt;', '&gt;'), $t);
    }
    if (isset($cS[$in])) {
        $inOk = $cS[$in];
    } elseif (isset($cI[$in])) {
        $inOk = $eI;
        $cI['del'] = 1;
        $cI['ins'] = 1;
    } elseif (isset($cF[$in])) {
        $inOk = $eF;
        unset($cI['del'], $cI['ins']);
    } elseif (isset($cB[$in])) {
        $inOk = $eB;
        unset($cI['del'], $cI['ins']);
    }
    if (isset($cO[$in])) {
        $inOk = $inOk + $cO[$in];

The variable $inOk does not seem to be defined for all execution paths leading up to this point.

    }
    if (isset($cN[$in])) {
        $inOk = array_diff_assoc($inOk, $cN[$in]);
    }

    $t = explode('<', $t);
    $ok = $q = array(); // $q seq list of open non-empty ele
    ob_start();

    for ($i = -1, $ci = count($t); ++$i < $ci;) {
        // allowed $ok in parent $p
        if ($ql = count($q)) {

This code seems to be duplicated across your project.

            $p = array_pop($q);
            $q[] = $p;
            if (isset($cS[$p])) {
                $ok = $cS[$p];
            } elseif (isset($cI[$p])) {
                $ok = $eI;
                $cI['del'] = 1;
                $cI['ins'] = 1;
            } elseif (isset($cF[$p])) {
                $ok = $eF;
                unset($cI['del'], $cI['ins']);
            } elseif (isset($cB[$p])) {
                $ok = $eB;
                unset($cI['del'], $cI['ins']);
            }
            if (isset($cO[$p])) {
                $ok = $ok + $cO[$p];
            }
            if (isset($cN[$p])) {
                $ok = array_diff_assoc($ok, $cN[$p]);
            }
        } else {
            $ok = $inOk;
            unset($cI['del'], $cI['ins']);
        }
        // bad tags, & ele content
        if (isset($e) && (1 == $do or (isset($ok['#pcdata']) && (3 == $do or 5 == $do)))) {

This code seems to be duplicated across your project.

            echo '<', $s, $e, $a, '>';

The variable $s does not seem to be defined for all execution paths leading up to this point.

The variable $a does not seem to be defined for all execution paths leading up to this point.

        }
        if (isset($x[0])) {

This code seems to be duplicated across your project.

            if (strlen(trim($x)) && (($ql && isset($cB[$p])) or (isset($cB[$in]) && !$ql))) {
                echo '<div>', $x, '</div>';
            } elseif ($do < 3 or isset($ok['#pcdata'])) {
                echo $x;
            } elseif (strpos($x, "\x02\x04")) {
                foreach (preg_split('`(\x01\x02[^\x01\x02]+\x02\x01)`', $x, -1, PREG_SPLIT_DELIM_CAPTURE | PREG_SPLIT_NO_EMPTY) as $v) {
                    echo "\x01\x02" == substr($v, 0, 2) ? $v : ($do > 4 ? preg_replace('`\S`', '', $v) : '');
                }
            } elseif ($do > 4) {
                echo preg_replace('`\S`', '', $x);
            }
        }
        // get markup
        if (!preg_match('`^(/?)([a-z1-6]+)([^>]*)>(.*)`sm', $t[$i], $r)) {
            $x = $t[$i];
            continue;
        }
        $s = null;
        $e = null;
        $a = null;
        $x = null;
        list($all, $s, $e, $a, $x) = $r;

The assignment to $all is unused. Consider omitting it like so list($first,,$third).

        // close tag
        if ($s) {
            if (isset($cE[$e]) or !in_array($e, $q)) {
                continue;
            } // Empty/unopen
            if ($p == $e) {

The variable $p does not seem to be defined for all execution paths leading up to this point.

                array_pop($q);
                echo '</', $e, '>';
                unset($e);
                continue;
            } // Last open
            $add = ''; // Nesting - close open tags that need to be
            for ($j = -1, $cj = count($q); ++$j < $cj;) {
                if (($d = array_pop($q)) == $e) {
                    break;
                } else {
                    $add .= "</{$d}>";
                }
            }
            echo $add, '</', $e, '>';
            unset($e);
            continue;
        }
        // open tag
        // $cB ele needs $eB ele as child
        if (isset($cB[$e]) && strlen(trim($x))) {
            $t[$i] = "{$e}{$a}>";
            array_splice($t, $i + 1, 0, 'div>'.$x);
            unset($e, $x);
            ++$ci;
            --$i;
            continue;
        }
        if ((($ql && isset($cB[$p])) or (isset($cB[$in]) && !$ql)) && !isset($eB[$e]) && !isset($ok[$e])) {
            array_splice($t, $i, 0, 'div>');
            unset($e, $x);
            ++$ci;
            --$i;
            continue;
        }
        // if no open ele, $in = parent; mostly immediate parent-child relation should hold
        if (!$ql or !isset($eN[$e]) or !array_intersect($q, $cN2)) {
            if (!isset($ok[$e])) {
                if ($ql && isset($cT[$p])) {
                    echo '</', array_pop($q), '>';
                    unset($e, $x);
                    --$i;
                }
                continue;
            }
            if (!isset($cE[$e])) {
                $q[] = $e;
            }
            echo '<', $e, $a, '>';
            unset($e);
            continue;
        }
        // specific parent-child
        if (isset($cS[$p][$e])) {
            if (!isset($cE[$e])) {
                $q[] = $e;
            }
            echo '<', $e, $a, '>';
            unset($e);
            continue;
        }
        // nesting
        $add = '';
        $q2 = array();
        for ($k = -1, $kc = count($q); ++$k < $kc;) {
            $d = $q[$k];
            $ok2 = array();

$ok2 is not used, you could remove the assignment.

            if (isset($cS[$d])) {
                $q2[] = $d;
                continue;
            }
            $ok2 = isset($cI[$d]) ? $eI : $eF;
            if (isset($cO[$d])) {
                $ok2 = $ok2 + $cO[$d];
            }
            if (isset($cN[$d])) {
                $ok2 = array_diff_assoc($ok2, $cN[$d]);
            }
            if (!isset($ok2[$e])) {
                if (!$k && !isset($inOk[$e])) {
                    continue 2;
                }
                $add = "</{$d}>";
                for (; ++$k < $kc;
                ) {
                    $add = "</{$q[$k]}>{$add}";
                }
                break;
            } else {
                $q2[] = $d;
            }
        }
        $q = $q2;
        if (!isset($cE[$e])) {
            $q[] = $e;
        }
        echo $add, '<', $e, $a, '>';
        unset($e);
        continue;
    }

    // end
    if ($ql = count($q)) {

This code seems to be duplicated across your project.

        $p = array_pop($q);
        $q[] = $p;
        if (isset($cS[$p])) {
            $ok = $cS[$p];
        } elseif (isset($cI[$p])) {
            $ok = $eI;
            $cI['del'] = 1;
            $cI['ins'] = 1;
        } elseif (isset($cF[$p])) {
            $ok = $eF;
            unset($cI['del'], $cI['ins']);
        } elseif (isset($cB[$p])) {
            $ok = $eB;
            unset($cI['del'], $cI['ins']);
        }
        if (isset($cO[$p])) {
            $ok = $ok + $cO[$p];
        }
        if (isset($cN[$p])) {
            $ok = array_diff_assoc($ok, $cN[$p]);
        }
    } else {
        $ok = $inOk;
        unset($cI['del'], $cI['ins']);
    }
    if (isset($e) && (1 == $do or (isset($ok['#pcdata']) && (3 == $do or 5 == $do)))) {

This code seems to be duplicated across your project.

        echo '<', $s, $e, $a, '>';
    }
    if (isset($x[0])) {

This code seems to be duplicated across your project.

        if (strlen(trim($x)) && (($ql && isset($cB[$p])) or (isset($cB[$in]) && !$ql))) {
            echo '<div>', $x, '</div>';
        } elseif ($do < 3 or isset($ok['#pcdata'])) {
            echo $x;
        } elseif (strpos($x, "\x02\x04")) {
            foreach (preg_split('`(\x01\x02[^\x01\x02]+\x02\x01)`', $x, -1, PREG_SPLIT_DELIM_CAPTURE | PREG_SPLIT_NO_EMPTY) as $v) {
                echo "\x01\x02" == substr($v, 0, 2) ? $v : ($do > 4 ? preg_replace('`\S`', '', $v) : '');
            }
        } elseif ($do > 4) {
            echo preg_replace('`\S`', '', $x);
        }
    }
    while (!empty($q) && ($e = array_pop($q))) {
        echo '</', $e, '>';
    }
    $o = ob_get_contents();
    ob_end_clean();

    return $o;
}