PHPIDSConverter::convertJSUnicode() - Code Metrics - IonutBajescu/sylar - Measure and Improve Code Quality continuously with Scrutinizer

PHPIDSConverter::convertJSUnicode() A
last analyzed 2016-05-08 10:19 UTC

↳ Parent: PHPIDSConverter

Complexity

Conditions	3
Paths	2

Size

Total Lines	14
Code Lines	9

Duplication

Lines	0
Ratio	0 %

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
c	1
b	0
f	0
dl	0
loc	14
rs	9.4285
cc	3
eloc	9
nc	2
nop	1

<?php


namespace Ionut\Sylar\Normalizers;


use Ionut\Sylar\NormalizedValue;
use Ionut\Sylar\NormalizedValueVariant;

class PHPIDSConverter implements NormalizerInterface
{

    /**
     * @var array
     */
    protected $notConverters = ['normalize', 'replacePairs'];

    /**
     * @param  array $parameters
     * @return array
     */
    public function normalize(array $parameters)
    {
        $converters = array_diff(get_class_methods(__CLASS__), $this->notConverters);

        array_walk_recursive($parameters, function (&$value) use ($converters) {
            $value = $value instanceof NormalizedValue ? $value : new NormalizedValue($value);

            $variant = $value->getOriginal();
            foreach ($converters as $method) {
                $variant = $this->$method($variant);
            }

            $value->variants[static::class] = new NormalizedValueVariant($variant);
        });

        return $parameters;
    }

    protected function replacePairs(array $patternReplacementPair, $value)
    {
        return preg_replace(
            array_keys($patternReplacementPair),
            array_values($patternReplacementPair),
            $value
        );
    }

    /**
     * Erases comments.
     *
     * @param  string $value
     * @return string
     */
    public function convertCommented($value)
    {
        if (preg_match('/(?:\<!-|-->|\/\*|\*\/|\/\/\W*\w+\s*$)|(?:--[^-]*-)/ms', $value)) {
            $pattern = [
                '/(?:(?:<!)(?:(?:--(?:[^-]*(?:-[^-]+)*)--\s*)*)(?:>))/ms',
                '/(?:(?:\/\*\/*[^\/\*]*)+\*\/)/ms',
                '/(?:--[^-]*-)/ms'
            ];

            return $value."\n".preg_replace($pattern, ';', $value);
        }

        return $value;
    }

    /**
     * @param  string $value
     * @return string
     */
    public function convertInlineComments($value)
    {
        return $this->replacePairs([
            '/(<\w+)\/+(\w+=?)/m'      => '$1/$2',
            '/[^\\\:]\/\/(.*)$/m'      => '/**/$1',
            '/([^\-&])#.*[\r\n\v\f]/m' => '$1',
            '/([^&\-])#.*\n/m'         => '$1',
            '/^#.*\n/m'                => ' '
        ], $value);
    }

    /**
     * Replaces newlines with semicolons.
     *
     * @param  string $value
     * @return string
     */
    public function convertWhiteSpace($value)
    {
        $value = str_replace(['\r', '\n', '\f', '\t', '\v'], ';', $value);
        $value = str_replace('�', ' ', $value);
        return preg_replace('/(?:\n|\r|\v)/m', '  ', $value);
    }

    /**
     * Decodes charcodes.
     *
     * @param  string $value
     * @return string
     */
    public function convertJSCharcode($value)
    {
        $matches = [];

        // Check for typical charcode pattern
        if (preg_match_all('/(?:[\d+-=\/\* ]+(?:\s?,\s?[\d+-=\/\* ]+)){4,}/ms', $value, $matches)) {
            $converted = '';
            $string = implode(',', $matches[0]);
            $string = preg_replace('/\s/', '', $string);
            $string = preg_replace('/\w+=/', '', $string);
            $charcode = explode(',', $string);
            foreach ($charcode as $char) {
                $char = preg_replace('/\W0/s', '', $char);
                if (preg_match_all('/\d*[+-\/\* ]\d+/', $char, $matches)) {
                    $match = preg_split('/(\W?\d+)/', implode('', $matches[0]), null, PREG_SPLIT_DELIM_CAPTURE);
                    if (array_sum($match) >= 20 && array_sum($match) <= 127) {
                        $converted .= chr(array_sum($match));
                    }
                } elseif (!empty($char) && $char >= 20 && $char <= 127) {
                    $converted .= chr($char);
                }
            }
            $value .= "\n".$converted;
        }

        // Check for octal charcode pattern
        if (preg_match_all('/(?:(?:[\\\]+\d+[ \t]*){8,})/ims', $value, $matches)) {
            $converted = '';
            $charcode = explode('\\', preg_replace('/\s/', '', implode(',', $matches[0])));
            foreach (array_map('octdec', array_filter($charcode)) as $char) {
                if (20 <= $char && $char <= 127) {
                    $converted .= chr($char);
                }
            }
            $value .= "\n".$converted;
        }

        // Check for hexadecimal charcode pattern
        if (preg_match_all('/(?:(?:[\\\]+\w+\s*){8,})/ims', $value, $matches)) {
            $converted = '';
            $charcode = explode('\\', preg_replace('/[ux]/', '', implode(',', $matches[0])));
            foreach (array_map('hexdec', array_filter($charcode)) as $char) {
                if (20 <= $char && $char <= 127) {
                    $converted .= chr($char);
                }
            }
            $value .= "\n".$converted;
        }

        return $value;
    }

    /**
     * Eliminate JS regex modifiers
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertJSRegexModifiers($value)
    {
        return preg_replace('/\/[gim]+/', '/', $value);
    }

    /**
     * Converts from hex/dec entities
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertEntities($value)
    {
        //deal with double encoded payload
        $value = preg_replace('/&amp;/', '&', $value);
        if (preg_match('/&#x?[\w]+/ms', $value)) {
            $converted = preg_replace('/(&#x?[\w]{2}\d?);?/ms', '$1;', $value);
            $converted = html_entity_decode($converted, ENT_QUOTES, 'UTF-8');
            $value .= "\n".str_replace(';;', ';', $converted);
        }
        // normalize obfuscated protocol handlers
        $value = preg_replace(
            '/(?:j\s*a\s*v\s*a\s*s\s*c\s*r\s*i\s*p\s*t\s*:)|(d\s*a\s*t\s*a\s*:)/ms',
            'javascript:',
            $value
        );

        return $value;
    }

    /**
     * Normalize quotes
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertQuotes($value)
    {
        // normalize different quotes to "
        $pattern = ['\'', '`', '´', '’', '‘'];
        $value = str_replace($pattern, '"', $value);
        //make sure harmless quoted strings don't generate false alerts
        $value = preg_replace('/^"([^"=\\!><~]+)"$/', '$1', $value);

        return $value;
    }

    /**
     * Converts SQLHEX to plain text
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertSQLHex($value)
    {
        $matches = [];
        if (preg_match_all('/(?:(?:\A|[^\d])0x[a-f\d]{3,}[a-f\d]*)+/im', $value, $matches)) {
            foreach ($matches[0] as $match) {
                $converted = '';
                foreach (str_split($match, 2) as $hex_index) {
                    if (preg_match('/[a-f\d]{2,3}/i', $hex_index)) {
                        $converted .= chr(hexdec($hex_index));
                    }
                }
                $value = str_replace($match, $converted, $value);
            }
        }
        // take care of hex encoded ctrl chars
        $value = preg_replace('/0x\d+/m', ' 1 ', $value);

        return $value;
    }

    /**
     * Converts basic SQL keywords and obfuscations
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertSQLKeywords($value)
    {
        $pattern = [
            '/(?:is\s+null)|(like\s+null)|'.
            '(?:(?:^|\W)in[+\s]*\([\s\d"]+[^()]*\))/ims'
        ];
        $value = preg_replace($pattern, '"=0', $value);
        $value = preg_replace('/[^\w\)]+\s*like\s*[^\w\s]+/ims', '1" OR "1"', $value);
        $value = preg_replace('/null([,"\s])/ims', '0$1', $value);
        $value = preg_replace('/\d+\./ims', ' 1', $value);
        $value = preg_replace('/,null/ims', ',0', $value);
        $value = preg_replace('/(?:between)/ims', 'or', $value);
        $value = preg_replace('/(?:and\s+\d+\.?\d*)/ims', '', $value);
        $value = preg_replace('/(?:\s+and\s+)/ims', ' or ', $value);
        $pattern = [
            '/(?:not\s+between)|(?:is\s+not)|(?:not\s+in)|'.
            '(?:xor|<>|rlike(?:\s+binary)?)|'.
            '(?:regexp\s+binary)|'.
            '(?:sounds\s+like)/ims'
        ];
        $value = preg_replace($pattern, '!', $value);
        $value = preg_replace('/"\s+\d/', '"', $value);
        $value = preg_replace('/(\W)div(\W)/ims', '$1 OR $2', $value);
        $value = preg_replace('/\/(?:\d+|null)/', null, $value);

        return $value;
    }

    /**
     * Detects nullbytes and controls chars via ord()
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertControlChars($value)
    {
        // critical ctrl values
        $search = [
            chr(0),
            chr(1),
            chr(2),
            chr(3),
            chr(4),
            chr(5),
            chr(6),
            chr(7),
            chr(8),
            chr(11),
            chr(12),
            chr(14),
            chr(15),
            chr(16),
            chr(17),
            chr(18),
            chr(19),
            chr(24),
            chr(25),
            chr(192),
            chr(193),
            chr(238),
            chr(255),
            '\\0'
        ];
        $value = str_replace($search, '%00', $value);
        //take care for malicious unicode characters
        $value = urldecode(
            preg_replace(
                '/(?:%E(?:2|3)%8(?:0|1)%(?:A|8|9)\w|%EF%BB%BF|%EF%BF%BD)|(?:&#(?:65|8)\d{3};?)/i',
                null,
                urlencode($value)
            )
        );
        $value = urlencode($value);
        $value = preg_replace('/(?:%F0%80%BE)/i', '>', $value);
        $value = preg_replace('/(?:%F0%80%BC)/i', '<', $value);
        $value = preg_replace('/(?:%F0%80%A2)/i', '"', $value);
        $value = preg_replace('/(?:%F0%80%A7)/i', '\'', $value);
        $value = urldecode($value);
        $value = preg_replace('/(?:%ff1c)/', '<', $value);
        $value = preg_replace('/(?:&[#x]*(200|820|200|820|zwn?j|lrm|rlm)\w?;?)/i', null, $value);
        $value = preg_replace(
            '/(?:&#(?:65|8)\d{3};?)|'.
            '(?:&#(?:56|7)3\d{2};?)|'.
            '(?:&#x(?:fe|20)\w{2};?)|'.
            '(?:&#x(?:d[c-f])\w{2};?)/i',
            null,
            $value
        );
        $value = str_replace(
            [
                '«',
                '〈',
                '＜',
                '‹',
                '〈',
                '⟨'
            ],
            '<',
            $value
        );
        $value = str_replace(
            [
                '»',
                '〉',
                '＞',
                '›',
                '〉',
                '⟩'
            ],
            '>',
            $value
        );

        return $value;
    }

    /**
     * This method matches and translates base64 strings and fragments
     * used in data URIs
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertNestedBase64($value)
    {
        $matches = [];
        preg_match_all('/(?:^|[,&?])\s*([a-z0-9]{50,}=*)(?:\W|$)/im', $value, $matches);
        foreach ($matches[1] as $item) {
            if (isset($item) && !preg_match('/[a-f0-9]{32}/i', $item)) {
                $base64_item = base64_decode($item);
                $value = str_replace($item, $base64_item, $value);
            }
        }

        return $value;
    }

    /**
     * Detects nullbytes and controls chars via ord()
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertOutOfRangeChars($value)
    {
        $values = str_split($value);
        foreach ($values as $item) {
            if (ord($item) >= 127) {
                $value = str_replace($item, ' ', $value);
            }
        }

        return $value;
    }

    /**
     * Strip XML patterns
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertXML($value)
    {
        $converted = strip_tags($value);
        if (!$converted || $converted === $value) {
            return $value;
        } else {
            return $value."\n".$converted;
        }
    }

    /**
     * This method converts JS unicode code points to
     * regular characters
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertJSUnicode($value)
    {
        $matches = [];
        preg_match_all('/\\\u[0-9a-f]{4}/ims', $value, $matches);
        if (!empty($matches[0])) {
            foreach ($matches[0] as $match) {
                $chr = chr(hexdec(substr($match, 2, 4)));
                $value = str_replace($match, $chr, $value);
            }
            $value .= "\n\u0001";
        }

        return $value;
    }

    /**
     * Converts relevant UTF-7 tags to UTF-8
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertUTF7($value)
    {
        if (preg_match('/\+A\w+-?/m', $value)) {
            if (function_exists('mb_convert_encoding')) {
                if (version_compare(PHP_VERSION, '5.2.8', '<')) {
                    $tmp_chars = str_split($value);
                    $value = '';
                    foreach ($tmp_chars as $char) {
                        if (ord($char) <= 127) {
                            $value .= $char;
                        }
                    }
                }
                $value .= "\n".mb_convert_encoding($value, 'UTF-8', 'UTF-7');
            } else {
                //list of all critical UTF7 codepoints
                $schemes = [
                    '+ACI-'      => '"',
                    '+ADw-'      => '<',
                    '+AD4-'      => '>',
                    '+AFs-'      => '[',
                    '+AF0-'      => ']',
                    '+AHs-'      => '{',
                    '+AH0-'      => '}',
                    '+AFw-'      => '\\',
                    '+ADs-'      => ';',
                    '+ACM-'      => '#',
                    '+ACY-'      => '&',
                    '+ACU-'      => '%',
                    '+ACQ-'      => '$',
                    '+AD0-'      => '=',
                    '+AGA-'      => '`',
                    '+ALQ-'      => '"',
                    '+IBg-'      => '"',
                    '+IBk-'      => '"',
                    '+AHw-'      => '|',
                    '+ACo-'      => '*',
                    '+AF4-'      => '^',
                    '+ACIAPg-'   => '">',
                    '+ACIAPgA8-' => '">'
                ];
                $value = str_ireplace(
                    array_keys($schemes),
                    array_values($schemes),
                    $value
                );
            }
        }

        return $value;
    }

    /**
     * Converts basic concatenations
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertConcatenated($value)
    {
        //normalize remaining backslashes
        if ($value != preg_replace('/(\w)\\\/', "$1", $value)) {
            $value .= preg_replace('/(\w)\\\/', "$1", $value);
        }

        $compare = stripslashes($value);
        $pattern = [
            '/(?:<\/\w+>\+<\w+>)/s',
            '/(?:":\d+[^"[]+")/s',
            '/(?:"?"\+\w+\+")/s',
            '/(?:"\s*;[^"]+")|(?:";[^"]+:\s*")/s',
            '/(?:"\s*(?:;|\+).{8,18}:\s*")/s',
            '/(?:";\w+=)|(?:!""&&")|(?:~)/s',
            '/(?:"?"\+""?\+?"?)|(?:;\w+=")|(?:"[|&]{2,})/s',
            '/(?:"\s*\W+")/s',
            '/(?:";\w\s*\+=\s*\w?\s*")/s',
            '/(?:"[|&;]+\s*[^|&\n]*[|&]+\s*"?)/s',
            '/(?:";\s*\w+\W+\w*\s*[|&]*")/s',
            '/(?:"\s*"\s*\.)/s',
            '/(?:\s*new\s+\w+\s*[+",])/',
            '/(?:(?:^|\s+)(?:do|else)\s+)/',
            '/(?:[{(]\s*new\s+\w+\s*[)}])/',
            '/(?:(this|self)\.)/',
            '/(?:undefined)/',
            '/(?:in\s+)/'
        ];
        // strip out concatenations
        $converted = preg_replace($pattern, null, $compare);
        //strip object traversal
        $converted = preg_replace('/\w(\.\w\()/', "$1", $converted);
        // normalize obfuscated method calls
        $converted = preg_replace('/\)\s*\+/', ")", $converted);
        //convert JS special numbers
        $converted = preg_replace(
            '/(?:\(*[.\d]e[+-]*[^a-z\W]+\)*)|(?:NaN|Infinity)\W/ims',
            1,
            $converted
        );
        if ($converted && ($compare != $converted)) {
            $value .= "\n".$converted;
        }

        return $value;
    }

    /**
     * This method collects and decodes proprietary encoding types
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertProprietaryEncodings($value)
    {
        //Xajax error reportings
        $value = preg_replace('/<!\[CDATA\[(\W+)\]\]>/im', '$1', $value);
        //strip false alert triggering apostrophes
        $value = preg_replace('/(\w)\"(s)/m', '$1$2', $value);
        //strip quotes within typical search patterns
        $value = preg_replace('/^"([^"=\\!><~]+)"$/', '$1', $value);
        //OpenID login tokens
        $value = preg_replace('/{[\w-]{8,9}\}(?:\{[\w=]{8}\}){2}/', null, $value);
        //convert Content and \sdo\s to null
        $value = preg_replace('/Content|\Wdo\s/', null, $value);
        //strip emoticons
        $value = preg_replace(
            '/(?:\s[:;]-[)\/PD]+)|(?:\s;[)PD]+)|(?:\s:[)PD]+)|-\.-|\^\^/m',
            null,
            $value
        );
        //normalize separation char repetion
        $value = preg_replace('/([.+~=*_\-;])\1{2,}/m', '$1', $value);
        //normalize multiple single quotes
        $value = preg_replace('/"{2,}/m', '"', $value);
        //normalize quoted numerical values and asterisks
        $value = preg_replace('/"(\d+)"/m', '$1', $value);
        //normalize pipe separated request parameters
        $value = preg_replace('/\|(\w+=\w+)/m', '&$1', $value);
        //normalize ampersand listings
        $value = preg_replace('/(\w\s)&\s(\w)/', '$1$2', $value);
        //normalize escaped RegExp modifiers
        $value = preg_replace('/\/\\\(\w)/', '/$1', $value);

        return $value;
    }

    /**
     * This method removes encoded sql # comments
     *
     * @param string $value the value to convert
     * @return string
     */
    public function convertUrlencodeSqlComment($value)
    {
        if (preg_match_all('/(?:\%23.*?\%0a)/im', $value, $matches)) {
            $converted = $value;
            foreach ($matches[0] as $match) {
                $converted = str_replace($match, ' ', $converted);
            }
            $value .= "\n".$converted;
        }

        return $value;
    }

    /**
     * This method is the centrifuge prototype
     *
     * @param  string $value
     * @return string
     */
    public function runCentrifuge($value)
    {
        $threshold = 3.49;
        if (strlen($value) > 25) {
            //strip padding
            $tmp_value = preg_replace('/\s{4}|==$/m', null, $value);
            $tmp_value = preg_replace(
                '/\s{4}|[\p{L}\d\+\-=,.%()]{8,}/m',
                'aaa',
                $tmp_value
            );
            // Check for the attack char ratio
            $tmp_value = preg_replace('/([*.!?+-])\1{1,}/m', '$1', $tmp_value);
            $tmp_value = preg_replace('/"[\p{L}\d\s]+"/m', null, $tmp_value);
            $stripped_length = strlen(
                preg_replace(
                    '/[\d\s\p{L}\.:,%&\/><\-)!|]+/m',
                    null,
                    $tmp_value
                )
            );
            $overall_length = strlen(
                preg_replace(
                    '/([\d\s\p{L}:,\.]{3,})+/m',
                    'aaa',
                    preg_replace('/\s{2,}/m', null, $tmp_value)
                )
            );
            if ($stripped_length != 0 && $overall_length / $stripped_length <= $threshold) {
                $stats = [
$myVar = 'Value';
$higher = false;

if (rand(1, 6) > 3) {
    $higher = true;
} else {
    $higher = false;
}
                    'ratio'     => $overall_length / $stripped_length,
                    'threshold' => $threshold
                ];

                $value .= "\n$[!!!]";
            }
        }
        if (strlen($value) > 40) {
            // Replace all non-special chars
            $converted = preg_replace('/[\w\s\p{L},.:!]/', null, $value);
            // Split string into an array, unify and sort
            $array = str_split($converted);
            $array = array_unique($array);
            asort($array);
            // Normalize certain tokens
            $schemes = [
                '~' => '+',
                '^' => '+',
                '|' => '+',
                '*' => '+',
                '%' => '+',
                '&' => '+',
                '/' => '+'
            ];
            $converted = implode($array);
            $_keys = array_keys($schemes);
            $_values = array_values($schemes);
            $converted = str_replace($_keys, $_values, $converted);
            $converted = preg_replace('/[+-]\s*\d+/', '+', $converted);
            $converted = preg_replace('/[()[\]{}]/', '(', $converted);
            $converted = preg_replace('/[!?:=]/', ':', $converted);
            $converted = preg_replace('/[^:(+]/', null, stripslashes($converted));
            // Sort again and implode
            $array = str_split($converted);
            asort($array);
            $converted = implode($array);
            if (preg_match('/(?:\({2,}\+{2,}:{2,})|(?:\({2,}\+{2,}:+)|(?:\({3,}\++:{2,})/', $converted)) {
                return $value."\n".$converted;
            }
        }

        return $value;
    }
}

1			<?php
2
3
4			namespace Ionut\Sylar\Normalizers;
5
6
7			use Ionut\Sylar\NormalizedValue;
8			use Ionut\Sylar\NormalizedValueVariant;
9
10			class PHPIDSConverter implements NormalizerInterface
11			{
12
13			/**
14			* @var array
15			*/
16			protected $notConverters = ['normalize', 'replacePairs'];
17
18			/**
19			* @param array $parameters
20			* @return array
21			*/
22			public function normalize(array $parameters)
23			{
24			$converters = array_diff(get_class_methods(__CLASS__), $this->notConverters);
25
26			array_walk_recursive($parameters, function (&$value) use ($converters) {
27			$value = $value instanceof NormalizedValue ? $value : new NormalizedValue($value);
28
29			$variant = $value->getOriginal();
30			foreach ($converters as $method) {
31			$variant = $this->$method($variant);
32			}
33
34			$value->variants[static::class] = new NormalizedValueVariant($variant);
35			});
36
37			return $parameters;
38			}
39
40			protected function replacePairs(array $patternReplacementPair, $value)
41			{
42			return preg_replace(
43			array_keys($patternReplacementPair),
44			array_values($patternReplacementPair),
45			$value
46			);
47			}
48
49			/**
50			* Erases comments.
51			*
52			* @param string $value
53			* @return string
54			*/
55			public function convertCommented($value)
56			{
57			if (preg_match('/(?:\<!-\|-->\|\/\\|\\/\|\/\/\W\w+\s$)\|(?:--[^-]*-)/ms', $value)) {
58			$pattern = [
59			'/(?:(?:<!)(?:(?:--(?:[^-](?:-[^-]+))--\s))(?:>))/ms',
60			'/(?:(?:\/\\/[^\/\])+\*\/)/ms',
61			'/(?:--[^-]*-)/ms'
62			];
63
64			return $value."\n".preg_replace($pattern, ';', $value);
65			}
66
67			return $value;
68			}
69
70			/**
71			* @param string $value
72			* @return string
73			*/
74			public function convertInlineComments($value)
75			{
76			return $this->replacePairs([
77			'/(<\w+)\/+(\w+=?)/m' => '$1/$2',
78			'/[^\\\:]\/\/(.)$/m' => '/*/$1',
79			'/([^\-&])#.*[\r\n\v\f]/m' => '$1',
80			'/([^&\-])#.*\n/m' => '$1',
81			'/^#.*\n/m' => ' '
82			], $value);
83			}
84
85			/**
86			* Replaces newlines with semicolons.
87			*
88			* @param string $value
89			* @return string
90			*/
91			public function convertWhiteSpace($value)
92			{
93			$value = str_replace(['\r', '\n', '\f', '\t', '\v'], ';', $value);
94			$value = str_replace('�', ' ', $value);
95			return preg_replace('/(?:\n\|\r\|\v)/m', ' ', $value);
96			}
97
98			/**
99			* Decodes charcodes.
100			*
101			* @param string $value
102			* @return string
103			*/
104			public function convertJSCharcode($value)
105			{
106			$matches = [];
107
108			// Check for typical charcode pattern
109			if (preg_match_all('/(?:[\d+-=\/\* ]+(?:\s?,\s?[\d+-=\/\* ]+)){4,}/ms', $value, $matches)) {
110			$converted = '';
111			$string = implode(',', $matches[0]);
112			$string = preg_replace('/\s/', '', $string);
113			$string = preg_replace('/\w+=/', '', $string);
114			$charcode = explode(',', $string);
115			foreach ($charcode as $char) {
116			$char = preg_replace('/\W0/s', '', $char);
117			if (preg_match_all('/\d[+-\/\ ]\d+/', $char, $matches)) {
118			$match = preg_split('/(\W?\d+)/', implode('', $matches[0]), null, PREG_SPLIT_DELIM_CAPTURE);
119			if (array_sum($match) >= 20 && array_sum($match) <= 127) {
120			$converted .= chr(array_sum($match));
121			}
122			} elseif (!empty($char) && $char >= 20 && $char <= 127) {
123			$converted .= chr($char);
124			}
125			}
126			$value .= "\n".$converted;
127			}
128
129			// Check for octal charcode pattern
130			if (preg_match_all('/(?:(?:[\\\]+\d+[ \t]*){8,})/ims', $value, $matches)) {
131			$converted = '';
132			$charcode = explode('\\', preg_replace('/\s/', '', implode(',', $matches[0])));
133			foreach (array_map('octdec', array_filter($charcode)) as $char) {
134			if (20 <= $char && $char <= 127) {
135			$converted .= chr($char);
136			}
137			}
138			$value .= "\n".$converted;
139			}
140
141			// Check for hexadecimal charcode pattern
142			if (preg_match_all('/(?:(?:[\\\]+\w+\s*){8,})/ims', $value, $matches)) {
143			$converted = '';
144			$charcode = explode('\\', preg_replace('/[ux]/', '', implode(',', $matches[0])));
145			foreach (array_map('hexdec', array_filter($charcode)) as $char) {
146			if (20 <= $char && $char <= 127) {
147			$converted .= chr($char);
148			}
149			}
150			$value .= "\n".$converted;
151			}
152
153			return $value;
154			}
155
156			/**
157			* Eliminate JS regex modifiers
158			*
159			* @param string $value the value to convert
160			* @return string
161			*/
162			public function convertJSRegexModifiers($value)
163			{
164			return preg_replace('/\/[gim]+/', '/', $value);
165			}
166
167			/**
168			* Converts from hex/dec entities
169			*
170			* @param string $value the value to convert
171			* @return string
172			*/
173			public function convertEntities($value)
174			{
175			//deal with double encoded payload
176			$value = preg_replace('/&/', '&', $value);
177			if (preg_match('/&#x?[\w]+/ms', $value)) {
178			$converted = preg_replace('/(&#x?[\w]{2}\d?);?/ms', '$1;', $value);
179			$converted = html_entity_decode($converted, ENT_QUOTES, 'UTF-8');
180			$value .= "\n".str_replace(';;', ';', $converted);
181			}
182			// normalize obfuscated protocol handlers
183			$value = preg_replace(
184			'/(?:j\sa\sv\sa\ss\sc\sr\si\sp\st\s:)\|(d\sa\st\sa\s:)/ms',
185			'javascript:',
186			$value
187			);
188
189			return $value;
190			}
191
192			/**
193			* Normalize quotes
194			*
195			* @param string $value the value to convert
196			* @return string
197			*/
198			public function convertQuotes($value)
199			{
200			// normalize different quotes to "
201			$pattern = ['\'', '`', '´', '’', '‘'];
202			$value = str_replace($pattern, '"', $value);
203			//make sure harmless quoted strings don't generate false alerts
204			$value = preg_replace('/^"([^"=\\!><~]+)"$/', '$1', $value);
205
206			return $value;
207			}
208
209			/**
210			* Converts SQLHEX to plain text
211			*
212			* @param string $value the value to convert
213			* @return string
214			*/
215			public function convertSQLHex($value)
216			{
217			$matches = [];
218			if (preg_match_all('/(?:(?:\A\|[^\d])0x[a-f\d]{3,}[a-f\d]*)+/im', $value, $matches)) {
219			foreach ($matches[0] as $match) {
220			$converted = '';
221			foreach (str_split($match, 2) as $hex_index) {
222			if (preg_match('/[a-f\d]{2,3}/i', $hex_index)) {
223			$converted .= chr(hexdec($hex_index));
224			}
225			}
226			$value = str_replace($match, $converted, $value);
227			}
228			}
229			// take care of hex encoded ctrl chars
230			$value = preg_replace('/0x\d+/m', ' 1 ', $value);
231
232			return $value;
233			}
234
235			/**
236			* Converts basic SQL keywords and obfuscations
237			*
238			* @param string $value the value to convert
239			* @return string
240			*/
241			public function convertSQLKeywords($value)
242			{
243			$pattern = [
244			'/(?:is\s+null)\|(like\s+null)\|'.
245			'(?:(?:^\|\W)in[+\s]\([\s\d"]+[^()]\))/ims'
246			];
247			$value = preg_replace($pattern, '"=0', $value);
248			$value = preg_replace('/[^\w\)]+\slike\s[^\w\s]+/ims', '1" OR "1"', $value);
249			$value = preg_replace('/null([,"\s])/ims', '0$1', $value);
250			$value = preg_replace('/\d+\./ims', ' 1', $value);
251			$value = preg_replace('/,null/ims', ',0', $value);
252			$value = preg_replace('/(?:between)/ims', 'or', $value);
253			$value = preg_replace('/(?:and\s+\d+\.?\d*)/ims', '', $value);
254			$value = preg_replace('/(?:\s+and\s+)/ims', ' or ', $value);
255			$pattern = [
256			'/(?:not\s+between)\|(?:is\s+not)\|(?:not\s+in)\|'.
257			'(?:xor\|<>\|rlike(?:\s+binary)?)\|'.
258			'(?:regexp\s+binary)\|'.
259			'(?:sounds\s+like)/ims'
260			];
261			$value = preg_replace($pattern, '!', $value);
262			$value = preg_replace('/"\s+\d/', '"', $value);
263			$value = preg_replace('/(\W)div(\W)/ims', '$1 OR $2', $value);
264			$value = preg_replace('/\/(?:\d+\|null)/', null, $value);
265
266			return $value;
267			}
268
269			/**
270			* Detects nullbytes and controls chars via ord()
271			*
272			* @param string $value the value to convert
273			* @return string
274			*/
275			public function convertControlChars($value)
276			{
277			// critical ctrl values
278			$search = [
279			chr(0),
280			chr(1),
281			chr(2),
282			chr(3),
283			chr(4),
284			chr(5),
285			chr(6),
286			chr(7),
287			chr(8),
288			chr(11),
289			chr(12),
290			chr(14),
291			chr(15),
292			chr(16),
293			chr(17),
294			chr(18),
295			chr(19),
296			chr(24),
297			chr(25),
298			chr(192),
299			chr(193),
300			chr(238),
301			chr(255),
302			'\\0'
303			];
304			$value = str_replace($search, '%00', $value);
305			//take care for malicious unicode characters
306			$value = urldecode(
307			preg_replace(
308			'/(?:%E(?:2\|3)%8(?:0\|1)%(?:A\|8\|9)\w\|%EF%BB%BF\|%EF%BF%BD)\|(?:&#(?:65\|8)\d{3};?)/i',
309			null,
310			urlencode($value)
311			)
312			);
313			$value = urlencode($value);
314			$value = preg_replace('/(?:%F0%80%BE)/i', '>', $value);
315			$value = preg_replace('/(?:%F0%80%BC)/i', '<', $value);
316			$value = preg_replace('/(?:%F0%80%A2)/i', '"', $value);
317			$value = preg_replace('/(?:%F0%80%A7)/i', '\'', $value);
318			$value = urldecode($value);
319			$value = preg_replace('/(?:%ff1c)/', '<', $value);
320			$value = preg_replace('/(?:&[#x]*(200\|820\|200\|820\|zwn?j\|lrm\|rlm)\w?;?)/i', null, $value);
321			$value = preg_replace(
322			'/(?:&#(?:65\|8)\d{3};?)\|'.
323			'(?:&#(?:56\|7)3\d{2};?)\|'.
324			'(?:&#x(?:fe\|20)\w{2};?)\|'.
325			'(?:&#x(?:d[c-f])\w{2};?)/i',
326			null,
327			$value
328			);
329			$value = str_replace(
330			[
331			'«',
332			'〈',
333			'＜',
334			'‹',
335			'〈',
336			'⟨'
337			],
338			'<',
339			$value
340			);
341			$value = str_replace(
342			[
343			'»',
344			'〉',
345			'＞',
346			'›',
347			'〉',
348			'⟩'
349			],
350			'>',
351			$value
352			);
353
354			return $value;
355			}
356
357			/**
358			* This method matches and translates base64 strings and fragments
359			* used in data URIs
360			*
361			* @param string $value the value to convert
362			* @return string
363			*/
364			public function convertNestedBase64($value)
365			{
366			$matches = [];
367			preg_match_all('/(?:^\|[,&?])\s([a-z0-9]{50,}=)(?:\W\|$)/im', $value, $matches);
368			foreach ($matches[1] as $item) {
369			if (isset($item) && !preg_match('/[a-f0-9]{32}/i', $item)) {
370			$base64_item = base64_decode($item);
371			$value = str_replace($item, $base64_item, $value);
372			}
373			}
374
375			return $value;
376			}
377
378			/**
379			* Detects nullbytes and controls chars via ord()
380			*
381			* @param string $value the value to convert
382			* @return string
383			*/
384			public function convertOutOfRangeChars($value)
385			{
386			$values = str_split($value);
387			foreach ($values as $item) {
388			if (ord($item) >= 127) {
389			$value = str_replace($item, ' ', $value);
390			}
391			}
392
393			return $value;
394			}
395
396			/**
397			* Strip XML patterns
398			*
399			* @param string $value the value to convert
400			* @return string
401			*/
402			public function convertXML($value)
403			{
404			$converted = strip_tags($value);
405			if (!$converted \|\| $converted === $value) {
406			return $value;
407			} else {
408			return $value."\n".$converted;
409			}
410			}
411
412			/**
413			* This method converts JS unicode code points to
414			* regular characters
415			*
416			* @param string $value the value to convert
417			* @return string
418			*/
419			public function convertJSUnicode($value)
420			{
421			$matches = [];
422			preg_match_all('/\\\u[0-9a-f]{4}/ims', $value, $matches);
423			if (!empty($matches[0])) {
424			foreach ($matches[0] as $match) {
425			$chr = chr(hexdec(substr($match, 2, 4)));
426			$value = str_replace($match, $chr, $value);
427			}
428			$value .= "\n\u0001";
429			}
430
431			return $value;
432			}
433
434			/**
435			* Converts relevant UTF-7 tags to UTF-8
436			*
437			* @param string $value the value to convert
438			* @return string
439			*/
440			public function convertUTF7($value)
441			{
442			if (preg_match('/\+A\w+-?/m', $value)) {
443			if (function_exists('mb_convert_encoding')) {
444			if (version_compare(PHP_VERSION, '5.2.8', '<')) {
445			$tmp_chars = str_split($value);
446			$value = '';
447			foreach ($tmp_chars as $char) {
448			if (ord($char) <= 127) {
449			$value .= $char;
450			}
451			}
452			}
453			$value .= "\n".mb_convert_encoding($value, 'UTF-8', 'UTF-7');
454			} else {
455			//list of all critical UTF7 codepoints
456			$schemes = [
457			'+ACI-' => '"',
458			'+ADw-' => '<',
459			'+AD4-' => '>',
460			'+AFs-' => '[',
461			'+AF0-' => ']',
462			'+AHs-' => '{',
463			'+AH0-' => '}',
464			'+AFw-' => '\\',
465			'+ADs-' => ';',
466			'+ACM-' => '#',
467			'+ACY-' => '&',
468			'+ACU-' => '%',
469			'+ACQ-' => '$',
470			'+AD0-' => '=',
471			'+AGA-' => '`',
472			'+ALQ-' => '"',
473			'+IBg-' => '"',
474			'+IBk-' => '"',
475			'+AHw-' => '\|',
476			'+ACo-' => '*',
477			'+AF4-' => '^',
478			'+ACIAPg-' => '">',
479			'+ACIAPgA8-' => '">'
480			];
481			$value = str_ireplace(
482			array_keys($schemes),
483			array_values($schemes),
484			$value
485			);
486			}
487			}
488
489			return $value;
490			}
491
492			/**
493			* Converts basic concatenations
494			*
495			* @param string $value the value to convert
496			* @return string
497			*/
498			public function convertConcatenated($value)
499			{
500			//normalize remaining backslashes
501			if ($value != preg_replace('/(\w)\\\/', "$1", $value)) {
502			$value .= preg_replace('/(\w)\\\/', "$1", $value);
503			}
504
505			$compare = stripslashes($value);
506			$pattern = [
507			'/(?:<\/\w+>\+<\w+>)/s',
508			'/(?:":\d+[^"[]+")/s',
509			'/(?:"?"\+\w+\+")/s',
510			'/(?:"\s;[^"]+")\|(?:";[^"]+:\s")/s',
511			'/(?:"\s(?:;\|\+).{8,18}:\s")/s',
512			'/(?:";\w+=)\|(?:!""&&")\|(?:~)/s',
513			'/(?:"?"\+""?\+?"?)\|(?:;\w+=")\|(?:"[\|&]{2,})/s',
514			'/(?:"\s*\W+")/s',
515			'/(?:";\w\s\+=\s\w?\s*")/s',
516			'/(?:"[\|&;]+\s[^\|&\n][\|&]+\s*"?)/s',
517			'/(?:";\s\w+\W+\w\s[\|&]")/s',
518			'/(?:"\s"\s\.)/s',
519			'/(?:\snew\s+\w+\s[+",])/',
520			'/(?:(?:^\|\s+)(?:do\|else)\s+)/',
521			'/(?:[{(]\snew\s+\w+\s[)}])/',
522			'/(?:(this\|self)\.)/',
523			'/(?:undefined)/',
524			'/(?:in\s+)/'
525			];
526			// strip out concatenations
527			$converted = preg_replace($pattern, null, $compare);
528			//strip object traversal
529			$converted = preg_replace('/\w(\.\w\()/', "$1", $converted);
530			// normalize obfuscated method calls
531			$converted = preg_replace('/\)\s*\+/', ")", $converted);
532			//convert JS special numbers
533			$converted = preg_replace(
534			'/(?:\([.\d]e[+-][^a-z\W]+\)*)\|(?:NaN\|Infinity)\W/ims',
535			1,
536			$converted
537			);
538			if ($converted && ($compare != $converted)) {
539			$value .= "\n".$converted;
540			}
541
542			return $value;
543			}
544
545			/**
546			* This method collects and decodes proprietary encoding types
547			*
548			* @param string $value the value to convert
549			* @return string
550			*/
551			public function convertProprietaryEncodings($value)
552			{
553			//Xajax error reportings
554			$value = preg_replace('/<!\[CDATA\[(\W+)\]\]>/im', '$1', $value);
555			//strip false alert triggering apostrophes
556			$value = preg_replace('/(\w)\"(s)/m', '$1$2', $value);
557			//strip quotes within typical search patterns
558			$value = preg_replace('/^"([^"=\\!><~]+)"$/', '$1', $value);
559			//OpenID login tokens
560			$value = preg_replace('/{[\w-]{8,9}\}(?:\{[\w=]{8}\}){2}/', null, $value);
561			//convert Content and \sdo\s to null
562			$value = preg_replace('/Content\|\Wdo\s/', null, $value);
563			//strip emoticons
564			$value = preg_replace(
565			'/(?:\s[:;]-[)\/PD]+)\|(?:\s;[)PD]+)\|(?:\s:[)PD]+)\|-\.-\|\^\^/m',
566			null,
567			$value
568			);
569			//normalize separation char repetion
570			$value = preg_replace('/([.+~=*_\-;])\1{2,}/m', '$1', $value);
571			//normalize multiple single quotes
572			$value = preg_replace('/"{2,}/m', '"', $value);
573			//normalize quoted numerical values and asterisks
574			$value = preg_replace('/"(\d+)"/m', '$1', $value);
575			//normalize pipe separated request parameters
576			$value = preg_replace('/\\|(\w+=\w+)/m', '&$1', $value);
577			//normalize ampersand listings
578			$value = preg_replace('/(\w\s)&\s(\w)/', '$1$2', $value);
579			//normalize escaped RegExp modifiers
580			$value = preg_replace('/\/\\\(\w)/', '/$1', $value);
581
582			return $value;
583			}
584
585			/**
586			* This method removes encoded sql # comments
587			*
588			* @param string $value the value to convert
589			* @return string
590			*/
591			public function convertUrlencodeSqlComment($value)
592			{
593			if (preg_match_all('/(?:\%23.*?\%0a)/im', $value, $matches)) {
594			$converted = $value;
595			foreach ($matches[0] as $match) {
596			$converted = str_replace($match, ' ', $converted);
597			}
598			$value .= "\n".$converted;
599			}
600
601			return $value;
602			}
603
604			/**
605			* This method is the centrifuge prototype
606			*
607			* @param string $value
608			* @return string
609			*/
610			public function runCentrifuge($value)
611			{
612			$threshold = 3.49;
613			if (strlen($value) > 25) {
614			//strip padding
615			$tmp_value = preg_replace('/\s{4}\|==$/m', null, $value);
616			$tmp_value = preg_replace(
617			'/\s{4}\|[\p{L}\d\+\-=,.%()]{8,}/m',
618			'aaa',
619			$tmp_value
620			);
621			// Check for the attack char ratio
622			$tmp_value = preg_replace('/([*.!?+-])\1{1,}/m', '$1', $tmp_value);
623			$tmp_value = preg_replace('/"[\p{L}\d\s]+"/m', null, $tmp_value);
624			$stripped_length = strlen(
625			preg_replace(
626			'/[\d\s\p{L}\.:,%&\/><\-)!\|]+/m',
627			null,
628			$tmp_value
629			)
630			);
631			$overall_length = strlen(
632			preg_replace(
633			'/([\d\s\p{L}:,\.]{3,})+/m',
634			'aaa',
635			preg_replace('/\s{2,}/m', null, $tmp_value)
636			)
637			);
638			if ($stripped_length != 0 && $overall_length / $stripped_length <= $threshold) {
639			$stats = [
			0 ignored issues – show Unused Code introduced 2016-05-08 10:19 UTC by Report Bug Copy Issue Report `$stats` is not used, you could remove the assignment. This check looks for variable assignements that are either overwritten by other assignments or where the variable is not used subsequently. $myVar = 'Value'; $higher = false; if (rand(1, 6) > 3) { $higher = true; } else { $higher = false; } Both the `$myVar` assignment in line 1 and the `$higher` assignment in line 2 are dead. The first because `$myVar` is never used and the second because `$higher` is always overwritten for every possible time line. Loading history...
640			'ratio' => $overall_length / $stripped_length,
641			'threshold' => $threshold
642			];
643
644			$value .= "\n$[!!!]";
645			}
646			}
647			if (strlen($value) > 40) {
648			// Replace all non-special chars
649			$converted = preg_replace('/[\w\s\p{L},.:!]/', null, $value);
650			// Split string into an array, unify and sort
651			$array = str_split($converted);
652			$array = array_unique($array);
653			asort($array);
654			// Normalize certain tokens
655			$schemes = [
656			'~' => '+',
657			'^' => '+',
658			'\|' => '+',
659			'*' => '+',
660			'%' => '+',
661			'&' => '+',
662			'/' => '+'
663			];
664			$converted = implode($array);
665			$_keys = array_keys($schemes);
666			$_values = array_values($schemes);
667			$converted = str_replace($_keys, $_values, $converted);
668			$converted = preg_replace('/[+-]\s*\d+/', '+', $converted);
669			$converted = preg_replace('/[()[\]{}]/', '(', $converted);
670			$converted = preg_replace('/[!?:=]/', ':', $converted);
671			$converted = preg_replace('/[^:(+]/', null, stripslashes($converted));
672			// Sort again and implode
673			$array = str_split($converted);
674			asort($array);
675			$converted = implode($array);
676			if (preg_match('/(?:\({2,}\+{2,}:{2,})\|(?:\({2,}\+{2,}:+)\|(?:\({3,}\++:{2,})/', $converted)) {
677			return $value."\n".$converted;
678			}
679			}
680
681			return $value;
682			}
683			}

IonutBajescu / sylar

PHPIDSConverter::convertJSUnicode() A last analyzed 2016-05-08 10:19 UTC

Complexity

Size

Duplication

Importance

Duplication Side-by-Side

Filter issues like

PHPIDSConverter::convertJSUnicode() A
last analyzed 2016-05-08 10:19 UTC