Charset::instance() - Code Metrics - Inspection of "add test on php 7.4" - gggeek/phpxmlrpc - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 87b7a4...6ce28d )

by Gaetano

created 2020-03-04 09:10 UTC

Charset::instance() A

↳ Parent: Charset

Complexity

Conditions	2
Paths	2

Size

Total Lines	7
Code Lines	3

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	4
CRAP Score	2

Importance

Changes	1
Bugs	0	Features	0

Metric	Value
cc	2
eloc	3
c	1
b	0
f	0
nc	2
nop	0
dl	0
loc	7
rs	10
ccs	4
cts	4
cp	1
crap	2

<?php

namespace PhpXmlRpc\Helper;

use PhpXmlRpc\PhpXmlRpc;

class Charset
{
    // tables used for transcoding different charsets into us-ascii xml
    protected $xml_iso88591_Entities = array("in" => array(), "out" => array());

    /// @todo add to iso table the characters from cp_1252 range, i.e. 128 to 159?
    /// These will NOT be present in true ISO-8859-1, but will save the unwary windows user from sending junk
    /// (though no luck when receiving them...)
    /*
    protected $xml_cp1252_Entities = array('in' => array(), out' => array(
        '&#x20AC;', '?',        '&#x201A;', '&#x0192;',
        '&#x201E;', '&#x2026;', '&#x2020;', '&#x2021;',
        '&#x02C6;', '&#x2030;', '&#x0160;', '&#x2039;',
        '&#x0152;', '?',        '&#x017D;', '?',
        '?',        '&#x2018;', '&#x2019;', '&#x201C;',
        '&#x201D;', '&#x2022;', '&#x2013;', '&#x2014;',
        '&#x02DC;', '&#x2122;', '&#x0161;', '&#x203A;',
        '&#x0153;', '?',        '&#x017E;', '&#x0178;'
    ));
    */

    protected $charset_supersets = array(
        'US-ASCII' => array('ISO-8859-1', 'ISO-8859-2', 'ISO-8859-3', 'ISO-8859-4',
            'ISO-8859-5', 'ISO-8859-6', 'ISO-8859-7', 'ISO-8859-8',
            'ISO-8859-9', 'ISO-8859-10', 'ISO-8859-11', 'ISO-8859-12',
            'ISO-8859-13', 'ISO-8859-14', 'ISO-8859-15', 'UTF-8',
            'EUC-JP', 'EUC-', 'EUC-KR', 'EUC-CN',),
    );

    protected static $instance = null;

    /**
     * This class is singleton for performance reasons.
     *
     * @return Charset
     */
    public static function instance()
    {
        if (self::$instance === null) {
            self::$instance = new self();
        }

        return self::$instance;
    }

    private function __construct()
    {
        for ($i = 0; $i < 32; $i++) {
            $this->xml_iso88591_Entities["in"][] = chr($i);
            $this->xml_iso88591_Entities["out"][] = "&#{$i};";
        }

        for ($i = 160; $i < 256; $i++) {
            $this->xml_iso88591_Entities["in"][] = chr($i);
            $this->xml_iso88591_Entities["out"][] = "&#{$i};";
        }

        /*for ($i = 128; $i < 160; $i++)
        {
            $this->xml_cp1252_Entities['in'][] = chr($i);
        }*/
    }

    /**
     * Convert a string to the correct XML representation in a target charset.
     *
     * To help correct communication of non-ascii chars inside strings, regardless of the charset used when sending
     * requests, parsing them, sending responses and parsing responses, an option is to convert all non-ascii chars
     * present in the message into their equivalent 'charset entity'. Charset entities enumerated this way are
     * independent of the charset encoding used to transmit them, and all XML parsers are bound to understand them.
     * Note that in the std case we are not sending a charset encoding mime type along with http headers, so we are
     * bound by RFC 3023 to emit strict us-ascii.
     *
     * @todo do a bit of basic benchmarking (strtr vs. str_replace)
     * @todo make usage of iconv() or recode_string() or mb_string() where available
     *
     * @param string $data
     * @param string $srcEncoding
     * @param string $destEncoding
     *
     * @return string
     */
    public function encodeEntities($data, $srcEncoding = '', $destEncoding = '')
    {
        if ($srcEncoding == '') {
            // lame, but we know no better...
            $srcEncoding = PhpXmlRpc::$xmlrpc_internalencoding;
        }

        $conversion = strtoupper($srcEncoding . '_' . $destEncoding);
        switch ($conversion) {
            case 'ISO-8859-1_':
            case 'ISO-8859-1_US-ASCII':
                $escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&amp;', '&quot;', '&apos;', '&lt;', '&gt;'), $data);
                $escapedData = str_replace($this->xml_iso88591_Entities['in'], $this->xml_iso88591_Entities['out'], $escapedData);
                break;

            case 'ISO-8859-1_UTF-8':
                $escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&amp;', '&quot;', '&apos;', '&lt;', '&gt;'), $data);
                $escapedData = utf8_encode($escapedData);
                break;

            case 'ISO-8859-1_ISO-8859-1':
            case 'US-ASCII_US-ASCII':
            case 'US-ASCII_UTF-8':
            case 'US-ASCII_':
            case 'US-ASCII_ISO-8859-1':
            case 'UTF-8_UTF-8':
            //case 'CP1252_CP1252':
                $escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&amp;', '&quot;', '&apos;', '&lt;', '&gt;'), $data);
                break;

            case 'UTF-8_':
            case 'UTF-8_US-ASCII':
            case 'UTF-8_ISO-8859-1':
                // NB: this will choke on invalid UTF-8, going most likely beyond EOF
                $escapedData = '';
                // be kind to users creating string xmlrpc values out of different php types
                $data = (string)$data;
                $ns = strlen($data);
                for ($nn = 0; $nn < $ns; $nn++) {
                    $ch = $data[$nn];
                    $ii = ord($ch);
                    // 7 bits: 0bbbbbbb (127)
                    if ($ii < 128) {
                        /// @todo shall we replace this with a (supposedly) faster str_replace?
                        switch ($ii) {
                            case 34:
                                $escapedData .= '&quot;';
                                break;
                            case 38:
                                $escapedData .= '&amp;';
                                break;
                            case 39:
                                $escapedData .= '&apos;';
                                break;
                            case 60:
                                $escapedData .= '&lt;';
                                break;
                            case 62:
                                $escapedData .= '&gt;';
                                break;
                            default:
                                $escapedData .= $ch;
                        } // switch
                    } // 11 bits: 110bbbbb 10bbbbbb (2047)
                    elseif ($ii >> 5 == 6) {
                        $b1 = ($ii & 31);
                        $ii = ord($data[$nn + 1]);
                        $b2 = ($ii & 63);
                        $ii = ($b1 * 64) + $b2;
                        $ent = sprintf('&#%d;', $ii);
                        $escapedData .= $ent;
                        $nn += 1;
                    } // 16 bits: 1110bbbb 10bbbbbb 10bbbbbb
                    elseif ($ii >> 4 == 14) {
                        $b1 = ($ii & 15);
                        $ii = ord($data[$nn + 1]);
                        $b2 = ($ii & 63);
                        $ii = ord($data[$nn + 2]);
                        $b3 = ($ii & 63);
                        $ii = ((($b1 * 64) + $b2) * 64) + $b3;
                        $ent = sprintf('&#%d;', $ii);
                        $escapedData .= $ent;
                        $nn += 2;
                    } // 21 bits: 11110bbb 10bbbbbb 10bbbbbb 10bbbbbb
                    elseif ($ii >> 3 == 30) {
                        $b1 = ($ii & 7);
                        $ii = ord($data[$nn + 1]);
                        $b2 = ($ii & 63);
                        $ii = ord($data[$nn + 2]);
                        $b3 = ($ii & 63);
                        $ii = ord($data[$nn + 3]);
                        $b4 = ($ii & 63);
                        $ii = ((((($b1 * 64) + $b2) * 64) + $b3) * 64) + $b4;
                        $ent = sprintf('&#%d;', $ii);
                        $escapedData .= $ent;
                        $nn += 3;
                    }
                }

                // when converting to latin-1, do not be so eager with using entities for characters 160-255
                if ($conversion == 'UTF-8_ISO-8859-1') {
                    $escapedData = str_replace(array_slice($this->xml_iso88591_Entities['out'], 32), array_slice($this->xml_iso88591_Entities['in'], 32), $escapedData);
                }
                break;

            /*
            case 'CP1252_':
            case 'CP1252_US-ASCII':
                $escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&amp;', '&quot;', '&apos;', '&lt;', '&gt;'), $data);
                $escapedData = str_replace($this->xml_iso88591_Entities']['in'], $this->xml_iso88591_Entities['out'], $escapedData);
                $escapedData = str_replace($this->xml_cp1252_Entities['in'], $this->xml_cp1252_Entities['out'], $escapedData);
                break;
            case 'CP1252_UTF-8':
                $escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&amp;', '&quot;', '&apos;', '&lt;', '&gt;'), $data);
                /// @todo we could use real UTF8 chars here instead of xml entities... (note that utf_8 encode all allone will NOT convert them)
                $escapedData = str_replace($this->xml_cp1252_Entities['in'], $this->xml_cp1252_Entities['out'], $escapedData);
                $escapedData = utf8_encode($escapedData);
                break;
            case 'CP1252_ISO-8859-1':
                $escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&amp;', '&quot;', '&apos;', '&lt;', '&gt;'), $data);
                // we might as well replace all funky chars with a '?' here, but we are kind and leave it to the receiving application layer to decide what to do with these weird entities...
                $escapedData = str_replace($this->xml_cp1252_Entities['in'], $this->xml_cp1252_Entities['out'], $escapedData);
                break;
            */

            default:
                $escapedData = '';
                Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ": Converting from $srcEncoding to $destEncoding: not supported...");
        }

        return $escapedData;
    }

    /**
     * Checks if a given charset encoding is present in a list of encodings or if it is a valid subset of any encoding
     * in the list.
     *
     * @param string $encoding charset to be tested
     * @param string|array $validList comma separated list of valid charsets (or array of charsets)
     *
     * @return bool
     */
    public function isValidCharset($encoding, $validList)
    {
        if (is_string($validList)) {
            $validList = explode(',', $validList);
        }
        if (@in_array(strtoupper($encoding), $validList)) {
            return true;
        } else {
            if (array_key_exists($encoding, $this->charset_supersets)) {
                foreach ($validList as $allowed) {
                    if (in_array($allowed, $this->charset_supersets[$encoding])) {
                        return true;
                    }
                }
            }

            return false;
        }
    }

    /**
     * Used only for backwards compatibility
     * @deprecated
     *
     * @param string $charset
     *
     * @return array
     *
     * @throws \Exception for unknown/unsupported charsets
     */
    public function getEntities($charset)
    {
        //trigger_error('Method ' . __METHOD__ . ' is deprecated', E_USER_DEPRECATED);

        switch ($charset)
        {
            case 'iso88591':
                return $this->xml_iso88591_Entities;
            default:
                throw new \Exception('Unsupported charset: ' . $charset);
        }
    }

}


1		<?php
2
3		namespace PhpXmlRpc\Helper;
4
5		use PhpXmlRpc\PhpXmlRpc;
6
7		class Charset
8		{
9		// tables used for transcoding different charsets into us-ascii xml
10		protected $xml_iso88591_Entities = array("in" => array(), "out" => array());
11
12		/// @todo add to iso table the characters from cp_1252 range, i.e. 128 to 159?
13		/// These will NOT be present in true ISO-8859-1, but will save the unwary windows user from sending junk
14		/// (though no luck when receiving them...)
15		/*
16		protected $xml_cp1252_Entities = array('in' => array(), out' => array(
17		'€', '?', '‚', 'ƒ',
18		'„', '…', '†', '‡',
19		'ˆ', '‰', 'Š', '‹',
20		'Œ', '?', 'Ž', '?',
21		'?', '‘', '’', '“',
22		'”', '•', '–', '—',
23		'˜', '™', 'š', '›',
24		'œ', '?', 'ž', 'Ÿ'
25		));
26		*/
27
28		protected $charset_supersets = array(
29		'US-ASCII' => array('ISO-8859-1', 'ISO-8859-2', 'ISO-8859-3', 'ISO-8859-4',
30		'ISO-8859-5', 'ISO-8859-6', 'ISO-8859-7', 'ISO-8859-8',
31		'ISO-8859-9', 'ISO-8859-10', 'ISO-8859-11', 'ISO-8859-12',
32		'ISO-8859-13', 'ISO-8859-14', 'ISO-8859-15', 'UTF-8',
33		'EUC-JP', 'EUC-', 'EUC-KR', 'EUC-CN',),
34		);
35
36		protected static $instance = null;
37
38		/**
39		* This class is singleton for performance reasons.
40		*
41		* @return Charset
42		*/
43	596	public static function instance()
44		{
45	596	if (self::$instance === null) {
46	374	self::$instance = new self();
47		}
48
49	596	return self::$instance;
50		}
51
52	374	private function __construct()
53		{
54	374	for ($i = 0; $i < 32; $i++) {
55	374	$this->xml_iso88591_Entities["in"][] = chr($i);
56	374	$this->xml_iso88591_Entities["out"][] = "&#{$i};";
57		}
58
59	374	for ($i = 160; $i < 256; $i++) {
60	374	$this->xml_iso88591_Entities["in"][] = chr($i);
61	374	$this->xml_iso88591_Entities["out"][] = "&#{$i};";
62		}
63
64		/*for ($i = 128; $i < 160; $i++)
65		{
66		$this->xml_cp1252_Entities['in'][] = chr($i);
67		}*/
68	374	}
69
70		/**
71		* Convert a string to the correct XML representation in a target charset.
72		*
73		* To help correct communication of non-ascii chars inside strings, regardless of the charset used when sending
74		* requests, parsing them, sending responses and parsing responses, an option is to convert all non-ascii chars
75		* present in the message into their equivalent 'charset entity'. Charset entities enumerated this way are
76		* independent of the charset encoding used to transmit them, and all XML parsers are bound to understand them.
77		* Note that in the std case we are not sending a charset encoding mime type along with http headers, so we are
78		* bound by RFC 3023 to emit strict us-ascii.
79		*
80		* @todo do a bit of basic benchmarking (strtr vs. str_replace)
81		* @todo make usage of iconv() or recode_string() or mb_string() where available
82		*
83		* @param string $data
84		* @param string $srcEncoding
85		* @param string $destEncoding
86		*
87		* @return string
88		*/
89	577	public function encodeEntities($data, $srcEncoding = '', $destEncoding = '')
90		{
91	577	if ($srcEncoding == '') {
92		// lame, but we know no better...
93		$srcEncoding = PhpXmlRpc::$xmlrpc_internalencoding;
94		}
95
96	577	$conversion = strtoupper($srcEncoding . '_' . $destEncoding);
97	577	switch ($conversion) {
98	577	case 'ISO-8859-1_':
99	384	case 'ISO-8859-1_US-ASCII':
100	487	$escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&', '"', ''', '<', '>'), $data);
101	487	$escapedData = str_replace($this->xml_iso88591_Entities['in'], $this->xml_iso88591_Entities['out'], $escapedData);
102	487	break;
103
104	384	case 'ISO-8859-1_UTF-8':
105	28	$escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&', '"', ''', '<', '>'), $data);
106	28	$escapedData = utf8_encode($escapedData);
107	28	break;
108
109	373	case 'ISO-8859-1_ISO-8859-1':
110	362	case 'US-ASCII_US-ASCII':
111	362	case 'US-ASCII_UTF-8':
112	362	case 'US-ASCII_':
113	362	case 'US-ASCII_ISO-8859-1':
114	362	case 'UTF-8_UTF-8':
115		//case 'CP1252_CP1252':
116	46	$escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&', '"', ''', '<', '>'), $data);
117	46	break;
118
119	344	case 'UTF-8_':
120	24	case 'UTF-8_US-ASCII':
121	24	case 'UTF-8_ISO-8859-1':
122		// NB: this will choke on invalid UTF-8, going most likely beyond EOF
123	344	$escapedData = '';
124		// be kind to users creating string xmlrpc values out of different php types
125	344	$data = (string)$data;
126	344	$ns = strlen($data);
127	344	for ($nn = 0; $nn < $ns; $nn++) {
128	344	$ch = $data[$nn];
129	344	$ii = ord($ch);
130		// 7 bits: 0bbbbbbb (127)
131	344	if ($ii < 128) {
132		/// @todo shall we replace this with a (supposedly) faster str_replace?
133	337	switch ($ii) {
134	337	case 34:
135	20	$escapedData .= '"';
136	20	break;
137	337	case 38:
138	20	$escapedData .= '&';
139	20	break;
140	337	case 39:
141	22	$escapedData .= ''';
142	22	break;
143	337	case 60:
144	20	$escapedData .= '<';
145	20	break;
146	337	case 62:
147	20	$escapedData .= '>';
148	20	break;
149		default:
150	337	$escapedData .= $ch;
151		} // switch
152		} // 11 bits: 110bbbbb 10bbbbbb (2047)
153	70	elseif ($ii >> 5 == 6) {
154	66	$b1 = ($ii & 31);
155	66	$ii = ord($data[$nn + 1]);
156	66	$b2 = ($ii & 63);
157	66	$ii = ($b1 * 64) + $b2;
158	66	$ent = sprintf('&#%d;', $ii);
159	66	$escapedData .= $ent;
160	66	$nn += 1;
161		} // 16 bits: 1110bbbb 10bbbbbb 10bbbbbb
162	30	elseif ($ii >> 4 == 14) {
163	30	$b1 = ($ii & 15);
164	30	$ii = ord($data[$nn + 1]);
165	30	$b2 = ($ii & 63);
166	30	$ii = ord($data[$nn + 2]);
167	30	$b3 = ($ii & 63);
168	30	$ii = ((($b1 * 64) + $b2) * 64) + $b3;
169	30	$ent = sprintf('&#%d;', $ii);
170	30	$escapedData .= $ent;
171	30	$nn += 2;
172		} // 21 bits: 11110bbb 10bbbbbb 10bbbbbb 10bbbbbb
173		elseif ($ii >> 3 == 30) {
174		$b1 = ($ii & 7);
175		$ii = ord($data[$nn + 1]);
176		$b2 = ($ii & 63);
177		$ii = ord($data[$nn + 2]);
178		$b3 = ($ii & 63);
179		$ii = ord($data[$nn + 3]);
180		$b4 = ($ii & 63);
181		$ii = ((((($b1 * 64) + $b2) * 64) + $b3) * 64) + $b4;
182		$ent = sprintf('&#%d;', $ii);
183		$escapedData .= $ent;
184		$nn += 3;
185		}
186		}
187
188		// when converting to latin-1, do not be so eager with using entities for characters 160-255
189	344	if ($conversion == 'UTF-8_ISO-8859-1') {
190	24	$escapedData = str_replace(array_slice($this->xml_iso88591_Entities['out'], 32), array_slice($this->xml_iso88591_Entities['in'], 32), $escapedData);
191		}
192	344	break;
193
194		/*
195		case 'CP1252_':
196		case 'CP1252_US-ASCII':
197		$escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&', '"', ''', '<', '>'), $data);
198		$escapedData = str_replace($this->xml_iso88591_Entities']['in'], $this->xml_iso88591_Entities['out'], $escapedData);
199		$escapedData = str_replace($this->xml_cp1252_Entities['in'], $this->xml_cp1252_Entities['out'], $escapedData);
200		break;
201		case 'CP1252_UTF-8':
202		$escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&', '"', ''', '<', '>'), $data);
203		/// @todo we could use real UTF8 chars here instead of xml entities... (note that utf_8 encode all allone will NOT convert them)
204		$escapedData = str_replace($this->xml_cp1252_Entities['in'], $this->xml_cp1252_Entities['out'], $escapedData);
205		$escapedData = utf8_encode($escapedData);
206		break;
207		case 'CP1252_ISO-8859-1':
208		$escapedData = str_replace(array('&', '"', "'", '<', '>'), array('&', '"', ''', '<', '>'), $data);
209		// we might as well replace all funky chars with a '?' here, but we are kind and leave it to the receiving application layer to decide what to do with these weird entities...
210		$escapedData = str_replace($this->xml_cp1252_Entities['in'], $this->xml_cp1252_Entities['out'], $escapedData);
211		break;
212		*/
213
214		default:
215		$escapedData = '';
216		Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ": Converting from $srcEncoding to $destEncoding: not supported...");
217		}
218
219	577	return $escapedData;
220		}
221
222		/**
223		* Checks if a given charset encoding is present in a list of encodings or if it is a valid subset of any encoding
224		* in the list.
225		*
226		* @param string $encoding charset to be tested
227		* @param string\|array $validList comma separated list of valid charsets (or array of charsets)
228		*
229		* @return bool
230		*/
231		public function isValidCharset($encoding, $validList)
232		{
233		if (is_string($validList)) {
234		$validList = explode(',', $validList);
235		}
236		if (@in_array(strtoupper($encoding), $validList)) {
237		return true;
238		} else {
239		if (array_key_exists($encoding, $this->charset_supersets)) {
240		foreach ($validList as $allowed) {
241		if (in_array($allowed, $this->charset_supersets[$encoding])) {
242		return true;
243		}
244		}
245		}
246
247		return false;
248		}
249		}
250
251		/**
252		* Used only for backwards compatibility
253		* @deprecated
254		*
255		* @param string $charset
256		*
257		* @return array
258		*
259		* @throws \Exception for unknown/unsupported charsets
260		*/
261		public function getEntities($charset)
262		{
263		//trigger_error('Method ' . __METHOD__ . ' is deprecated', E_USER_DEPRECATED);
264
265		switch ($charset)
266		{
267		case 'iso88591':
268		return $this->xml_iso88591_Entities;
269		default:
270		throw new \Exception('Unsupported charset: ' . $charset);
271		}
272		}
273
274		}
275

gggeek / phpxmlrpc

Push — master ( 87b7a4...6ce28d )

Charset::instance() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like