StringEncoder::iterateNamespaces() - Code Metrics - Inspection of "Added support for class resolution operator string..." - Riimu/Kit-PHPEncoder - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — master (#6)

by Riikka

created 2018-07-03 12:41 UTC

StringEncoder::iterateNamespaces() A

↳ Parent: StringEncoder

Complexity

Conditions	2
Paths	2

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	6
CRAP Score	2

Importance

Changes

Metric	Value
cc	2
nc	2
nop	1
dl	0
loc	9
ccs	6
cts	6
cp	1
crap	2
rs	9.9666
c	0
b	0
f	0

<?php

namespace Riimu\Kit\PHPEncoder\Encoder;

/**
 * Encoder for string values.
 * @author Riikka Kalliomäki <[email protected]>
 * @copyright Copyright (c) 2014-2018 Riikka Kalliomäki
 * @license http://opensource.org/licenses/mit-license.php MIT License
 */
class StringEncoder implements Encoder
{
    /** @var array Default values for options in the encoder */
    private static $defaultOptions = [
        'string.escape' => true,
        'string.binary' => false,
        'string.utf8' => false,
        'string.classes' => [],
        'string.imports' => [],
    ];

    public function getDefaultOptions()
    {
        return self::$defaultOptions;
    }

    public function supports($value)
    {
        return \is_string($value);
    }

    public function encode($value, $depth, array $options, callable $encode)
    {
        $value = (string) $value;

        if ($this->isClassName($value, $options)) {
            return $this->getClassName($value, $options);
        }

        if (preg_match('/[^\x20-\x7E]/', $value)) {
            return $this->getComplexString($value, $options);
        }

        return $this->getSingleQuotedString($value);
    }

    /**
     * Tests if the given value is a string that could be encoded as a class name constant.
     * @param string $value The string to test
     * @param array $options The string encoding options
     * @return bool True if string can be encoded as class constant, false if not
     */
    private function isClassName($value, array $options)
    {
        if (preg_match('/^([a-zA-Z_\x7f-\xff][a-zA-Z0-9_\x7f-\xff]*)(\\\\(?1))*$/', $value) !== 1) {
            return false;
        }

        return array_intersect(iterator_to_array($this->iterateNamespaces($value)), $options['string.classes']) !== [];
    }

    /**
     * Encodes the given string as a class name constant based on used imports.
     * @param string $value The string to encode
     * @param array $options The string encoding options
     * @return string The class constant PHP code representation
     */
    private function getClassName($value, array $options)
    {
        foreach ($this->iterateNamespaces($value) as $partial) {
            if (isset($options['string.imports'][$partial])) {
                $trimmed = substr($value, \strlen(rtrim($partial, '\\')));
                return ltrim(sprintf('%s%s::class', rtrim($options['string.imports'][$partial], '\\'), $trimmed), '\\');
            }
        }

        return sprintf('\\%s::class', $value);
    }

    /**
     * Iterates over the variations of the namespace for the given class name.
     * @param string $value The class name to iterate over
     * @return \Generator|string[] The namespace parts of the string
     */
    private function iterateNamespaces($value)
    {
        yield $value;

        $parts = explode('\\', '\\' . $value);
        $count = \count($parts);

        for ($i = 1; $i < $count; $i++) {
            yield ltrim(implode('\\', \array_slice($parts, 0, -$i)), '\\') . '\\';
        }
    }

    /**
     * Returns the PHP code representation for the string that is not just simple ascii characters.
     * @param string $value The string to encode
     * @param array $options The string encoding options
     * @return string The PHP code representation for the complex string
     */
    private function getComplexString($value, array $options)
    {
        if ($this->isBinaryString($value, $options)) {
            return $this->encodeBinaryString($value);
        }

        if ($options['string.escape']) {
            return $this->getDoubleQuotedString($value, $options);
        }

        return $this->getSingleQuotedString($value);
    }

    /**
     * Tells if the string is not a valid UTF-8 string.
     * @param string $string The string to test
     * @param array $options The string encoding options
     * @return bool True if the string is not valid UTF-8 and false if it is
     */
    private function isBinaryString($string, $options)
    {
        if (!$options['string.binary']) {
            return false;
        }

        // UTF-8 validity test without mbstring extension
        $pattern =
            '/^(?>
                [\x00-\x7F]+                       # ASCII
              | [\xC2-\xDF][\x80-\xBF]             # non-overlong 2-byte
              |  \xE0[\xA0-\xBF][\x80-\xBF]        # excluding over longs
              | [\xE1-\xEC\xEE\xEF][\x80-\xBF]{2}  # straight 3-byte
              |  \xED[\x80-\x9F][\x80-\xBF]        # excluding surrogates
              |  \xF0[\x90-\xBF][\x80-\xBF]{2}     # planes 1-3
              | [\xF1-\xF3][\x80-\xBF]{3}          # planes 4-15
              |  \xF4[\x80-\x8F][\x80-\xBF]{2}     # plane 16
            )*$/x';

        return !preg_match($pattern, $string);
    }

    /**
     * Encodes the given string into base 64 encoded format.
     * @param string $string The string to encode
     * @return string A base 64 PHP code representation for the string
     */
    private function encodeBinaryString($string)
    {
        return sprintf("base64_decode('%s')", base64_encode($string));
    }

    /**
     * Returns the string wrapped in single quotes and escape appropriately.
     * @param string $string String to wrap
     * @return string The string wrapped in single quotes
     */
    private function getSingleQuotedString($string)
    {
        return sprintf("'%s'", strtr($string, ["'" => "\\'", '\\' => '\\\\']));
    }

    /**
     * Returns the string wrapped in double quotes and all but print characters escaped.
     * @param string $string String to wrap and escape
     * @param array $options The string encoding options
     * @return string The string wrapped in double quotes and escape correctly
     */
    private function getDoubleQuotedString($string, $options)
    {
        $string = strtr($string, [
            "\n" => '\n',
            "\r" => '\r',
            "\t" => '\t',
            '$' => '\$',
            '"' => '\"',
            '\\' => '\\\\',
        ]);

        if ($options['string.utf8']) {
            $string = $this->encodeUtf8($string, $options);
        }

        $hexFormat = function ($matches) use ($options) {
            return sprintf($options['hex.capitalize'] ? '\x%02X' : '\x%02x', \ord($matches[0]));
        };

        return sprintf('"%s"', preg_replace_callback('/[^\x20-\x7E]/', $hexFormat, $string));
    }

    /**
     * Encodes all multibyte UTF-8 characters into PHP7 string encoding.
     * @param string $string The string to encoder
     * @param array $options The string encoding options
     * @return string The string with all the multibyte characters encoded
     */
    private function encodeUtf8($string, $options)
    {
        $pattern =
            '/  [\xC2-\xDF][\x80-\xBF]
              |  \xE0[\xA0-\xBF][\x80-\xBF]
              | [\xE1-\xEC\xEE\xEF][\x80-\xBF]{2}
              |  \xED[\x80-\x9F][\x80-\xBF]
              |  \xF0[\x90-\xBF][\x80-\xBF]{2}
              | [\xF1-\xF3][\x80-\xBF]{3}
              |  \xF4[\x80-\x8F][\x80-\xBF]{2}/x';

        return preg_replace_callback($pattern, function ($match) use ($options) {
            return sprintf($options['hex.capitalize'] ? '\u{%X}' : '\u{%x}', $this->getCodePoint($match[0]));
        }, $string);
    }

    /**
     * Returns the unicode code point for the given multibyte UTF-8 character.
     * @param string $bytes The multibyte character
     * @return int The code point for the multibyte character
     */
    private function getCodePoint($bytes)
    {
        if (\strlen($bytes) === 2) {
            return ((\ord($bytes[0]) & 0b11111) << 6)
                | (\ord($bytes[1]) & 0b111111);
        }

        if (\strlen($bytes) === 3) {
            return ((\ord($bytes[0]) & 0b1111) << 12)
                | ((\ord($bytes[1]) & 0b111111) << 6)
                | (\ord($bytes[2]) & 0b111111);
        }

        return ((\ord($bytes[0]) & 0b111) << 18)
            | ((\ord($bytes[1]) & 0b111111) << 12)
            | ((\ord($bytes[2]) & 0b111111) << 6)
            | (\ord($bytes[3]) & 0b111111);
    }
}


1		<?php
2
3		namespace Riimu\Kit\PHPEncoder\Encoder;
4
5		/**
6		* Encoder for string values.
7		* @author Riikka Kalliomäki <[email protected]>
8		* @copyright Copyright (c) 2014-2018 Riikka Kalliomäki
9		* @license http://opensource.org/licenses/mit-license.php MIT License
10		*/
11		class StringEncoder implements Encoder
12		{
13		/** @var array Default values for options in the encoder */
14		private static $defaultOptions = [
15		'string.escape' => true,
16		'string.binary' => false,
17		'string.utf8' => false,
18		'string.classes' => [],
19		'string.imports' => [],
20		];
21
22	228	public function getDefaultOptions()
23		{
24	228	return self::$defaultOptions;
25		}
26
27	140	public function supports($value)
28		{
29	140	return \is_string($value);
30		}
31
32	88	public function encode($value, $depth, array $options, callable $encode)
33		{
34	88	$value = (string) $value;
35
36	88	if ($this->isClassName($value, $options)) {
37	8	return $this->getClassName($value, $options);
38		}
39
40	84	if (preg_match('/[^\x20-\x7E]/', $value)) {
41	20	return $this->getComplexString($value, $options);
42		}
43
44	76	return $this->getSingleQuotedString($value);
45		}
46
47		/**
48		* Tests if the given value is a string that could be encoded as a class name constant.
49		* @param string $value The string to test
50		* @param array $options The string encoding options
51		* @return bool True if string can be encoded as class constant, false if not
52		*/
53	88	private function isClassName($value, array $options)
54		{
55	88	if (preg_match('/^([a-zA-Z_\x7f-\xff][a-zA-Z0-9_\x7f-\xff])(\\\\(?1))$/', $value) !== 1) {
56	40	return false;
57		}
58
59	72	return array_intersect(iterator_to_array($this->iterateNamespaces($value)), $options['string.classes']) !== [];
60		}
61
62		/**
63		* Encodes the given string as a class name constant based on used imports.
64		* @param string $value The string to encode
65		* @param array $options The string encoding options
66		* @return string The class constant PHP code representation
67		*/
68	8	private function getClassName($value, array $options)
69		{
70	8	foreach ($this->iterateNamespaces($value) as $partial) {
71	8	if (isset($options['string.imports'][$partial])) {
72	4	$trimmed = substr($value, \strlen(rtrim($partial, '\\')));
73	7	return ltrim(sprintf('%s%s::class', rtrim($options['string.imports'][$partial], '\\'), $trimmed), '\\');
74		}
75	2	}
76
77	4	return sprintf('\\%s::class', $value);
78		}
79
80		/**
81		* Iterates over the variations of the namespace for the given class name.
82		* @param string $value The class name to iterate over
83		* @return \Generator\|string[] The namespace parts of the string
84		*/
85	72	private function iterateNamespaces($value)
86		{
87	72	yield $value;
88
89	72	$parts = explode('\\', '\\' . $value);
90	72	$count = \count($parts);
91
92	72	for ($i = 1; $i < $count; $i++) {
93	72	yield ltrim(implode('\\', \array_slice($parts, 0, -$i)), '\\') . '\\';
94	18	}
95	72	}
96
97		/**
98		* Returns the PHP code representation for the string that is not just simple ascii characters.
99		* @param string $value The string to encode
100		* @param array $options The string encoding options
101		* @return string The PHP code representation for the complex string
102		*/
103	20	private function getComplexString($value, array $options)
104		{
105	20	if ($this->isBinaryString($value, $options)) {
106	4	return $this->encodeBinaryString($value);
107		}
108
109	20	if ($options['string.escape']) {
110	12	return $this->getDoubleQuotedString($value, $options);
111		}
112
113	12	return $this->getSingleQuotedString($value);
114		}
115
116		/**
117		* Tells if the string is not a valid UTF-8 string.
118		* @param string $string The string to test
119		* @param array $options The string encoding options
120		* @return bool True if the string is not valid UTF-8 and false if it is
121		*/
122	20	private function isBinaryString($string, $options)
123		{
124	20	if (!$options['string.binary']) {
125	16	return false;
126		}
127
128		// UTF-8 validity test without mbstring extension
129		$pattern =
130	3	'/^(?>
131		[\x00-\x7F]+ # ASCII
132		\| [\xC2-\xDF][\x80-\xBF] # non-overlong 2-byte
133		\| \xE0[\xA0-\xBF][\x80-\xBF] # excluding over longs
134		\| [\xE1-\xEC\xEE\xEF][\x80-\xBF]{2} # straight 3-byte
135		\| \xED[\x80-\x9F][\x80-\xBF] # excluding surrogates
136		\| \xF0[\x90-\xBF][\x80-\xBF]{2} # planes 1-3
137		\| [\xF1-\xF3][\x80-\xBF]{3} # planes 4-15
138		\| \xF4[\x80-\x8F][\x80-\xBF]{2} # plane 16
139	1	)*$/x';
140
141	4	return !preg_match($pattern, $string);
142		}
143
144		/**
145		* Encodes the given string into base 64 encoded format.
146		* @param string $string The string to encode
147		* @return string A base 64 PHP code representation for the string
148		*/
149	4	private function encodeBinaryString($string)
150		{
151	4	return sprintf("base64_decode('%s')", base64_encode($string));
152		}
153
154		/**
155		* Returns the string wrapped in single quotes and escape appropriately.
156		* @param string $string String to wrap
157		* @return string The string wrapped in single quotes
158		*/
159	80	private function getSingleQuotedString($string)
160		{
161	80	return sprintf("'%s'", strtr($string, ["'" => "\\'", '\\' => '\\\\']));
162		}
163
164		/**
165		* Returns the string wrapped in double quotes and all but print characters escaped.
166		* @param string $string String to wrap and escape
167		* @param array $options The string encoding options
168		* @return string The string wrapped in double quotes and escape correctly
169		*/
170	12	private function getDoubleQuotedString($string, $options)
171		{
172	12	$string = strtr($string, [
173	12	"\n" => '\n',
174	3	"\r" => '\r',
175	3	"\t" => '\t',
176	3	'$' => '\$',
177	3	'"' => '\"',
178	3	'\\' => '\\\\',
179	3	]);
180
181	12	if ($options['string.utf8']) {
182	4	$string = $this->encodeUtf8($string, $options);
183	1	}
184
185	3	$hexFormat = function ($matches) use ($options) {
186	12	return sprintf($options['hex.capitalize'] ? '\x%02X' : '\x%02x', \ord($matches[0]));
187	12	};
188
189	12	return sprintf('"%s"', preg_replace_callback('/[^\x20-\x7E]/', $hexFormat, $string));
190		}
191
192		/**
193		* Encodes all multibyte UTF-8 characters into PHP7 string encoding.
194		* @param string $string The string to encoder
195		* @param array $options The string encoding options
196		* @return string The string with all the multibyte characters encoded
197		*/
198	4	private function encodeUtf8($string, $options)
199		{
200		$pattern =
201	3	'/ [\xC2-\xDF][\x80-\xBF]
202		\| \xE0[\xA0-\xBF][\x80-\xBF]
203		\| [\xE1-\xEC\xEE\xEF][\x80-\xBF]{2}
204		\| \xED[\x80-\x9F][\x80-\xBF]
205		\| \xF0[\x90-\xBF][\x80-\xBF]{2}
206		\| [\xF1-\xF3][\x80-\xBF]{3}
207	1	\| \xF4[\x80-\x8F][\x80-\xBF]{2}/x';
208
209	2	return preg_replace_callback($pattern, function ($match) use ($options) {
210	4	return sprintf($options['hex.capitalize'] ? '\u{%X}' : '\u{%x}', $this->getCodePoint($match[0]));
211	4	}, $string);
212		}
213
214		/**
215		* Returns the unicode code point for the given multibyte UTF-8 character.
216		* @param string $bytes The multibyte character
217		* @return int The code point for the multibyte character
218		*/
219	4	private function getCodePoint($bytes)
220		{
221	4	if (\strlen($bytes) === 2) {
222	4	return ((\ord($bytes[0]) & 0b11111) << 6)
223	4	\| (\ord($bytes[1]) & 0b111111);
224		}
225
226	4	if (\strlen($bytes) === 3) {
227	4	return ((\ord($bytes[0]) & 0b1111) << 12)
228	4	\| ((\ord($bytes[1]) & 0b111111) << 6)
229	4	\| (\ord($bytes[2]) & 0b111111);
230		}
231
232	4	return ((\ord($bytes[0]) & 0b111) << 18)
233	4	\| ((\ord($bytes[1]) & 0b111111) << 12)
234	4	\| ((\ord($bytes[2]) & 0b111111) << 6)
235	4	\| (\ord($bytes[3]) & 0b111111);
236		}
237		}
238

Riimu / Kit-PHPEncoder

Pull Request — master (#6)

StringEncoder::iterateNamespaces() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like