Utf8::cp() - Code Metrics - Inspection of "Input\Utf8: simplified codepoint algorithm" - s9e/RegexpBuilder - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 50e3b5...fbe966 )

by Josh

created 2016-09-15 10:14 UTC

Utf8::cp() A

↳ Parent: Utf8

Complexity

Conditions	4
Paths	4

Size

Total Lines	18
Code Lines	9

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	9
CRAP Score	4

Importance

Changes

Metric	Value
dl	0
loc	18
ccs	9
cts	9
cp	1
rs	9.2
c	0
b	0
f	0
cc	4
eloc	9
nc	4
nop	1
crap	4

<?php

/**
* @package   s9e\RegexpBuilder
* @copyright Copyright (c) 2016 The s9e Authors
* @license   http://www.opensource.org/licenses/mit-license.php The MIT License
*/
namespace s9e\RegexpBuilder\Input;

use InvalidArgumentException;

class Utf8 implements InputInterface
{
	/**
	* {@inheritdoc}
	*/
	public function split($string)
	{
		if (preg_match_all('(.)us', $string, $matches) === false)
		{
			throw new InvalidArgumentException('Invalid UTF-8 string');
		}

		return $this->charsToCodepoints($matches[0]);
	}

	/**
	* Convert a list of UTF-8 characters to a list of Unicode codepoint
	*
	* @param  string[]  $chars
	* @return integer[]
	*/
	protected function charsToCodepoints(array $chars)
	{
		return array_map([$this, 'cp'], $chars);
	}

	/**
	* Compute and return the Unicode codepoint for given UTF-8 char
	*
	* @param  string  $char UTF-8 char
	* @return integer
	*/
	protected function cp($char)
	{
		$cp = ord($char[0]);
		if ($cp >= 0xF0)
		{
			$cp = ($cp << 18) + (ord($char[1]) << 12) + (ord($char[2]) << 6) + ord($char[3]) - 0x3C82080;
		}
		elseif ($cp >= 0xE0)
		{
			$cp = ($cp << 12) + (ord($char[1]) << 6) + ord($char[2]) - 0xE2080;
		}
		elseif ($cp >= 0xC0)
		{
			$cp = ($cp << 6) + ord($char[1]) - 0x3080;
		}

		return $cp;
	}
}

1		<?php
2
3		/**
4		* @package s9e\RegexpBuilder
5		* @copyright Copyright (c) 2016 The s9e Authors
6		* @license http://www.opensource.org/licenses/mit-license.php The MIT License
7		*/
8		namespace s9e\RegexpBuilder\Input;
9
10		use InvalidArgumentException;
11
12		class Utf8 implements InputInterface
13		{
14		/**
15		* {@inheritdoc}
16		*/
17	6	public function split($string)
18		{
19	6	if (preg_match_all('(.)us', $string, $matches) === false)
20		{
21	1	throw new InvalidArgumentException('Invalid UTF-8 string');
22		}
23
24	5	return $this->charsToCodepoints($matches[0]);
25		}
26
27		/**
28		* Convert a list of UTF-8 characters to a list of Unicode codepoint
29		*
30		* @param string[] $chars
31		* @return integer[]
32		*/
33	5	protected function charsToCodepoints(array $chars)
34		{
35	5	return array_map([$this, 'cp'], $chars);
36		}
37
38		/**
39		* Compute and return the Unicode codepoint for given UTF-8 char
40		*
41		* @param string $char UTF-8 char
42		* @return integer
43		*/
44	4	protected function cp($char)
45		{
46	4	$cp = ord($char[0]);
47	4	if ($cp >= 0xF0)
48		{
49	1	$cp = ($cp << 18) + (ord($char[1]) << 12) + (ord($char[2]) << 6) + ord($char[3]) - 0x3C82080;
50		}
51	4	elseif ($cp >= 0xE0)
52		{
53	1	$cp = ($cp << 12) + (ord($char[1]) << 6) + ord($char[2]) - 0xE2080;
54		}
55	3	elseif ($cp >= 0xC0)
56		{
57	1	$cp = ($cp << 6) + ord($char[1]) - 0x3080;
58		}
59
60	4	return $cp;
61		}
62		}

s9e / RegexpBuilder

Push — master ( 50e3b5...fbe966 )

Utf8::cp() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like