Completed
Push — master ( 695952...ad19ae )
by Nazar
04:24
created

Assets_processing::file_put_contents_with_hash()   A

Complexity

Conditions 1
Paths 1

Size

Total Lines 5
Code Lines 4

Duplication

Lines 0
Ratio 0 %

Code Coverage

Tests 4
CRAP Score 1

Importance

Changes 0
Metric Value
cc 1
eloc 4
nc 1
nop 3
dl 0
loc 5
ccs 4
cts 4
cp 1
crap 1
rs 9.4285
c 0
b 0
f 0
1
<?php
2
/**
3
 * @package   CleverStyle Framework
4
 * @author    Nazar Mokrynskyi <[email protected]>
5
 * @copyright Copyright (c) 2014-2016, Nazar Mokrynskyi
6
 * @license   MIT License, see license.txt
7
 */
8
namespace cs\Page;
9
10
/**
11
 * Class includes few methods used for processing CSS, JS and HTML files before putting into cache
12
 *
13
 * This is because CSS and HTML files may include other CSS, JS files, images, fonts and so on with absolute and relative paths.
14
 * Methods of this class handle all this assets, applies basic minification to CSS and JS files and produce single resulting file (relative paths to
15
 * files that can't be embedded are converted to absolute). This allows to decrease number of HTTP requests on page and avoid breaking of relative paths for
16
 * fonts, images and other assets after putting them into cache directory as well as minimize contents size by removing comments and other redundant stuff.
17
 */
18
class Assets_processing {
19
	/**
20
	 * Do not inline files bigger than 4 KiB
21
	 */
22
	const MAX_EMBEDDING_SIZE = 4096;
23
	protected static $extension_to_mime = [
24
		'jpeg' => 'image/jpg',
25
		'jpe'  => 'image/jpg',
26
		'jpg'  => 'image/jpg',
27
		'gif'  => 'image/gif',
28
		'png'  => 'image/png',
29
		'svg'  => 'image/svg+xml',
30
		'svgz' => 'image/svg+xml',
31
		'woff' => 'application/font-woff',
32
		//'woff2' => 'application/font-woff2'
33
	];
34
	/**
35
	 * Analyses file for images, fonts and css links and include they content into single resulting css file.
36
	 *
37
	 * Supports next file extensions for possible assets:
38
	 * jpeg, jpe, jpg, gif, png, ttf, ttc, svg, svgz, woff, css
39
	 *
40
	 * @param string   $data                   Content of processed file
41
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
42
	 * @param string   $target_directory_path  Target directory for resulting combined files
43
	 * @param string[] $not_embedded_resources Some resources like images and fonts might not be embedded into resulting CSS because of their size
44
	 *
45
	 * @return string    $data
46
	 */
47 10
	public static function css ($data, $file, $target_directory_path = PUBLIC_CACHE, &$not_embedded_resources = []) {
48 10
		$dir = dirname($file);
49
		/**
50
		 * Remove comments, tabs and new lines
51
		 */
52 10
		$data = preg_replace('#(/\*.*?\*/)|\t|\n|\r#s', ' ', $data);
53
		/**
54
		 * Remove unnecessary spaces
55
		 */
56 10
		$data = preg_replace('/\s*([,;>{}(])\s*/', '$1', $data);
57 10
		$data = preg_replace('/\s+/', ' ', $data);
58
		/**
59
		 * Return spaces required in media queries
60
		 */
61 10
		$data = preg_replace('/\s(and|or)\(/', ' $1 (', $data);
62
		/**
63
		 * Duplicated semicolons
64
		 */
65 10
		$data = preg_replace('/;+/m', ';', $data);
66
		/**
67
		 * Minify rgb colors declarations
68
		 */
69 10
		$data = preg_replace_callback(
70 10
			'/rgb\(([0-9,.]+)\)/i',
71
			function ($rgb) {
72 6
				$rgb = explode(',', $rgb[1]);
73
				return
74
					'#'.
75 6
					str_pad(dechex($rgb[0]), 2, 0, STR_PAD_LEFT).
76 6
					str_pad(dechex($rgb[1]), 2, 0, STR_PAD_LEFT).
77 6
					str_pad(dechex($rgb[2]), 2, 0, STR_PAD_LEFT);
78 10
			},
79
			$data
80
		);
81
		/**
82
		 * Minify repeated colors declarations
83
		 */
84 10
		$data = preg_replace('/#([0-9a-f])\1([0-9a-f])\2([0-9a-f])\3/i', '#$1$2$3', $data);
85
		/**
86
		 * Remove unnecessary zeros
87
		 */
88 10
		$data = preg_replace('/(\D)0\.(\d+)/i', '$1.$2', $data);
89
		/**
90
		 * Unnecessary spaces around colons (should have whitespace character after, otherwise might be `.c :disabled` and will be handled incorrectly)
91
		 */
92 10
		$data = preg_replace('/\s*:\s+/', ':', $data);
93
		/**
94
		 * Assets processing
95
		 */
96 10
		$data = preg_replace_callback(
97 10
			'/url\((.*)\)|@import\s*(?:url\()?\s*([\'"].*[\'"])\s*\)??(.*);/U',
98
			function ($match) use ($dir, $target_directory_path, &$not_embedded_resources) {
99 10
				$path_matched = @$match[2] ?: $match[1];
100 10
				$path         = trim($path_matched, '\'" ');
101 10
				$link         = explode('?', $path, 2)[0];
102 10
				if (!static::is_relative_path_and_exists($link, $dir)) {
103 4
					return $match[0];
104
				}
105 10
				$extension     = file_extension($link);
106 10
				$absolute_path = static::absolute_path($link, $dir);
107
				/**
108
				 * Only process CSS imports without media queries, imports with media queries will just be corrected to absolute paths
109
				 */
110 10
				if ($extension == 'css' && @$match[2] && !trim(@$match[3])) {
111
					/**
112
					 * For recursive stylesheets processing, if CSS file includes others CSS files
113
					 */
114 6
					return static::css(file_get_contents($absolute_path), $absolute_path, $target_directory_path, $not_embedded_resources);
115
				}
116 10
				$content = file_get_contents($absolute_path);
117 10
				if (!isset(static::$extension_to_mime[$extension]) || filesize($absolute_path) > static::MAX_EMBEDDING_SIZE) {
118 10
					$filename = static::file_put_contents_with_hash($target_directory_path, $extension, $content);
119 10
					if (isset(static::$extension_to_mime[$extension]) && strpos($path, '?') === false) {
120 6
						$not_embedded_resources[] = str_replace(getcwd(), '', "$target_directory_path/$filename");
121
					}
122 10
					return str_replace($path_matched, "'./$filename'", $match[0]);
123
				}
124 6
				$mime_type = static::$extension_to_mime[$extension];
125 6
				$content   = base64_encode($content);
126 6
				return str_replace($path, "data:$mime_type;charset=utf-8;base64,$content", $match[0]);
127 10
			},
128
			$data
129
		);
130 10
		return trim($data);
131
	}
132
	/**
133
	 * Put `$content` into `$dir` where filename is `md5($content)` with specified extension
134
	 *
135
	 * @param string $dir
136
	 * @param string $extension
137
	 * @param string $content
138
	 *
139
	 * @return string Filename (without full path)
140
	 */
141 10
	protected static function file_put_contents_with_hash ($dir, $extension, $content) {
142 10
		$hash = md5($content);
143 10
		file_put_contents("$dir/$hash.$extension", $content, LOCK_EX | FILE_BINARY);
144 10
		return "$hash.$extension";
145
	}
146
	/**
147
	 * Simple and fast JS minification
148
	 *
149
	 * @param string $data
150
	 *
151
	 * @return string
152
	 */
153 10
	public static function js ($data) {
154
		/**
155
		 * Split into array of lines
156
		 */
157 10
		$data = explode("\n", $data);
158
		/**
159
		 * Flag that is `true` when inside comment
160
		 */
161 10
		$in_comment              = false;
162 10
		$continue_after_position = -1;
163 10
		foreach ($data as $index => &$current_line) {
164 10
			if ($continue_after_position >= $index) {
165 6
				continue;
166
			}
167 10
			$next_line = isset($data[$index + 1]) ? trim($data[$index + 1]) : '';
168
			/**
169
			 * Remove starting and trailing spaces
170
			 */
171 10
			$current_line = trim($current_line);
172
			/**
173
			 * Remove single-line comments
174
			 */
175 10
			if (mb_strpos($current_line, '//') === 0) {
176 10
				$current_line = '';
177 10
				continue;
178
			}
179
			/**
180
			 * Starts with multi-line comment
181
			 */
182 10
			if (mb_strpos($current_line, '/*') === 0) {
183 10
				$in_comment = true;
184
			}
185 10
			if (!$in_comment) {
186 10
				$backticks_position = strpos($current_line, '`');
187
				/**
188
				 * Handling template strings can be tricky (since they might be multi-line), so let's fast-forward to the last backticks position and continue
189
				 * from there
190
				 */
191 10
				if ($backticks_position !== false) {
192 6
					$last_item_with_backticks = array_keys(
193
						array_filter(
194
							$data,
195
							function ($d) {
196 6
								return strpos($d, '`') !== false;
197 6
							}
198
						)
199
					);
200 6
					$last_item_with_backticks = array_pop($last_item_with_backticks);
201 6
					if ($last_item_with_backticks > $index) {
202 6
						$continue_after_position = $last_item_with_backticks;
203 6
						continue;
204
					}
205
				}
206
				/**
207
				 * Add new line at the end if only needed
208
				 */
209 10
				if (static::new_line_needed($current_line, $next_line)) {
210 10
					$current_line .= "\n";
211
				}
212
				/**
213
				 * Single-line comment
214
				 */
215 10
				$current_line = preg_replace('#^\s*//[^\'"]+$#', '', $current_line);
216
				/**
217
				 * If we are not sure - just add new line afterwards
218
				 */
219 10
				$current_line = preg_replace('#//.*$#', "\\0\n", $current_line);
220
			} else {
221
				/**
222
				 * End of multi-line comment
223
				 */
224 10
				if (strpos($current_line, '*/') !== false) {
225 10
					$current_line = explode('*/', $current_line)[1];
226 10
					$in_comment   = false;
227
				} else {
228 10
					$current_line = '';
229
				}
230
			}
231
		}
232 10
		$data = implode('', $data);
233 10
		$data = str_replace('</script>', '<\/script>', $data);
234 10
		return trim($data, ';').';';
235
	}
236
	/**
237
	 * @param string $current_line
238
	 * @param string $next_line
239
	 *
240
	 * @return bool
241
	 */
242 10
	protected static function new_line_needed ($current_line, $next_line) {
243
		/**
244
		 * Set of symbols that are safe to be concatenated without new line with anything else
245
		 */
246
		$regexp = /** @lang PhpRegExp */
247 10
			'[:;,.+\-*/{}?><^\'"\[\]=&(]';
248
		return
249 10
			$current_line &&
250 10
			$next_line &&
251 10
			!preg_match("#$regexp\$#", $current_line) &&
252 10
			!preg_match("#^$regexp#", $next_line);
253
	}
254
	/**
255
	 * Analyses file for scripts and styles, combines them into resulting files in order to optimize loading process
256
	 * (files with combined scripts and styles will be created)
257
	 *
258
	 * @param string   $data                   Content of processed file
259
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
260
	 * @param string   $target_directory_path  Target directory for resulting combined files
261
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
262
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
263
	 *
264
	 * @return string
265
	 */
266 8
	public static function html ($data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources = []) {
267 8
		static::html_process_links_and_styles($data, $file, $target_directory_path, $vulcanization, $not_embedded_resources);
268 8
		static::html_process_scripts($data, $file, $target_directory_path, $vulcanization, $not_embedded_resources);
269
		// Removing HTML comments (those that are mostly likely comments, to avoid problems)
270 8
		$data = preg_replace_callback(
271 8
			'/^\s*<!--([^>-].*[^-])?-->/Ums',
272 8
			function ($matches) {
273 8
				return mb_strpos('--', $matches[1]) === false ? '' : $matches[0];
274 8
			},
275
			$data
276
		);
277 8
		return preg_replace("/\n+/", "\n", $data);
278
	}
279
	/**
280
	 * @param string   $data                   Content of processed file
281
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
282
	 * @param string   $target_directory_path  Target directory for resulting combined files
283
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
284
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
285
	 */
286 8
	protected static function html_process_scripts (&$data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources) {
287 8
		if (!preg_match_all('/<script(.*)<\/script>/Uims', $data, $scripts)) {
288 8
			return;
289
		}
290 8
		$scripts_content    = '';
291 8
		$scripts_to_replace = [];
292 8
		$dir                = dirname($file);
293 8
		foreach ($scripts[1] as $index => $script) {
294 8
			$script = explode('>', $script, 2);
295 8
			if (preg_match('/src\s*=\s*[\'"](.*)[\'"]/Uims', $script[0], $url)) {
296 8
				$url = $url[1];
297 8
				if (!static::is_relative_path_and_exists($url, $dir)) {
298 4
					continue;
299
				}
300 8
				$scripts_to_replace[] = $scripts[0][$index];
301 8
				$scripts_content .= file_get_contents("$dir/$url").";\n";
302
			} else {
303 8
				$scripts_to_replace[] = $scripts[0][$index];
304 8
				$scripts_content .= "$script[1];\n";
305
			}
306
		}
307 8
		$scripts_content = static::js($scripts_content);
308 8
		if (!$scripts_to_replace) {
309 4
			return;
310
		}
311
		// Remove all scripts
312 8
		$data = str_replace($scripts_to_replace, '', $data);
313
		/**
314
		 * If vulcanization is not used - put contents into separate file, and put link to it, otherwise put minified content back
315
		 */
316 8
		if (!$vulcanization) {
317
			// TODO: Remove in 7.x; For backward compatibility, since some modules might use this by specifying file path
318 2
			if (!is_dir($target_directory_path)) {
319
				$target_directory_path = dirname($target_directory_path);
320
			}
321 2
			$filename = static::file_put_contents_with_hash($target_directory_path, 'js', $scripts_content);
322
			// Add script with combined content file to the end
323 2
			$data .= "<script src=\"./$filename\"></script>";
324 2
			$not_embedded_resources[] = str_replace(getcwd(), '', "$target_directory_path/$filename");
325
		} else {
326
			// Add combined content inline script to the end
327 6
			$data .= "<script>$scripts_content</script>";
328
		}
329 8
	}
330
	/**
331
	 * @param string   $data                   Content of processed file
332
	 * @param string   $file                   Path to file, that contains specified in previous parameter content
333
	 * @param string   $target_directory_path  Target directory for resulting combined files
334
	 * @param bool     $vulcanization          Whether to put combined files separately or to make included assets built-in (vulcanization)
335
	 * @param string[] $not_embedded_resources Resources like images/fonts might not be embedded into resulting CSS because of big size or CSS/JS because of CSP
336
	 */
337 8
	protected static function html_process_links_and_styles (&$data, $file, $target_directory_path, $vulcanization, &$not_embedded_resources) {
338
		// Drop Polymer inclusion, since it is already present
339 8
		$data = str_replace('<link rel="import" href="../polymer/polymer.html">', '', $data);
340 8
		if (!preg_match_all('/<link(.*)>|<style(.*)<\/style>/Uims', $data, $links_and_styles)) {
341 8
			return;
342
		}
343 8
		$dir = dirname($file);
344 8
		foreach ($links_and_styles[1] as $index => $link) {
345
			/**
346
			 * Check for custom styles `is="custom-style"` or styles includes `include=".."` - we'll skip them
347
			 * Or if content is plain CSS
348
			 */
349
			if (
350 8
				preg_match('/^[^>]*(is="custom-style"|include=)[^>]*>/Uim', $links_and_styles[2][$index]) ||
351 8
				mb_strpos($links_and_styles[0][$index], '</style>') > 0
352
			) {
353 8
				$content = explode('>', $links_and_styles[2][$index], 2)[1];
354 8
				$data    = str_replace(
355
					$content,
356 8
					static::css($content, $file, $target_directory_path, $not_embedded_resources),
357
					$data
358
				);
359 8
				continue;
360
			}
361 8
			if (!static::has_relative_href($link, $url, $dir)) {
362 4
				continue;
363
			}
364 8
			$import = preg_match('/rel\s*=\s*[\'"]import[\'"]/Uim', $link);
365
			/**
366
			 * CSS imports are available in Polymer alongside with HTML imports
367
			 */
368 8
			$css_import = $import && preg_match('/type\s*=\s*[\'"]css[\'"]/Uim', $link);
369 8
			$stylesheet = preg_match('/rel\s*=\s*[\'"]stylesheet[\'"]/Uim', $link);
370
			// TODO: Polymer only supports `style[is=custom-style]`, but no `link`-based counterpart, so we can't provide CSP-compatibility for CSS anyway
371 8
			if ($css_import || $stylesheet) {
372
				/**
373
				 * If content is link to CSS file
374
				 */
375 8
				$css  = static::css(
376 8
					file_get_contents("$dir/$url"),
377 8
					"$dir/$url",
378
					$target_directory_path,
379
					$not_embedded_resources
380
				);
381 8
				$data = preg_replace(
382 8
					'/'.$links_and_styles[0][$index].'.*<template>/Uims',
383 8
					"<template><style>$css</style>",
384
					$data
385
				);
386 4
			} elseif ($import) {
387
				/**
388
				 * If content is HTML import
389
				 */
390 4
				$data = str_replace(
391 4
					$links_and_styles[0][$index],
392 4
					static::html(
393 4
						file_get_contents("$dir/$url"),
394 8
						"$dir/$url",
395
						$target_directory_path,
396
						$vulcanization,
397
						$not_embedded_resources
398
					),
399
					$data
400
				);
401
			}
402
		}
403 8
	}
404
	/**
405
	 * @param string $link
406
	 * @param string $url
407
	 * @param string $dir
408
	 *
409
	 * @return bool
410
	 */
411 8
	protected static function has_relative_href ($link, &$url, $dir) {
412
		$result =
413 8
			$link &&
414 8
			preg_match('/href\s*=\s*[\'"](.*)[\'"]/Uims', $link, $url);
415 8
		if ($result && static::is_relative_path_and_exists($url[1], $dir)) {
416 8
			$url = $url[1];
417 8
			return true;
418
		}
419 4
		return false;
420
	}
421
	/**
422
	 * @param string $path
423
	 * @param string $dir
424
	 *
425
	 * @return bool
426
	 */
427 10
	protected static function is_relative_path_and_exists ($path, $dir) {
428 10
		return $dir && !preg_match('#^https?://#i', $path) && file_exists(static::absolute_path($path, $dir));
429
	}
430
	/**
431
	 * @param string $path
432
	 * @param string $dir
433
	 *
434
	 * @return string
435
	 */
436 10
	protected static function absolute_path ($path, $dir) {
437 10
		if (strpos($path, '/') === 0) {
438 6
			return realpath(getcwd().$path);
439
		}
440 10
		return realpath("$dir/$path");
441
	}
442
}
443