1 | <?php |
||
17 | abstract class Minify |
||
18 | { |
||
19 | /** |
||
20 | * The data to be minified. |
||
21 | * |
||
22 | * @var string[] |
||
23 | */ |
||
24 | protected $data = array(); |
||
25 | |||
26 | /** |
||
27 | * Array of patterns to match. |
||
28 | * |
||
29 | * @var string[] |
||
30 | */ |
||
31 | protected $patterns = array(); |
||
32 | |||
33 | /** |
||
34 | * This array will hold content of strings and regular expressions that have |
||
35 | * been extracted from the JS source code, so we can reliably match "code", |
||
36 | * without having to worry about potential "code-like" characters inside. |
||
37 | * |
||
38 | * @var string[] |
||
39 | */ |
||
40 | public $extracted = array(); |
||
41 | |||
42 | /** |
||
43 | * Init the minify class - optionally, code may be passed along already. |
||
44 | */ |
||
45 | public function __construct(/* $data = null, ... */) |
||
46 | { |
||
47 | // it's possible to add the source through the constructor as well ;) |
||
48 | if (func_num_args()) { |
||
49 | call_user_func_array(array($this, 'add'), func_get_args()); |
||
50 | } |
||
51 | } |
||
52 | |||
53 | /** |
||
54 | * Add a file or straight-up code to be minified. |
||
55 | * |
||
56 | * @param string $data |
||
57 | */ |
||
58 | public function add($data /* $data = null, ... */) |
||
59 | { |
||
60 | // bogus "usage" of parameter $data: scrutinizer warns this variable is |
||
61 | // not used (we're using func_get_args instead to support overloading), |
||
62 | // but it still needs to be defined because it makes no sense to have |
||
63 | // this function without argument :) |
||
64 | $args = array($data) + func_get_args(); |
||
65 | |||
66 | // this method can be overloaded |
||
67 | foreach ($args as $data) { |
||
68 | // redefine var |
||
69 | $data = (string) $data; |
||
70 | |||
71 | // load data |
||
72 | $value = $this->load($data); |
||
73 | $key = ($data != $value) ? $data : count($this->data); |
||
74 | |||
75 | // store data |
||
76 | $this->data[$key] = $value; |
||
77 | } |
||
78 | } |
||
79 | |||
80 | /** |
||
81 | * Minify the data & (optionally) saves it to a file. |
||
82 | * |
||
83 | * @param string[optional] $path Path to write the data to. |
||
84 | * |
||
85 | * @return string The minified data. |
||
86 | */ |
||
87 | public function minify($path = null) |
||
88 | { |
||
89 | $content = $this->execute($path); |
||
90 | |||
91 | // save to path |
||
92 | if ($path !== null) { |
||
93 | $this->save($content, $path); |
||
94 | } |
||
95 | |||
96 | return $content; |
||
97 | } |
||
98 | |||
99 | /** |
||
100 | * Minify & gzip the data & (optionally) saves it to a file. |
||
101 | * |
||
102 | * @param string[optional] $path Path to write the data to. |
||
103 | * @param int[optional] $level Compression level, from 0 to 9. |
||
104 | * |
||
105 | * @return string The minified & gzipped data. |
||
106 | */ |
||
107 | public function gzip($path = null, $level = 9) |
||
108 | { |
||
109 | $content = $this->execute($path); |
||
110 | $content = gzencode($content, $level, FORCE_GZIP); |
||
111 | |||
112 | // save to path |
||
113 | if ($path !== null) { |
||
114 | $this->save($content, $path); |
||
115 | } |
||
116 | |||
117 | return $content; |
||
118 | } |
||
119 | |||
120 | /** |
||
121 | * Minify the data & write it to a CacheItemInterface object. |
||
122 | * |
||
123 | * @param CacheItemInterface $item Cache item to write the data to. |
||
124 | * |
||
125 | * @return CacheItemInterface Cache item with the minifier data. |
||
126 | */ |
||
127 | public function cache(CacheItemInterface $item) |
||
128 | { |
||
129 | $content = $this->execute(); |
||
130 | $item->set($content); |
||
131 | |||
132 | return $item; |
||
133 | } |
||
134 | |||
135 | /** |
||
136 | * Minify the data. |
||
137 | * |
||
138 | * @param string[optional] $path Path to write the data to. |
||
139 | * |
||
140 | * @return string The minified data. |
||
141 | */ |
||
142 | abstract public function execute($path = null); |
||
143 | |||
144 | /** |
||
145 | * Load data. |
||
146 | * |
||
147 | * @param string $data Either a path to a file or the content itself. |
||
148 | * |
||
149 | * @return string |
||
150 | */ |
||
151 | protected function load($data) |
||
152 | { |
||
153 | // check if the data is a file |
||
154 | if ($this->canImportFile($data)) { |
||
155 | $data = file_get_contents($data); |
||
156 | |||
157 | // strip BOM, if any |
||
158 | if (substr($data, 0, 3) == "\xef\xbb\xbf") { |
||
159 | $data = substr($data, 3); |
||
160 | } |
||
161 | } |
||
162 | |||
163 | return $data; |
||
164 | } |
||
165 | |||
166 | /** |
||
167 | * Save to file. |
||
168 | * |
||
169 | * @param string $content The minified data. |
||
170 | * @param string $path The path to save the minified data to. |
||
171 | * |
||
172 | * @throws IOException |
||
173 | */ |
||
174 | protected function save($content, $path) |
||
175 | { |
||
176 | $handler = $this->openFileForWriting($path); |
||
177 | |||
178 | $this->writeToFile($handler, $content); |
||
179 | |||
180 | @fclose($handler); |
||
|
|||
181 | } |
||
182 | |||
183 | /** |
||
184 | * Register a pattern to execute against the source content. |
||
185 | * |
||
186 | * @param string $pattern PCRE pattern. |
||
187 | * @param string|callable $replacement Replacement value for matched pattern. |
||
188 | */ |
||
189 | protected function registerPattern($pattern, $replacement = '') |
||
190 | { |
||
191 | // study the pattern, we'll execute it more than once |
||
192 | $pattern .= 'S'; |
||
193 | |||
194 | $this->patterns[] = array($pattern, $replacement); |
||
195 | } |
||
196 | |||
197 | /** |
||
198 | * We can't "just" run some regular expressions against JavaScript: it's a |
||
199 | * complex language. E.g. having an occurrence of // xyz would be a comment, |
||
200 | * unless it's used within a string. Of you could have something that looks |
||
201 | * like a 'string', but inside a comment. |
||
202 | * The only way to accurately replace these pieces is to traverse the JS one |
||
203 | * character at a time and try to find whatever starts first. |
||
204 | * |
||
205 | * @param string $content The content to replace patterns in. |
||
206 | * |
||
207 | * @return string The (manipulated) content. |
||
208 | */ |
||
209 | protected function replace($content) |
||
210 | { |
||
211 | $processed = ''; |
||
212 | $positions = array_fill(0, count($this->patterns), -1); |
||
213 | $matches = array(); |
||
214 | |||
215 | while ($content) { |
||
216 | // find first match for all patterns |
||
217 | foreach ($this->patterns as $i => $pattern) { |
||
218 | list($pattern, $replacement) = $pattern; |
||
219 | |||
220 | // no need to re-run matches that are still in the part of the |
||
221 | // content that hasn't been processed |
||
222 | if ($positions[$i] >= 0) { |
||
223 | continue; |
||
224 | } |
||
225 | |||
226 | $match = null; |
||
227 | if (preg_match($pattern, $content, $match)) { |
||
228 | $matches[$i] = $match; |
||
229 | |||
230 | // we'll store the match position as well; that way, we |
||
231 | // don't have to redo all preg_matches after changing only |
||
232 | // the first (we'll still know where those others are) |
||
233 | $positions[$i] = strpos($content, $match[0]); |
||
234 | } else { |
||
235 | // if the pattern couldn't be matched, there's no point in |
||
236 | // executing it again in later runs on this same content; |
||
237 | // ignore this one until we reach end of content |
||
238 | unset($matches[$i]); |
||
239 | $positions[$i] = strlen($content); |
||
240 | } |
||
241 | } |
||
242 | |||
243 | // no more matches to find: everything's been processed, break out |
||
244 | if (!$matches) { |
||
245 | $processed .= $content; |
||
246 | break; |
||
247 | } |
||
248 | |||
249 | // see which of the patterns actually found the first thing (we'll |
||
250 | // only want to execute that one, since we're unsure if what the |
||
251 | // other found was not inside what the first found) |
||
252 | $discardLength = min($positions); |
||
253 | $firstPattern = array_search($discardLength, $positions); |
||
254 | $match = $matches[$firstPattern][0]; |
||
255 | |||
256 | // execute the pattern that matches earliest in the content string |
||
257 | list($pattern, $replacement) = $this->patterns[$firstPattern]; |
||
258 | $replacement = $this->replacePattern($pattern, $replacement, $content); |
||
259 | |||
260 | // figure out which part of the string was unmatched; that's the |
||
261 | // part we'll execute the patterns on again next |
||
262 | $content = substr($content, $discardLength); |
||
263 | $unmatched = (string) substr($content, strpos($content, $match) + strlen($match)); |
||
264 | |||
265 | // move the replaced part to $processed and prepare $content to |
||
266 | // again match batch of patterns against |
||
267 | $processed .= substr($replacement, 0, strlen($replacement) - strlen($unmatched)); |
||
268 | $content = $unmatched; |
||
269 | |||
270 | // first match has been replaced & that content is to be left alone, |
||
271 | // the next matches will start after this replacement, so we should |
||
272 | // fix their offsets |
||
273 | foreach ($positions as $i => $position) { |
||
274 | $positions[$i] -= $discardLength + strlen($match); |
||
275 | } |
||
276 | } |
||
277 | |||
278 | return $processed; |
||
279 | } |
||
280 | |||
281 | /** |
||
282 | * This is where a pattern is matched against $content and the matches |
||
283 | * are replaced by their respective value. |
||
284 | * This function will be called plenty of times, where $content will always |
||
285 | * move up 1 character. |
||
286 | * |
||
287 | * @param string $pattern Pattern to match. |
||
288 | * @param string|callable $replacement Replacement value. |
||
289 | * @param string $content Content to match pattern against. |
||
290 | * |
||
291 | * @return string |
||
292 | */ |
||
293 | protected function replacePattern($pattern, $replacement, $content) |
||
294 | { |
||
295 | if (is_callable($replacement)) { |
||
296 | return preg_replace_callback($pattern, $replacement, $content, 1, $count); |
||
297 | } else { |
||
298 | return preg_replace($pattern, $replacement, $content, 1, $count); |
||
299 | } |
||
300 | } |
||
301 | |||
302 | /** |
||
303 | * Strings are a pattern we need to match, in order to ignore potential |
||
304 | * code-like content inside them, but we just want all of the string |
||
305 | * content to remain untouched. |
||
306 | * |
||
307 | * This method will replace all string content with simple STRING# |
||
308 | * placeholder text, so we've rid all strings from characters that may be |
||
309 | * misinterpreted. Original string content will be saved in $this->extracted |
||
310 | * and after doing all other minifying, we can restore the original content |
||
311 | * via restoreStrings(). |
||
312 | * |
||
313 | * @param string[optional] $chars |
||
314 | */ |
||
315 | protected function extractStrings($chars = '\'"') |
||
316 | { |
||
317 | // PHP only supports $this inside anonymous functions since 5.4 |
||
318 | $minifier = $this; |
||
319 | $callback = function ($match) use ($minifier) { |
||
320 | // check the second index here, because the first always contains a quote |
||
321 | if (!$match[2]) { |
||
322 | /* |
||
323 | * Empty strings need no placeholder; they can't be confused for |
||
324 | * anything else anyway. |
||
325 | * But we still needed to match them, for the extraction routine |
||
326 | * to skip over this particular string. |
||
327 | */ |
||
328 | return $match[0]; |
||
329 | } |
||
330 | |||
331 | $count = count($minifier->extracted); |
||
332 | $placeholder = $match[1].$count.$match[1]; |
||
333 | $minifier->extracted[$placeholder] = $match[1].$match[2].$match[1]; |
||
334 | |||
335 | return $placeholder; |
||
336 | }; |
||
337 | |||
338 | /* |
||
339 | * The \\ messiness explained: |
||
340 | * * Don't count ' or " as end-of-string if it's escaped (has backslash |
||
341 | * in front of it) |
||
342 | * * Unless... that backslash itself is escaped (another leading slash), |
||
343 | * in which case it's no longer escaping the ' or " |
||
344 | * * So there can be either no backslash, or an even number |
||
345 | * * multiply all of that times 4, to account for the escaping that has |
||
346 | * to be done to pass the backslash into the PHP string without it being |
||
347 | * considered as escape-char (times 2) and to get it in the regex, |
||
348 | * escaped (times 2) |
||
349 | */ |
||
350 | $this->registerPattern('/(['.$chars.'])(.*?(?<!\\\\)(\\\\\\\\)*+)\\1/s', $callback); |
||
351 | } |
||
352 | |||
353 | /** |
||
354 | * This method will restore all extracted data (strings, regexes) that were |
||
355 | * replaced with placeholder text in extract*(). The original content was |
||
356 | * saved in $this->extracted. |
||
357 | * |
||
358 | * @param string $content |
||
359 | * |
||
360 | * @return string |
||
361 | */ |
||
362 | protected function restoreExtractedData($content) |
||
375 | |||
376 | /** |
||
377 | * Check if the path is a regular file and can be read. |
||
378 | * |
||
379 | * @param string $path |
||
380 | * |
||
381 | * @return bool |
||
382 | */ |
||
383 | protected function canImportFile($path) |
||
384 | { |
||
385 | return strlen($path) < PHP_MAXPATHLEN && is_file($path) && is_readable($path); |
||
386 | } |
||
387 | |||
388 | /** |
||
389 | * Attempts to open file specified by $path for writing. |
||
390 | * |
||
391 | * @param string $path The path to the file. |
||
392 | * |
||
393 | * @return resource Specifier for the target file. |
||
394 | * |
||
395 | * @throws IOException |
||
396 | */ |
||
397 | protected function openFileForWriting($path) |
||
405 | |||
406 | /** |
||
407 | * Attempts to write $content to the file specified by $handler. $path is used for printing exceptions. |
||
408 | * |
||
409 | * @param resource $handler The resource to write to. |
||
410 | * @param string $content The content to write. |
||
411 | * @param string $path The path to the file (for exception printing only). |
||
412 | * |
||
413 | * @throws IOException |
||
414 | */ |
||
415 | protected function writeToFile($handler, $content, $path = '') |
||
421 | } |
||
422 |
If you suppress an error, we recommend checking for the error condition explicitly: