raisePhpSettingLimits(); } } /** * Minify a string of CSS * @param string $css * @param int|bool $linebreak_pos * @return string */ public function run($css, $linebreak_pos = FALSE) { $this->comments = array(); $this->preserved_tokens = array(); $start_index = 0; $length = strlen($css); $css = $this->extract_data_urls($css); // collect all comment blocks... while (($start_index = $this->index_of($css, '/*', $start_index)) >= 0) { $end_index = $this->index_of($css, '*/', $start_index + 2); if ($end_index < 0) { $end_index = $length; } $this->comments[] = $this->str_slice($css, $start_index + 2, $end_index); $css = $this->str_slice($css, 0, $start_index + 2) . '___YUICSSMIN_PRESERVE_CANDIDATE_COMMENT_' . (count($this->comments) - 1) . '___' . $this->str_slice($css, $end_index); $start_index += 2; } // preserve strings so their content doesn't get accidentally minified $css = preg_replace_callback('/(?:"(?:[^\\\\"]|\\\\.|\\\\)*")|'."(?:'(?:[^\\\\']|\\\\.|\\\\)*')/", array($this, 'callback_one'), $css); // Let's divide css code in chunks of 25.000 chars aprox. // Reason: PHP's PCRE functions like preg_replace have a "backtrack limit" of 100.000 chars by default (php < 5.3.7) // so if we're dealing with really long strings and a (sub)pattern matches a number of chars greater than // the backtrack limit number (i.e. /(.*)/s) PCRE functions may fail silently returning NULL and // $css would be empty. $charset = ''; $charset_regexp = '/@charset [^;]+;/i'; $css_chunks = array(); $css_chunk_length = 25000; // aprox size, not exact $start_index = 0; $i = $css_chunk_length; // save initial iterations $l = strlen($css); // if the number of characters is 25000 or less, do not chunk if ($l <= $css_chunk_length) { $css_chunks[] = $css; } else { // chunk css code securely while ($i < $l) { $i += 50; // save iterations. 500 checks for a closing curly brace } if ($l - $start_index <= $css_chunk_length || $i >= $l) { $css_chunks[] = $this->str_slice($css, $start_index); break; } if ($css[$i - 1] === '}' && $i - $start_index > $css_chunk_length) { // If there are two ending curly braces }} separated or not by spaces, // join them in the same chunk (i.e. @media blocks) $next_chunk = substr($css, $i); if (preg_match('/^\s*\}/', $next_chunk)) { $i = $i + $this->index_of($next_chunk, '}') + 1; } $css_chunks[] = $this->str_slice($css, $start_index, $i); $start_index = $i; } } } // Minify each chunk for ($i = 0, $n = count($css_chunks); $i < $n; $i++) { $css_chunks[$i] = $this->minify($css_chunks[$i], $linebreak_pos); // If there is a @charset in a css chunk... if (empty($charset) && preg_match($charset_regexp, $css_chunks[$i], $matches)) { // delete all of them no matter the chunk $css_chunks[$i] = preg_replace($charset_regexp, '', $css_chunks[$i]); $charset = $matches[0]; } } // Update the first chunk and put the charset to the top of the file. $css_chunks[0] = $charset . $css_chunks[0]; return implode('', $css_chunks); } /** * Get the minimum PHP setting values suggested for CSSmin * @return array */ public function getSuggestedPhpLimits() { return array( 'memory_limit' => '128M', 'pcre.backtrack_limit' => 1000 * 1000, 'pcre.recursion_limit' => 500 * 1000, ); } /** * Configure PHP to use at least the suggested minimum settings * * @todo Move this functionality to separate class. */ public function raisePhpSettingLimits() { foreach ($this->getSuggestedPhpLimits() as $key => $val) { $current = $this->normalizeInt(ini_get($key)); $suggested = $this->normalizeInt($val); if ($current < $suggested) { ini_set($key, $val); } } } /** * Does bulk of the minification * @param string $css * @param int|bool $linebreak_pos * @return string */ private function minify($css, $linebreak_pos) { // strings are safe, now wrestle the comments for ($i = 0, $max = count($this->comments); $i < $max; $i++) { $token = $this->comments[$i]; $placeholder = '/___YUICSSMIN_PRESERVE_CANDIDATE_COMMENT_' . $i . '___/'; // ! in the first position of the comment means preserve // so push to the preserved tokens keeping the ! if (substr($token, 0, 1) === '!') { $this->preserved_tokens[] = $token; $css = preg_replace($placeholder, '___YUICSSMIN_PRESERVED_TOKEN_' . (count($this->preserved_tokens) - 1) . '___', $css, 1); continue; } // \ in the last position looks like hack for Mac/IE5 // shorten that to /*\*/ and the next one to /**/ if (substr($token, (strlen($token) - 1), 1) === '\\') { $this->preserved_tokens[] = '\\'; $css = preg_replace($placeholder, '___YUICSSMIN_PRESERVED_TOKEN_' . (count($this->preserved_tokens) - 1) . '___', $css, 1); $i = $i + 1; // attn: advancing the loop $this->preserved_tokens[] = ''; $css = preg_replace('/___YUICSSMIN_PRESERVE_CANDIDATE_COMMENT_' . $i . '___/', '___YUICSSMIN_PRESERVED_TOKEN_' . (count($this->preserved_tokens) - 1) . '___', $css, 1); continue; } // keep empty comments after child selectors (IE7 hack) // e.g. html >/**/ body if (strlen($token) === 0) { $start_index = $this->index_of($css, $this->str_slice($placeholder, 1, -1)); if ($start_index > 2) { if (substr($css, $start_index - 3, 1) === '>') { $this->preserved_tokens[] = ''; $css = preg_replace($placeholder, '___YUICSSMIN_PRESERVED_TOKEN_' . (count($this->preserved_tokens) - 1) . '___', $css, 1); } } } // in all other cases kill the comment $css = preg_replace('/\/\*' . $this->str_slice($placeholder, 1, -1) . '\*\//', '', $css, 1); } // Normalize all whitespace strings to single spaces. Easier to work with that way. $css = preg_replace('/\s+/', ' ', $css); // Remove the spaces before the things that should not have spaces before them. // But, be careful not to turn "p :link {...}" into "p:link{...}" // Swap out any pseudo-class colons with the token, and then swap back. $css = preg_replace_callback('/(?:^|\})(?:(?:[^\{\:])+\:)+(?:[^\{]*\{)/', array($this, 'callback_two'), $css); $css = preg_replace('/\s+([\!\{\}\;\:\>\+\(\)\],])/', '$1', $css); $css = preg_replace('/___YUICSSMIN_PSEUDOCLASSCOLON___/', ':', $css); // retain space for special IE6 cases $css = preg_replace('/\:first\-(line|letter)(\{|,)/', ':first-$1 $2', $css); // no space after the end of a preserved comment $css = preg_replace('/\*\/ /', '*/', $css); // Put the space back in some cases, to support stuff like // @media screen and (-webkit-min-device-pixel-ratio:0){ $css = preg_replace('/\band\(/i', 'and (', $css); // Remove the spaces after the things that should not have spaces after them. $css = preg_replace('/([\!\{\}\:;\>\+\(\[,])\s+/', '$1', $css); // remove unnecessary semicolons $css = preg_replace('/;+\}/', '}', $css); // Replace 0(px,em,%) with 0. $css = preg_replace('/([\s\:])(0)(?:px|em|%|in|cm|mm|pc|pt|ex)/i', '$1$2', $css); // Replace 0 0 0 0; with 0. $css = preg_replace('/\:0 0 0 0(;|\})/', ':0$1', $css); $css = preg_replace('/\:0 0 0(;|\})/', ':0$1', $css); $css = preg_replace('/\:0 0(;|\})/', ':0$1', $css); // Replace background-position:0; with background-position:0 0; // same for transform-origin $css = preg_replace_callback('/(background\-position|transform\-origin|webkit\-transform\-origin|moz\-transform\-origin|o-transform\-origin|ms\-transform\-origin)\:0(;|\})/i', array($this, 'callback_three'), $css); // Replace 0.6 to .6, but only when preceded by : or a white-space $css = preg_replace('/(\:|\s)0+\.(\d+)/', '$1.$2', $css); // Shorten colors from rgb(51,102,153) to #336699 // This makes it more likely that it'll get further compressed in the next step. $css = preg_replace_callback('/rgb\s*\(\s*([0-9,\s]+)\s*\)/i', array($this, 'callback_four'), $css); // Shorten colors from #AABBCC to #ABC. $css = $this->compress_hex_colors($css); // border: none -> border:0 $css = preg_replace_callback('/(border|border\-top|border\-right|border\-bottom|border\-right|outline|background)\:none(;|\})/i', array($this, 'callback_five'), $css); // shorter opacity IE filter $css = preg_replace('/progid\:DXImageTransform\.Microsoft\.Alpha\(Opacity\=/i', 'alpha(opacity=', $css); // Remove empty rules. $css = preg_replace('/[^\};\{\/]+\{\}/', '', $css); // Some source control tools don't like it when files containing lines longer // than, say 8000 characters, are checked in. The linebreak option is used in // that case to split long lines after a specific column. if ($linebreak_pos !== FALSE && (int) $linebreak_pos >= 0) { $linebreak_pos = (int) $linebreak_pos; $start_index = $i = 0; while ($i < strlen($css)) { $i++; if ($css[$i - 1] === '}' && $i - $start_index > $linebreak_pos) { $css = $this->str_slice($css, 0, $i) . "\n" . $this->str_slice($css, $i); $start_index = $i; } } } // Replace multiple semi-colons in a row by a single one // See SF bug #1980989 $css = preg_replace('/;;+/', ';', $css); // restore preserved comments and strings for ($i = 0, $max = count($this->preserved_tokens); $i < $max; $i++) { $css = preg_replace('/___YUICSSMIN_PRESERVED_TOKEN_' . $i . '___/', $this->preserved_tokens[$i], $css, 1); } // Trim the final string (for any leading or trailing white spaces) $css = preg_replace('/^\s+|\s+$/', '', $css); return $css; } /** * Utility method to replace all data urls with tokens before we start * compressing, to avoid performance issues running some of the subsequent * regexes against large strings chunks. * * @param string $css * @return string */ private function extract_data_urls($css) { // Leave data urls alone to increase parse performance. $max_index = strlen($css) - 1; $append_index = $index = $last_index = $offset = 0; $sb = array(); $pattern = '/url\(\s*(["\']?)data\:/'; // Since we need to account for non-base64 data urls, we need to handle // ' and ) being part of the data string. Hence switching to indexOf, // to determine whether or not we have matching string terminators and // handling sb appends directly, instead of using matcher.append* methods. while (preg_match($pattern, $css, $m, 0, $offset)) { $index = $this->index_of($css, $m[0], $offset); $last_index = $index + strlen($m[0]); $start_index = $index + 4; // "url(".length() $end_index = $last_index - 1; $terminator = $m[1]; // ', " or empty (not quoted) $found_terminator = FALSE; if (strlen($terminator) === 0) { $terminator = ')'; } while ($found_terminator === FALSE && $end_index+1 <= $max_index) { $end_index = $this->index_of($css, $terminator, $end_index + 1); // endIndex == 0 doesn't really apply here if ($end_index > 0 && substr($css, $end_index - 1, 1) !== '\\') { $found_terminator = TRUE; if (')' != $terminator) { $end_index = $this->index_of($css, ')', $end_index); } } } // Enough searching, start moving stuff over to the buffer $sb[] = $this->substring($css, $append_index, $index); if ($found_terminator) { $token = $this->substring($css, $start_index, $end_index); $token = preg_replace('/\s+/', '', $token); $this->preserved_tokens[] = $token; $preserver = 'url(___YUICSSMIN_PRESERVED_TOKEN_' . (count($this->preserved_tokens) - 1) . '___)'; $sb[] = $preserver; $append_index = $end_index + 1; } else { // No end terminator found, re-add the whole match. Should we throw/warn here? $sb[] = $this->substring($css, $index, $last_index); $append_index = $last_index; } $offset = $last_index; } $sb[] = $this->substring($css, $append_index); return implode('', $sb); } /** * Utility method to compress hex color values of the form #AABBCC to #ABC. * * DOES NOT compress CSS ID selectors which match the above pattern (which would break things). * e.g. #AddressForm { ... } * * DOES NOT compress IE filters, which have hex color values (which would break things). * e.g. filter: chroma(color="#FFFFFF"); * * DOES NOT compress invalid hex values. * e.g. background-color: #aabbccdd * * @param string $css * @return string */ private function compress_hex_colors($css) { // Look for hex colors inside { ... } (to avoid IDs) and which don't have a =, or a " in front of them (to avoid filters) $pattern = '/(\=\s*?["\']?)?#([0-9a-f])([0-9a-f])([0-9a-f])([0-9a-f])([0-9a-f])([0-9a-f])(\}|[^0-9a-f{][^{]*?\})/i'; $_index = $index = $last_index = $offset = 0; $sb = array(); while (preg_match($pattern, $css, $m, 0, $offset)) { $index = $this->index_of($css, $m[0], $offset); $last_index = $index + strlen($m[0]); $is_filter = (bool) $m[1]; $sb[] = $this->substring($css, $_index, $index); if ($is_filter) { // Restore, maintain case, otherwise filter will break $sb[] = $m[1] . '#' . $m[2] . $m[3] . $m[4] . $m[5] . $m[6] . $m[7]; } else { if (strtolower($m[2]) == strtolower($m[3]) && strtolower($m[4]) == strtolower($m[5]) && strtolower($m[6]) == strtolower($m[7])) { // Compress. $sb[] = '#' . strtolower($m[3] . $m[5] . $m[7]); } else { // Non compressible color, restore but lower case. $sb[] = '#' . strtolower($m[2] . $m[3] . $m[4] . $m[5] . $m[6] . $m[7]); } } $_index = $offset = $last_index - strlen($m[8]); } $sb[] = $this->substring($css, $_index); return implode('', $sb); } /* CALLBACKS * --------------------------------------------------------------------------------------------- */ private function callback_one($matches) { $match = $matches[0]; $quote = substr($match, 0, 1); // Must use addcslashes in PHP to avoid parsing of backslashes $match = addcslashes($this->str_slice($match, 1, -1), '\\'); // maybe the string contains a comment-like substring? // one, maybe more? put'em back then if (($pos = $this->index_of($match, '___YUICSSMIN_PRESERVE_CANDIDATE_COMMENT_')) >= 0) { for ($i = 0, $max = count($this->comments); $i < $max; $i++) { $match = preg_replace('/___YUICSSMIN_PRESERVE_CANDIDATE_COMMENT_' . $i . '___/', $this->comments[$i], $match, 1); } } // minify alpha opacity in filter strings $match = preg_replace('/progid\:DXImageTransform\.Microsoft\.Alpha\(Opacity\=/i', 'alpha(opacity=', $match); $this->preserved_tokens[] = $match; return $quote . '___YUICSSMIN_PRESERVED_TOKEN_' . (count($this->preserved_tokens) - 1) . '___' . $quote; } private function callback_two($matches) { return preg_replace('/\:/', '___YUICSSMIN_PSEUDOCLASSCOLON___', $matches[0]); } private function callback_three($matches) { return strtolower($matches[1]) . ':0 0' . $matches[2]; } private function callback_four($matches) { $rgbcolors = explode(',', $matches[1]); for ($i = 0; $i < count($rgbcolors); $i++) { $rgbcolors[$i] = base_convert(strval(intval($rgbcolors[$i], 10)), 10, 16); if (strlen($rgbcolors[$i]) === 1) { $rgbcolors[$i] = '0' . $rgbcolors[$i]; } } return '#' . implode('', $rgbcolors); } private function callback_five($matches) { return strtolower($matches[1]) . ':0' . $matches[2]; } /* HELPERS * --------------------------------------------------------------------------------------------- */ /** * PHP port of Javascript's "indexOf" function for strings only * Author: Tubal Martin http://blog.margenn.com * * @param string $haystack * @param string $needle * @param int $offset index (optional) * @return int */ private function index_of($haystack, $needle, $offset = 0) { $index = strpos($haystack, $needle, $offset); return ($index !== FALSE) ? $index : -1; } /** * PHP port of Javascript's "substring" function * Author: Tubal Martin http://blog.margenn.com * Tests: http://margenn.com/tubal/substring/ * * @param string $str * @param int $from index * @param int|bool $to index (optional) * @return string */ private function substring($str, $from = 0, $to = FALSE) { if ($to !== FALSE) { if ($from == $to || ($from <= 0 && $to < 0)) { return ''; } if ($from > $to) { $from_copy = $from; $from = $to; $to = $from_copy; } } if ($from < 0) { $from = 0; } $substring = ($to === FALSE) ? substr($str, $from) : substr($str, $from, $to - $from); return ($substring === FALSE) ? '' : $substring; } /** * PHP port of Javascript's "slice" function for strings only * Author: Tubal Martin http://blog.margenn.com * Tests: http://margenn.com/tubal/str_slice/ * * @param string $str * @param int $start index * @param int|bool $end index (optional) * @return string */ private function str_slice($str, $start = 0, $end = FALSE) { if ($end !== FALSE && ($start < 0 || $end <= 0)) { $max = strlen($str); if ($start < 0) { if (($start = $max + $start) < 0) { return ''; } } if ($end < 0) { if (($end = $max + $end) < 0) { return ''; } } if ($end <= $start) { return ''; } } $slice = ($end === FALSE) ? substr($str, $start) : substr($str, $start, $end - $start); return ($slice === FALSE) ? '' : $slice; } /** * Convert strings like "64M" to int values * @param mixed $size * @return int */ private function normalizeInt($size) { if (is_string($size)) { switch (substr($size, -1)) { case 'M': case 'm': return (int)$size * 1048576; case 'K': case 'k': return (int)$size * 1024; case 'G': case 'g': return (int)$size * 1073741824; default: return (int) $size; } } return (int) $size; } }