Index: trunk/phase3/includes/libs/JavaScriptDistiller.php |
— | — | @@ -19,12 +19,20 @@ |
20 | 20 | * @param $stripVerticalSpace Boolean: Try to remove as much vertical whitespace as possible |
21 | 21 | */ |
22 | 22 | public static function stripWhiteSpace( $script, $stripVerticalSpace = false ) { |
| 23 | + // Try to avoid segfaulting |
| 24 | + // I saw segfaults with a limit of 10000, 1000 seems to work |
| 25 | + $oldLimit = ini_get( 'pcre.recursion_limit' ); |
| 26 | + if ( intval( $oldLimit ) > 1000 ) { |
| 27 | + ini_set( 'pcre.recursion_limit', '1000' ); |
| 28 | + } |
| 29 | + |
23 | 30 | $script = self::stripHorizontalSpace( $script ); |
24 | 31 | // If requested, make some vertical whitespace collapsing as well |
25 | 32 | if ( $stripVerticalSpace ) { |
26 | 33 | $script = self::stripVerticalSpace( $script ); |
27 | 34 | } |
28 | 35 | // Done |
| 36 | + ini_set( 'pcre.recursion_limit', $oldLimit ); |
29 | 37 | return $script; |
30 | 38 | } |
31 | 39 | |
— | — | @@ -74,55 +82,44 @@ |
75 | 83 | // to \s if we use a backslash as the escape character. We work around this by using an |
76 | 84 | // obscure escape character that we hope will never appear at the end of a line. |
77 | 85 | $parser->escapeChar = chr( 1 ); |
| 86 | + |
| 87 | + // C-style comment: use non-greedy repetition to find the end |
| 88 | + $parser->add( '\/ \* .*? \* \/' ); |
| 89 | + |
| 90 | + // Preserve the newline after a C++-style comment -- bug 27046 |
| 91 | + $parser->add( '\/ \/ [^\r\n]* ( [\r\n] )', '$2' ); |
| 92 | + |
78 | 93 | // Protect strings. The original code had [^\'\\v] here, but that didn't armor multiline |
79 | 94 | // strings correctly. This also armors multiline strings that don't have backslashes at the |
80 | 95 | // end of the line (these are invalid), but that's fine because we're just armoring here. |
81 | 96 | |
82 | 97 | // Single quotes |
83 | | - $parser->add( |
84 | | - '\' (' . // start quote |
| 98 | + $parser->add( |
| 99 | + '\'' . // start quote |
85 | 100 | '[^\'\\\\]*' . // a run of non-special characters |
86 | | - '(' . |
87 | | - '\\\\ ( . | [\r\n] )' . // a backslash followed by a character or line ending |
| 101 | + '(?:' . |
| 102 | + '\\\\ .' . // a backslash followed by a character or line ending |
88 | 103 | '[^\'\\\\]*' . // a run of non-special characters |
89 | 104 | ')*' . // any number of the above |
90 | | - ') \'', // end quote |
| 105 | + '\'', // end quote |
91 | 106 | '$1' ); |
92 | 107 | |
93 | 108 | // Double quotes: same as above |
94 | | - $parser->add( '" ( [^"\\\\]* ( \\\\ ( . | [\r\n] ) [^"\\\\]* )* ) "', '$1' ); |
| 109 | + $parser->add( '" [^"\\\\]* (?: \\\\ . [^"\\\\]* )* "', '$1' ); |
95 | 110 | |
96 | 111 | // Protect regular expressions |
97 | 112 | // Regular expression with whitespace before it |
98 | 113 | $parser->add( |
99 | | - '[ \t]+ ( ( \/' . // whitespace then start slash |
| 114 | + '(?<= [ \t] | [^\w\$\/\'"*)\?:] )' . // assert that whitespace or punctuation precedes |
| 115 | + '\/' . // start slash |
100 | 116 | '[^\r\n\*]' . // not a comment-start or line ending |
101 | 117 | '[^\/\r\n\\\\]*' . // a sequence of non-special characters |
102 | | - '(' . |
| 118 | + '(?:' . |
103 | 119 | '\\\\.' . // an escaped dot |
104 | 120 | '[^\/\r\n\\\\]*' . // a sequence of non-special characters |
105 | 121 | ')*' . // any number of the above |
106 | | - '\/(i|g)*' . // pattern end, optional modifier |
107 | | - ') )', |
| 122 | + '\/[ig]*' , // pattern end, optional modifier |
108 | 123 | '$1' ); |
109 | | - // Regular expression with an operator before it |
110 | | - $parser->add( |
111 | | - '( [^\w\$\/\'"*)\?:] (\/' . // certain kinds of punctuation and then start slash |
112 | | - '[^\r\n\*]' . // not a comment-start or line ending |
113 | | - '[^\/\r\n\\\\]*' . // a sequence of non-special characters |
114 | | - '(' . |
115 | | - '\\\\.' . // an escaped dot |
116 | | - '[^\/\r\n\\\\]*' . // a sequence of non-special characters |
117 | | - ')*' . // any number of the above |
118 | | - '\/(i|g)*)' . // pattern end, optional modifier |
119 | | - ')', |
120 | | - '$1' ); |
121 | | - |
122 | | - // C-style comment: use non-greedy repetition to find the end |
123 | | - $parser->add( '\/ \* ( . | [\r\n] )*? \* \/' ); |
124 | | - |
125 | | - // Preserve the newline after a C++-style comment -- bug 27046 |
126 | | - $parser->add( '\/ \/ [^\r\n]* ( [\r\n] )', '$2' ); |
127 | 124 | return $parser; |
128 | 125 | } |
129 | 126 | } |
— | — | @@ -149,10 +146,9 @@ |
150 | 147 | const LENGTH = 2; |
151 | 148 | |
152 | 149 | // used to determine nesting levels |
153 | | - private $GROUPS = '/\(/';//g |
| 150 | + private $GROUPS = '/\( (?! \? ) /x';//g |
154 | 151 | private $SUB_REPLACE = '/\$\d/'; |
155 | 152 | private $INDEXED = '/^\$\d+$/'; |
156 | | - //private $TRIM = '/([\'"])\1\.(.*)\.\1\1$/'; |
157 | 153 | private $ESCAPE = '/\\\./';//g |
158 | 154 | private $QUOTE = '/\'/'; |
159 | 155 | private $DELETED = '/\x01[^\x01]*\x01/';//g |
— | — | @@ -197,9 +193,9 @@ |
198 | 194 | // simulate the _patterns.toSTring of Dean |
199 | 195 | $regexp = '/'; |
200 | 196 | foreach ($this->_patterns as $reg) { |
201 | | - $regexp .= '(' . $reg[self::EXPRESSION] . ')|'; |
| 197 | + $regexp .= '(' . $reg[self::EXPRESSION] . ")|\n"; |
202 | 198 | } |
203 | | - $regexp = substr($regexp, 0, -1) . '/Sx'; |
| 199 | + $regexp = substr($regexp, 0, -2) . '/Sxs'; |
204 | 200 | $regexp .= ($this->ignoreCase) ? 'i' : ''; |
205 | 201 | |
206 | 202 | $string = $this->_escape($string, $this->escapeChar); |