Improve regex extraction

Fixes issue #115
matthiasmullie · Jun 9, 2016 · dcdedb0 · dcdedb0
1 parent 2552412
commit dcdedb0
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 13 deletions.
diff --git a/src/JS.php b/src/JS.php
@@ -205,25 +205,27 @@ protected function extractRegex()
         $callback = function ($match) use ($minifier) {
             $count = count($minifier->extracted);
             $placeholder = '/'.$count.'/';
-            $minifier->extracted[$placeholder] = $match[1];
+            $minifier->extracted[$placeholder] = $match[0];
 
             return $placeholder;
         };
 
-        // it's a regex if we can find an opening and (not escaped) closing /,
-        // include \n because it may be there for a reason
+        $pattern = '\/.*?(?<!\\\\)(\\\\\\\\)*+\/[gimy]*';
+
+        // a regular expression can only be followed by a few operators or some
+        // of the RegExp methods (a `\` followed by a variable or value is
+        // likely part of a division, not a regex)
+        $after = '[,;\)]';
+        $methods = '\.(exec|test|match|search|replace|split)\(';
+        $this->registerPattern('/'.$pattern.'\s*(?=('.$after.'|'.$methods.'))/', $callback);
+
+        // 1 more edgecase: a regex can be followed by a lot more operators or
+        // keywords if there's a newline (ASI) in between, where the operator
+        // actually starts a new statement
         // (https://github.com/matthiasmullie/minify/issues/56)
-        $pattern = '(\/.*?(?<!\\\\)(\\\\\\\\)*+\/\n?)';
-
-        // / can't be preceded by variable, value, or similar because then
-        // it's going to be division
-        // checking for that is complex, so we'll do inverse:
-        // * at the beginning of the file, it's not division, but regex
-        $this->registerPattern('/^\s*\K'.$pattern.'/', $callback);
-        // * following another operator, it's not division, but regex
         $operators = $this->getOperatorsForRegex($this->operatorsBefore, '/');
-        $operators += $this->getKeywordsForRegex($this->keywordsReserved, '/');
-        $this->registerPattern('/(?:'.implode('|', $operators).')\s*\K'.$pattern.'/', $callback);
+        $operators += $this->getOperatorsForRegex($this->keywordsReserved, '/');
+        $this->registerPattern('/'.$pattern.'\s*\n?(?=\s*('.implode('|', $operators).'))/', $callback);
     }
 
     /**

diff --git a/tests/js/JSTest.php b/tests/js/JSTest.php
@@ -88,6 +88,10 @@ public function dataProvider()
             'a = b / c; d = e / f',
             'a=b/c;d=e/f',
         );
+        $tests[] = array(
+            '(2 + 4) / 3 + 5 / 1',
+            '(2+4)/3+5/1',
+        );
 
         $tests[] = array(
             'a=4/
@@ -704,6 +708,12 @@ function isJSON(){str.replace(/"[^"\\\n\r]*"|true|false|null|-?\d+(?:\.\d*)?(?:[
             'function isHtmlNamespace(node){var ns;return typeof node.namespaceURI==UNDEF||((ns=node.namespaceURI)===null||ns=="http://www.w3.org/1999/xhtml")}',
         );
 
+        // https://github.com/matthiasmullie/minify/issues/115
+        $tests[] = array(
+            'if(typeof i[s].token=="string")/keyword|support|storage/.test(i[s].token)&&n.push(i[s].regex);else if(typeof i[s].token=="object")for(var u=0,a=i[s].token.length;u<a;u++)if(/keyword|support|storage/.test(i[s].token[u])){}',
+            'if(typeof i[s].token=="string")/keyword|support|storage/.test(i[s].token)&&n.push(i[s].regex);else if(typeof i[s].token=="object")for(var u=0,a=i[s].token.length;u<a;u++)if(/keyword|support|storage/.test(i[s].token[u])){}',
+        );
+
         // update tests' expected results for cross-system compatibility
         foreach ($tests as &$test) {
             if (!empty($test[1])) {