Changeset 17354 in webkit for trunk/JavaScriptCore/kjs/lexer.cpp

Timestamp:

Oct 26, 2006, 10:30:25 PM (19 years ago)

Author:

bdash

Message:

2006-10-26 W. Andy Carrel <​[email protected]>

Reviewed by Maciej.

• Fix ​http://bugs.webkit.org/show_bug.cgi?id=7445 / <rdar://problem/4614195> (and 7253 / <rdar://4694011>) by changing inline regexps so that they can have \u escaped Unicode sequences and still work properly.

• kjs/lexer.cpp: (Lexer::Lexer): (Lexer::setCode): (Lexer::shift): Looking ahead one additional character for the benefit of scanRegExp (Lexer::scanRegExp): Change code to support unicode escapes in inline regexps.
• kjs/lexer.h: Extra lookahead added.
• tests/mozilla/ecma_2/RegExp/properties-001.js: Changed test to look for Unicode character rather than the '\u' escaped equivalent for .source and .toString().

File:

• trunk/JavaScriptCore/kjs/lexer.cpp (modified) (5 diffs)

trunk/JavaScriptCore/kjs/lexer.cpp

@@ -61 +61 @@
     bol(true),
 #endif
-    current(0), next1(0), next2(0), next3(0),
+    current(0), next1(0), next2(0), next3(0), next4(0),
     strings(0), numStrings(0), stringsCapacity(0),
     identifiers(0), numIdentifiers(0), identifiersCapacity(0)
@@ -119 +119 @@
   next2 = (length > 2) ? code[2].uc : -1;
   next3 = (length > 3) ? code[3].uc : -1;
+  next4 = (length > 4) ? code[4].uc : -1;
 }
 
@@ -130 +131 @@
     next1 = next2;
     next2 = next3;
-    next3 = (pos + 3 < length) ? code[pos+3].uc : -1;
+    next3 = next4;
+    next4 = (pos + 4 < length) ? code[pos+4].uc : -1;
   }
 }
@@ -837 +839 @@
     else if (current != '/' || lastWasEscape == true || inBrackets == true)
     {
-        // keep track of '[' and ']'
-        if ( !lastWasEscape ) {
+        if (lastWasEscape) {
+          // deal with unicode escapes in inline regexps
+          if (current == 'u') {
+            if (isHexDigit(next1) && isHexDigit(next2) && 
+                isHexDigit(next3) && isHexDigit(next4)) {
+              record16(convertUnicode(next1, next2, next3, next4));
+              shift(5);
+              lastWasEscape = false;
+              continue;
+            } else 
+              // this wasn't unicode after all
+              record16('\\');
+          }
+        } else {
+          // keep track of '[' and ']'
           if ( current == '[' && !inBrackets )
             inBrackets = true;
@@ -844 +859 @@
             inBrackets = false;
         }
-        record16(current);
+        // don't want to capture the '\' for unicode escapes
+        if (current != '\\' || next1 != 'u')
+          record16(current);
         lastWasEscape =
             !lastWasEscape && (current == '\\');