Changeset 18182 in webkit for trunk/JavaScriptCore/kjs

Timestamp:

Dec 12, 2006, 12:09:58 PM (18 years ago)

Author:

ap

Message:

Reviewed by Geoff. Based on a patch by Maks Orlovich.

​http://bugs.webkit.org/show_bug.cgi?id=6257
Throw errors on invalid expressions (KJS merge)

JavaScriptCore:

• kjs/regexp.cpp: (KJS::RegExp::RegExp): (KJS::RegExp::~RegExp): (KJS::RegExp::match):
• kjs/regexp.h: (KJS::RegExp::flags): (KJS::RegExp::isValid): (KJS::RegExp::errorMessage): (KJS::RegExp::subPatterns): Remember and report RegExp construction failures. Renamed data members not to start with underscores.

• kjs/regexp_object.cpp: (RegExpObjectImp::construct): Raise an exception if RegExp construction fails. (RegExpObjectImp::callAsFunction): Removed an obsolete comment.

• tests/mozilla/ecma_3/RegExp/regress-119909.js: Reduced the number of nested parentheses to a value supported by PCRE.

LayoutTests:

• fast/js/kde/RegExp-expected.txt: One more test passes.

Location:

trunk/JavaScriptCore/kjs

Files:

• regexp.cpp (modified) (10 diffs)
• regexp.h (modified) (1 diff)
• regexp_object.cpp (modified) (3 diffs)

trunk/JavaScriptCore/kjs/regexp.cpp

@@ -2 +2 @@
 /*
  *  This file is part of the KDE libraries
- *  Copyright (C) 1999-2001 Harri Porten ([email protected])
+ *  Copyright (C) 1999-2001,2004 Harri Porten ([email protected])
  *
  *  This library is free software; you can redistribute it and/or
@@ -23 +23 @@
 #include "regexp.h"
 
+#include "lexer.h"
+
 #include <assert.h>
 #include <stdio.h>
@@ -31 +33 @@
 
 RegExp::RegExp(const UString &p, int flags)
-  : _flags(flags), _numSubPatterns(0)
+  : m_flags(flags), m_constructionError(0), m_numSubPatterns(0)
 {
 #if HAVE(PCREPOSIX)
@@ -49 +51 @@
   
   pattern.append('\0');
-  _regex = pcre_compile(reinterpret_cast<const uint16_t*>(pattern.data()),
+  m_regex = pcre_compile(reinterpret_cast<const uint16_t*>(pattern.data()),
                         options, &errorMessage, &errorOffset, NULL);
-  if (!_regex) {
+  if (!m_regex) {
     // Try again, this time handle any \u we might find.
     UString uPattern = sanitizePattern(pattern);
-    _regex = pcre_compile(reinterpret_cast<const uint16_t*>(uPattern.data()),
+    m_regex = pcre_compile(reinterpret_cast<const uint16_t*>(uPattern.data()),
                           options, &errorMessage, &errorOffset, NULL);
-    if (!_regex) 
+    if (!m_regex) {
+      m_constructionError = strdup(errorMessage);
       return;
+    }
   }
 
 #ifdef PCRE_INFO_CAPTURECOUNT
   // Get number of subpatterns that will be returned.
-  pcre_fullinfo(_regex, NULL, PCRE_INFO_CAPTURECOUNT, &_numSubPatterns);
+  pcre_fullinfo(m_regex, NULL, PCRE_INFO_CAPTURECOUNT, &m_numSubPatterns);
 #endif
 
@@ -81 +85 @@
   // Note: the Global flag is already handled by RegExpProtoFunc::execute
 
-  regcomp(&_regex, p.ascii(), regflags);
-  /* TODO check for errors */
+  // FIXME: support \u Unicode escapes.
+
+  int errorCode = regcomp(&m_regex, intern.ascii(), regflags);
+  if (errorCode != 0) {
+    char errorMessage[80];
+    regerror(errorCode, &m_regex, errorMessage, sizeof errorMessage);
+    m_constructionError = strdup(errorMessage);
+  }
 
 #endif
@@ -90 +100 @@
 {
 #if HAVE(PCREPOSIX)
-  pcre_free(_regex);
+  pcre_free(m_regex);
 #else
   /* TODO: is this really okay after an error ? */
-  regfree(&_regex);
-#endif
+  regfree(&m_regex);
+#endif
+  free(m_constructionError);
 }
 
@@ -113 +124 @@
 #if HAVE(PCREPOSIX)
 
-  if (!_regex)
+  if (!m_regex)
     return UString::null();
 
@@ -125 +136 @@
     offsetVector = fixedSizeOffsetVector;
   } else {
-    offsetVectorSize = (_numSubPatterns + 1) * 3;
+    offsetVectorSize = (m_numSubPatterns + 1) * 3;
     offsetVector = new int [offsetVectorSize];
   }
 
-  const int numMatches = pcre_exec(_regex, NULL, reinterpret_cast<const uint16_t *>(s.data()), s.size(), i, 0, offsetVector, offsetVectorSize);
+  const int numMatches = pcre_exec(m_regex, NULL, reinterpret_cast<const uint16_t *>(s.data()), s.size(), i, 0, offsetVector, offsetVectorSize);
 
   if (numMatches < 0) {
@@ -152 +163 @@
 
   char *str = strdup(s.ascii()); // TODO: why ???
-  if (regexec(&_regex, str + i, maxMatch, rmatch, 0)) {
+  if (regexec(&m_regex, str + i, maxMatch, rmatch, 0)) {
     free(str);
     return UString::null();
@@ -164 +175 @@
 
   // map rmatch array to ovector used in PCRE case
-  _numSubPatterns = 0;
+  m_numSubPatterns = 0;
   for(unsigned j = 1; j < maxMatch && rmatch[j].rm_so >= 0 ; j++)
-      _numSubPatterns++;
-  int ovecsize = (_numSubPatterns+1)*3; // see above
+      m_numSubPatterns++;
+  int ovecsize = (m_numSubPatterns+1)*3; // see above
   *ovector = new int[ovecsize];
-  for (unsigned j = 0; j < _numSubPatterns + 1; j++) {
+  for (unsigned j = 0; j < m_numSubPatterns + 1; j++) {
     if (j>maxMatch)
       break;

trunk/JavaScriptCore/kjs/regexp.h

@@ -46 +46 @@
     ~RegExp();
 
-    int flags() const { return _flags; }
+    int flags() const { return m_flags; }
+    bool isValid() const { return !m_constructionError; }
+    const char* errorMessage() const { return m_constructionError; }
 
     UString match(const UString &s, int i, int *pos = 0, int **ovector = 0);
-    unsigned subPatterns() const { return _numSubPatterns; }
+    unsigned subPatterns() const { return m_numSubPatterns; }
 
   private:
 #if HAVE(PCREPOSIX)
-    pcre *_regex;
+    pcre *m_regex;
 #else
-    regex_t _regex;
+    regex_t m_regex;
 #endif
-    int _flags;
-    unsigned _numSubPatterns;
+    int m_flags;
+    char* m_constructionError;
+    unsigned m_numSubPatterns;
 
     RegExp(const RegExp &);

trunk/JavaScriptCore/kjs/regexp_object.cpp

@@ -383 +383 @@
   bool ignoreCase = (flags.find("i") >= 0);
   bool multiline = (flags.find("m") >= 0);
-  // TODO: throw a syntax error on invalid flags
 
   dat->putDirect("global", jsBoolean(global), DontDelete | ReadOnly | DontEnum);
@@ -399 +398 @@
   if (multiline)
       reflags |= RegExp::Multiline;
-  dat->setRegExp(new RegExp(p, reflags));
+  RegExp* re = new RegExp(p, reflags);
+  if (!re->isValid()) {
+    delete re;
+    return throwError(exec, SyntaxError, UString("Invalid regular expression: ").append(re->errorMessage()));
+  }
+  dat->setRegExp(re);
 
   return dat;
@@ -407 +411 @@
 JSValue *RegExpObjectImp::callAsFunction(ExecState *exec, JSObject * /*thisObj*/, const List &args)
 {
-  // TODO: handle RegExp argument case (15.10.3.1)
-
   return construct(exec, args);
 }