Issue 4125 - [emscripten] Convert filter classes to C++

compiled/RegExpFilter.cpp

 #include <climits>
-#include <unordered_map>
 
 #include <emscripten.h>
 
 #include "RegExpFilter.h"
-#include "WhiteListFilter.h"
-#include "InvalidFilter.h"
 #include "StringScanner.h"
+#include "StringMap.h"
 
 namespace
 {
   enum
   {
     TYPE_OTHER = 0x1,
     TYPE_SCRIPT = 0x2,
     TYPE_IMAGE = 0x4,
     TYPE_STYLESHEET = 0x8,
     TYPE_OBJECT = 0x10,
     TYPE_SUBDOCUMENT = 0x20,
     TYPE_DOCUMENT = 0x40,
     TYPE_PING = 0x400,
     TYPE_XMLHTTPREQUEST = 0x800,
     TYPE_OBJECT_SUBREQUEST = 0x1000,
     TYPE_MEDIA = 0x4000,
     TYPE_FONT = 0x8000,
     TYPE_POPUP = 0x8000000,
     TYPE_GENERICBLOCK = 0x10000000,
     TYPE_GENERICHIDE = 0x20000000,
     TYPE_ELEMHIDE = 0x40000000,
   };
 
-  std::unordered_map<std::u16string,int> typeMap({
-    {u"OTHER", TYPE_OTHER},
-    {u"SCRIPT", TYPE_SCRIPT},
-    {u"IMAGE", TYPE_IMAGE},
-    {u"STYLESHEET", TYPE_STYLESHEET},
-    {u"OBJECT", TYPE_OBJECT},
-    {u"SUBDOCUMENT", TYPE_SUBDOCUMENT},
-    {u"DOCUMENT", TYPE_DOCUMENT},
-    {u"XBL", TYPE_OTHER},          // Backwards compat
-    {u"PING", TYPE_PING},
-    {u"XMLHTTPREQUEST", TYPE_XMLHTTPREQUEST},
-    {u"OBJECT_SUBREQUEST", TYPE_OBJECT_SUBREQUEST},
-    {u"DTD", TYPE_OTHER},          // Backwards compat
-    {u"MEDIA", TYPE_MEDIA},
-    {u"FONT", TYPE_FONT},
-    {u"BACKGROUND", TYPE_IMAGE},   // Backwards compat
-
-    {u"POPUP", TYPE_POPUP},
-    {u"GENERICBLOCK", TYPE_GENERICBLOCK},
-    {u"GENERICHIDE", TYPE_GENERICHIDE},
-    {u"ELEMHIDE", TYPE_ELEMHIDE},
-  });
-
-  int defaultTypeMask = INT_MAX & ~(TYPE_DOCUMENT | TYPE_ELEMHIDE | TYPE_POPUP |
-      TYPE_GENERICBLOCK | TYPE_GENERICHIDE);
-
-  int GenerateRegExp(const std::u16string& regexp, bool matchCase)
+  StringMap<int> typeMap {
+    {u"other"_str, TYPE_OTHER},
+    {u"script"_str, TYPE_SCRIPT},
+    {u"image"_str, TYPE_IMAGE},
+    {u"stylesheet"_str, TYPE_STYLESHEET},
+    {u"object"_str, TYPE_OBJECT},
+    {u"subdocument"_str, TYPE_SUBDOCUMENT},
+    {u"document"_str, TYPE_DOCUMENT},
+    {u"xbl"_str, TYPE_OTHER},          // Backwards compat
+    {u"ping"_str, TYPE_PING},
+    {u"xmlhttprequest"_str, TYPE_XMLHTTPREQUEST},
+    {u"object-subrequest"_str, TYPE_OBJECT_SUBREQUEST},
+    {u"dtd"_str, TYPE_OTHER},          // Backwards compat
+    {u"media"_str, TYPE_MEDIA},
+    {u"font"_str, TYPE_FONT},
+    {u"background"_str, TYPE_IMAGE},   // Backwards compat
+
+    {u"popup"_str, TYPE_POPUP},
+    {u"genericblock"_str, TYPE_GENERICBLOCK},
+    {u"generichide"_str, TYPE_GENERICHIDE},
+    {u"elemhide"_str, TYPE_ELEMHIDE},
+  };
+
+  const int defaultTypeMask = INT_MAX & ~(TYPE_DOCUMENT | TYPE_ELEMHIDE |
+      TYPE_POPUP | TYPE_GENERICBLOCK | TYPE_GENERICHIDE);
+
+  int GenerateRegExp(const String& regexp, bool matchCase)
   {
     return EM_ASM_INT(return regexps.create($0, $1), &regexp, matchCase);
   }
-}
-
-RegExpFilter::RegExpFilter(const std::u16string& text,
-    const std::u16string& pattern, const std::u16string& options)
-    : ActiveFilter(text, true), regexpId(0), contentType(-1), matchCase(false),
-      thirdParty(TrippleState::ANY)
-{
-  int optionStart = 0;
-  int optionEnd = -1;
-  int valueStart = -1;
-  StringScanner scanner(options + u",");
-  while (!scanner.done())
-  {
-    switch (scanner.next())
-    {
-      case u'=':
-        if (optionEnd < 0)
-        {
-          optionEnd = scanner.position();
-          valueStart = optionEnd + 1;
-        }
+
+  void NormalizeWhitespace(DependentString& text)
+  {
+    // We want to remove all spaces but bail out early in the common scenario
+    // that the string contains no spaces.
+
+    // Look for the first space
+    String::size_type len = text.length();
+    String::size_type pos;
+    for (pos = 0; pos < len; pos++)
+      if (text[pos] == ' ')
         break;
-      case u',':
-        if (optionEnd < 0)
-          optionEnd = scanner.position();
-        ProcessOption(options, optionStart, optionEnd, valueStart, scanner.position());
-        optionStart = scanner.position() + 1;
-        optionEnd = -1;
-        valueStart = -1;
-        break;
-    }
-  }
-  if (contentType < 0)
-    contentType = defaultTypeMask;
-
-  size_t len = pattern.length();
-  if (len >= 2 && pattern[0] == u'/' && pattern[len - 1] == u'/')
-  {
-    regexpId = GenerateRegExp(pattern.substr(1, len - 2), matchCase);
-
-    std::u16string* error = reinterpret_cast<std::u16string*>(EM_ASM_INT(return regexps.getError($0), regexpId));
-    if (error)
-    {
-      EM_ASM_ARGS(regexps.delete($0), regexpId);
-      throw std::u16string(*error);
-    }
-  }
-  else
-    regexpSource = pattern;
+
+    if (pos >= len)
+      return;
+
+    // Found spaces, move characters to remove them
+    String::size_type delta = 1;
+    for (pos = pos + 1; pos < len; pos++)
+    {
+      if (text[pos] == ' ')
+        delta++;
+      else
+        text[pos - delta] = text[pos];
+    }
+    text.reset(text, 0, len - delta);
+  }
+
+  void ParseOption(String& text, DependentString& error, RegExpFilterData& data,
+      int optionStart, int optionEnd, int valueStart, int valueEnd)
+  {
+    if (optionEnd <= optionStart)
+      return;
+
+    bool reverse = false;
+    if (text[optionStart] == u'~')
+    {
+      reverse = true;
+      optionStart++;
+    }
+
+    DependentString name(text, optionStart, optionEnd - optionStart);
+    for (size_t i = 0; i < name.length(); ++i)
+    {
+      char16_t currChar = name[i];
+      if (currChar >= u'A' && currChar <= u'Z')
+        name[i] = currChar + u'a' - u'A';
+      else if (currChar == u'_')
+        name[i] = u'-';
+    }
+
+    auto it = typeMap.find(name);
+    if (it)
+    {
+      if (data.mContentType < 0)
+        data.mContentType = reverse ? defaultTypeMask : 0;
+      if (reverse)
+        data.mContentType &= ~it->second;
+      else
+        data.mContentType |= it->second;
+    }
+    else if (name.equals(u"domain"_str))
+    {
+      if (valueStart >= 0 && valueEnd > valueStart)
+      {
+        data.mDomainsStart = valueStart;
+        data.mDomainsEnd = valueEnd;
+        DependentString(text, valueStart, valueEnd - valueStart).toLower();
+      }
+    }
+    else if (name.equals(u"sitekey"_str))
+    {
+      if (valueStart >= 0 && valueEnd > valueStart)
+      {
+        data.mSitekeysStart = valueStart;
+        data.mSitekeysEnd = valueEnd;
+      }
+    }
+    else if (name.equals(u"match-case"_str))
+      data.mMatchCase = !reverse;
+    else if (name.equals(u"third-party"_str))
+      data.mThirdParty = reverse ? TrippleState::NO : TrippleState::YES;
+    else if (name.equals(u"collapse"_str))
+      data.mCollapse = reverse ? TrippleState::NO : TrippleState::YES;
+    else
+      error.reset(u"filter_unknown_option"_str);
+  }
+
+  void ParseOptions(String& text, DependentString& error, RegExpFilterData& data,
+      String::size_type optionsStart)
+  {
+    data.mMatchCase = false;
+    data.mThirdParty = TrippleState::ANY;
+    data.mCollapse = TrippleState::ANY;
+    data.mDomainsStart = String::npos;
+    data.mSitekeysStart = String::npos;
+    if (optionsStart >= text.length())
+    {
+      data.mContentType = defaultTypeMask;
+      return;
+    }
+
+    data.mContentType = -1;
+
+    int optionStart = data.mPatternEnd + 1;
+    int optionEnd = -1;
+    int valueStart = -1;
+
+    StringScanner scanner(text, optionStart, u',');
+    bool done = false;
+    while (!done)
+    {
+      done = scanner.done();
+      switch (scanner.next())
+      {
+        case u'=':
+          if (optionEnd < 0)
+          {
+            optionEnd = scanner.position();
+            valueStart = optionEnd + 1;
+          }
+          break;
+        case u',':
+          if (optionEnd < 0)
+            optionEnd = scanner.position();
+          ParseOption(text, error, data, optionStart, optionEnd, valueStart,
+              scanner.position());
+          if (!error.empty())
+            return;
+
+          optionStart = scanner.position() + 1;
+          optionEnd = -1;
+          valueStart = -1;
+          break;
+      }
+    }
+
+    if (data.mContentType < 0)
+      data.mContentType = defaultTypeMask;
+  }
+}
+
+RegExpFilter::RegExpFilter(Type type, const String& text, const RegExpFilterData& data)
+    : ActiveFilter(type, text, true), mData(data)
+{
 }
 
 RegExpFilter::~RegExpFilter()
 {
-  if (regexpId)
-    EM_ASM_ARGS(regexps.delete($0), regexpId);
-}
-
-void RegExpFilter::ProcessOption(const std::u16string& options,
-    int optionStart, int optionEnd, int valueStart, int valueEnd)
-{
-  if (optionEnd <= optionStart)
-    return;
-
-  bool reverse = false;
-  if (options[optionStart] == u'~')
-  {
-    reverse = true;
-    optionStart++;
-  }
-
-  std::u16string name(options.substr(optionStart, optionEnd - optionStart));
-  for (size_t i = 0, l = name.length(); i < l; ++i)
-  {
-    char16_t currChar = name[i];
-    if (currChar >= u'a' && currChar <= u'z')
-      name[i] = currChar + u'A' - u'a';
-    else if (currChar == u'-')
-      name[i] = u'_';
-  }
-
-  auto it = typeMap.find(name);
-  if (it != typeMap.end())
-  {
-    if (contentType < 0)
-      contentType = reverse ? defaultTypeMask : 0;
-    if (reverse)
-      contentType &= ~it->second;
-    else
-      contentType |= it->second;
-  }
-  else if (!name.compare(u"DOMAIN"))
-  {
-    if (valueStart >= 0 && valueEnd > valueStart)
-      ParseDomains(options.substr(valueStart, valueEnd - valueStart), u'|');
-  }
-  else if (!name.compare(u"SITEKEY"))
-  {
-    if (valueStart >= 0 && valueEnd > valueStart)
-    {
-      StringScanner scanner(options.substr(valueStart, valueEnd - valueStart) + u"|");
-      size_t start = 0;
-      while (!scanner.done())
-      {
-        if (scanner.next() == u'|')
-        {
-          if (scanner.position() > start)
-            sitekeys.insert(options.substr(valueStart + start, scanner.position() - start));
-          start = scanner.position() + 1;
-        }
-      }
-    }
-  }
-  else if (!name.compare(u"MATCH_CASE"))
-    matchCase = !reverse;
-  else if (!name.compare(u"THIRD_PARTY"))
-    thirdParty = reverse ? TrippleState::NO : TrippleState::YES;
-  else if (!name.compare(u"COLLAPSE"))
-    collapse = reverse ? TrippleState::NO : TrippleState::YES;
-  else
-    throw std::u16string(u"Unknown option " + name);
-}
-
-Filter* RegExpFilter::Create(const std::u16string& text)
-{
-  bool blocking = true;
-  size_t patternStart = 0;
-  if (!text.compare(0, 2, u"@@"))
-  {
-    blocking = false;
-    patternStart = 2;
-  }
-
-  size_t patternEnd = text.find(u'$', patternStart);
-  size_t patternLength = (patternEnd != std::u16string::npos ?
-      patternEnd - patternStart : patternEnd);
-  std::u16string pattern(text.substr(patternStart, patternLength));
-  std::u16string options(patternEnd != std::u16string::npos ?
-      text.substr(patternEnd + 1) : u"");
-
-  try
-  {
-    if (blocking)
-      return new RegExpFilter(text, pattern, options);
-    else
-      return new WhiteListFilter(text, pattern, options);
-  }
-  catch (const std::u16string& reason)
-  {
-    return new InvalidFilter(text, reason);
+  if (mData.HasRegExp())
+    EM_ASM_ARGS(regexps.delete($0), mData.mRegexpId);
+}
+
+Filter::Type RegExpFilter::Parse(DependentString& text, DependentString& error,
+    RegExpFilterData& data)
+{
+  NormalizeWhitespace(text);
+
+  Filter::Type type = Type::BLOCKING;
+
+  data.mPatternStart = 0;
+  if (text.length() >= 2 && text[0] == u'@' && text[1] == u'@')
+  {
+    type = Type::WHITELIST;
+    data.mPatternStart = 2;
+  }
+
+  data.mPatternEnd = text.find(u'$', data.mPatternStart);
+  if (data.mPatternEnd == text.npos)
+    data.mPatternEnd = text.length();
+
+  ParseOptions(text, error, data, data.mPatternEnd + 1);
+  if (!error.empty())
+    return Type::INVALID;
+
+  if (data.mPatternEnd - data.mPatternStart >= 2 &&
+      text[data.mPatternStart] == u'/' &&
+      text[data.mPatternEnd - 1] == u'/')
+  {
+    data.SetRegExp(GenerateRegExp(DependentString(text, data.mPatternStart + 1,
+        data.mPatternEnd - data.mPatternStart - 2), data.mMatchCase));
+    if (data.mRegexpId == -1)
+    {
+      error.reset(u"filter_invalid_regexp"_str);
+      return Type::INVALID;
+    }
+  }
+
+  return type;
+}
+
+void RegExpFilter::ParseSitekeys(const String& sitekeys) const
+{
+  StringScanner scanner(sitekeys, 0, u'|');
+  size_t start = 0;
+  bool done = false;
+  while (!done)
+  {
+    done = scanner.done();
+    if (scanner.next() == u'|')
+    {
+      if (scanner.position() > start)
+        AddSitekey(DependentString(sitekeys, start, scanner.position() - start));
+      start = scanner.position() + 1;
+    }
   }
 }
 
 void RegExpFilter::InitJSTypes()
 {
+  EM_ASM(exports.RegExpFilter.typeMap = {};);
   for (auto it = typeMap.begin(); it != typeMap.end(); ++it)
-    EM_ASM_ARGS(Module.RegExpFilter_typeMap[getStringData($0)] = $1, &(it->first), it->second);
-}
-
-const std::u16string RegExpFilter::RegExpFromSource(const std::u16string& source)
-{
+    EM_ASM_ARGS(exports.RegExpFilter.typeMap[readString($0).replace("-", "_").toUpperCase()] = $1, &(it->first), it->second);
+}
+
+OwnedString RegExpFilter::RegExpFromSource(const String& source)
+{
+  /* TODO: this is very inefficient */
+
   // Note: This doesn't remove trailing wildcards, otherwise the result should
   // be identical to Filter.toRegExp().
-  std::u16string result;
-  char16_t prevChar = u'*';
-  for (size_t i = 0, l = source.length(); i < l; ++i)
-  {
-    char16_t currChar = source[i];
+  OwnedString result;
+  String::value_type prevChar = u'*';
+  for (String::size_type i = 0; i < source.length(); ++i)
+  {
+    String::value_type currChar = source[i];
     switch (currChar)
     {
       case u'*':
         if (prevChar != u'*')
-          result += u".*";
+          result.append(u".*"_str);
         break;
       case u'^':
-        result += u"(?:[\\x00-\\x24\\x26-\\x2C\\x2F\\x3A-\\x40\\x5B-\\x5E\\x60\\x7B-\\x7F]|$)";
+        result.append(u"(?:[\\x00-\\x24\\x26-\\x2C\\x2F\\x3A-\\x40\\x5B-\\x5E\\x60\\x7B-\\x7F]|$)"_str);
         break;
       case u'|':
         if (i == 0)
         {
           // Anchor at expression start, maybe extended anchor?
-          if (i + 1 < l && source[i + 1] == u'|')
+          if (i + 1 < source.length() && source[i + 1] == u'|')
           {
-            result += u"^[\\w\\-]+:\\/+(?!\\/)(?:[^\\/]+\\.)?";
+            result.append(u"^[\\w\\-]+:\\/+(?!\\/)(?:[^\\/]+\\.)?"_str);
             ++i;
           }
           else
-            result += u"^";
+            result.append(u'^');
         }
-        else if (i == l - 1)
+        else if (i == source.length() - 1)
         {
           // Anchor at expression end, ignore if following separator placeholder
           if (prevChar != u'^')
-            result += u"$";
+            result.append(u'$');
         }
         else
         {
           // Not actually an anchor, escape it
-          result += u"\\|";
+          result.append(u"\\|"_str);
         }
         break;
       default:
-        if ((currChar >= u'a' && currChar <= u'z') ||
-            (currChar >= u'A' && currChar <= u'Z') ||
-            (currChar >= u'0' && currChar <= u'9') ||
-            currChar >= 128)
+        if (!(currChar >= u'a' && currChar <= u'z') &&
+            !(currChar >= u'A' && currChar <= u'Z') &&
+            !(currChar >= u'0' && currChar <= u'9') &&
+            currChar < 128)
         {
-          result += currChar;
+          result.append(u'\\');
         }
-        else
-        {
-          result += u"\\";
-          result.append(1, currChar);
-        }
+        result.append(currChar);
     }
     prevChar = currChar;
   }
   return result;
 }
 
-Filter::Type RegExpFilter::GetType() const
-{
-  return Type::BLOCKING;
-}
-
-bool RegExpFilter::Matches(const std::u16string& location, int typeMask,
-    const std::u16string& docDomain, bool thirdParty,
-    const std::u16string& sitekey)
-{
-  if (!(this->contentType & typeMask) ||
-      (this->thirdParty == TrippleState::YES && !thirdParty) ||
-      (this->thirdParty == TrippleState::NO && thirdParty) ||
+RegExpFilter::DomainMap* RegExpFilter::GetDomains() const
+{
+  if (!mData.DomainsParsingDone())
+  {
+    ParseDomains(mData.GetDomainsSource(mText), u'|');
+    mData.SetDomainsParsingDone();
+  }
+  return ActiveFilter::GetDomains();
+}
+
+RegExpFilter::SitekeySet* RegExpFilter::GetSitekeys() const
+{
+  if (!mData.SitekeyParsingDone())
+  {
+    ParseSitekeys(mData.GetSitekeysSource(mText));
+    mData.SetSitekeysParsingDone();
+  }
+  return ActiveFilter::GetSitekeys();
+}
+
+bool RegExpFilter::Matches(const String& location, int typeMask,
+    DependentString& docDomain, bool thirdParty, const String& sitekey) const
+{
+  if (!(mData.mContentType & typeMask) ||
+      (mData.mThirdParty == TrippleState::YES && !thirdParty) ||
+      (mData.mThirdParty == TrippleState::NO && thirdParty) ||
       !IsActiveOnDomain(docDomain, sitekey))
   {
     return false;
   }
 
-  if (!regexpId)
-  {
-    regexpId = GenerateRegExp(RegExpFromSource(regexpSource), matchCase);
-    regexpSource.resize(0);
-  }
-  return EM_ASM_INT(return regexps.test($0, $1), regexpId, &location);
-}
+  if (!mData.RegExpParsingDone())
+  {
+    const OwnedString pattern(mData.GetRegExpSource(mText));
+    mData.SetRegExp(GenerateRegExp(RegExpFromSource(pattern), mData.mMatchCase));
+  }
+  return EM_ASM_INT(return regexps.test($0, $1), mData.mRegexpId, &location);
+}