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

compiled/RegExpFilter.cpp

+#include <emscripten.h>
+
+#include "RegExpFilter.h"
+#include "WhiteListFilter.h"
+#include "InvalidFilter.h"
+
+namespace
+{
+  int GenerateRegExp(const std::u16string& source)
+  {
+    // 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];
+      switch (currChar)
+      {
+        case u'*':
+          if (prevChar != u'*')
+            result += u".*";
+          break;
+        case u'^':
+          result += u"(?:[\\x00-\\x24\\x26-\\x2C\\x2F\\x3A-\\x40\\x5B-\\x5E\\x60\\x7B-\\x7F]|$)";
+          break;
+        case u'|':
+          if (i == 0)
+          {
+            // Anchor at expression start, maybe extended anchor?
+            if (i + 1 < l && source[i + 1] == u'|')
+            {
+              result += u"^[\\w\\-]+:\\/+(?!\\/)(?:[^\\/]+\\.)?";
+              ++i;
+            }
+            else
+              result += u"^";
+          }
+          else if (i == l - 1)
+          {
+            // Anchor at expression end, ignore if following separator placeholder
+            if (prevChar != u'^')
+              result += u"$";
+          }
+          else
+          {
+            // Not actually an anchor, escape it
+            result += u"\\|";
+          }
+          break;
+        default:
+          if ((currChar >= u'a' && currChar <= u'z') ||
+              (currChar >= u'A' && currChar <= u'Z') ||
+              (currChar >= u'0' && currChar <= u'9') ||
+              currChar >= 128)
+          {
+            result += currChar;
+          }
+          else
+          {
+            result += u"\\";
+            result.append(1, currChar);
+          }
+      }
+      prevChar = currChar;
+    }
+    return EM_ASM_INT(return regexps.create($0, $1), &result, false);
+  }
+}
+
+RegExpFilter::RegExpFilter(const std::u16string& text,
+    const std::u16string& pattern, const std::u16string& options)
+    : ActiveFilter(text), regexpId(0)
+{
+  size_t len = pattern.length();
+  if (len >= 2 && pattern[0] == u'/' && pattern[len - 1] == u'/')
+  {
+    std::u16string param = pattern.substr(1, len - 2);
+    regexpId = EM_ASM_INT(return regexps.create($0, $1), &param, false);
+
+    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;
+}
+
+RegExpFilter::~RegExpFilter()
+{
+  if (regexpId)
+    EM_ASM_ARGS(regexps.delete($0), regexpId);
+}
+
+Filter* RegExpFilter::Create(const std::u16string& text)
+{
+  bool blocking = true;
+  size_t patternStart = 0;
+  if (!text.compare(0, 2, u"@@"))
+  {
+    blocking = false;
+    patternStart = 2;
+  }
+
+  size_t patternEnd = text.find(u'$', 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) : 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);
+  }
+}
+
+Filter::Type RegExpFilter::GetType() const
+{
+  return Type::BLOCKING;
+}
+
+bool RegExpFilter::Matches(const std::u16string& location)
+{
+  if (!regexpId)
+  {
+    regexpId = GenerateRegExp(regexpSource);
+    regexpSource.resize(0);
+  }
+  return EM_ASM_INT(return regexps.test($0, $1), regexpId, &location);
+}