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

compiled/Filter.cpp

Index: compiled/Filter.cpp
===================================================================
new file mode 100644
--- /dev/null
+++ b/compiled/Filter.cpp
@@ -0,0 +1,105 @@
+#include <unordered_map>
+
+#include "Filter.h"
+#include "CommentFilter.h"
+#include "RegExpFilter.h"
+#include "ElemHideFilter.h"
+#include "ElemHideException.h"
+
+namespace
+{
+  std::unordered_map<std::u16string,FilterPtr> knownFilters;
+
+  void trim_spaces(std::u16string& str)
+  {
+    size_t pos;
+
+    // Remove leading whitespace
+    pos = str.find_first_not_of(u' ');
+    if (pos > 0)
+      str.erase(0, pos);
+
+    // Remove trailing whitespace
+    pos = str.find_last_not_of(u' ');
+    if (pos < str.length() - 1)
+      str.erase(pos + 1);
+  }
+
+  void remove_spaces(std::u16string& str)
+  {
+    for (size_t i = 0, l = str.length(); i < l; ++i)
+    {
+      if (str[i] == u' ')
+      {
+        str.erase(i, 1);
+        --i;
+        --l;
+      }
+    }
+  }
+}
+
+Filter::Filter(const std::u16string& text)
+{
+  this->text = text;
+}
+
+const std::u16string Filter::Serialize()
+{
+  return (
+    u"[Filter]\n"
+    u"text=" + text + u"\n"
+  );
+}
+
+FilterPtr Filter::FromText(const std::u16string& text)
+{
+  auto it = knownFilters.find(text);
+  if (it != knownFilters.end())
+    return it->second;
+
+  FilterPtr filter = FilterPtr(CommentFilter::Create(text));
+  if (!filter)
+    filter.reset(ElemHideBase::Create(text));
+  if (!filter)
+    filter.reset(RegExpFilter::Create(text));
+  return knownFilters[text] = filter;
+}
+
+const std::u16string Filter::Normalize(const std::u16string& text)
+{
+  std::u16string result(text);
+
+  // Remove special characters like line breaks
+  for (size_t i = 0, l = result.length(); i < l; ++i)
+  {
+    if (result[i] < u' ')
+    {
+      result.erase(i, 1);
+      --i;
+      --l;
+    }
+  }
+
+  trim_spaces(result);
+
+  {
+    size_t domainsEnd;
+    size_t selectorStart;
+    Filter::Type type = ElemHideBase::Parse(result, &domainsEnd, &selectorStart);
+    if (type != Filter::Type::UNKNOWN)
+    {
+      std::u16string domains = result.substr(0, domainsEnd);
+      std::u16string selector = result.substr(selectorStart);
+      remove_spaces(domains);
+      trim_spaces(selector);
+      return domains + (
+        type == Filter::Type::ELEMHIDEEXCEPTION ? u"#@#" : u"##"
+      ) + selector;
+    }
+  }
+
+  if (CommentFilter::Parse(result) == Filter::Type::UNKNOWN)
+    remove_spaces(result);
+  return result;
+}