Issue 5141 - Convert filter match to C++

compiled/StringMap.h

 /*
  * This file is part of Adblock Plus <https://adblockplus.org/>,
  * Copyright (C) 2006-present eyeo GmbH
  *
  * Adblock Plus is free software: you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 3 as
  * published by the Free Software Foundation.
  *
  * Adblock Plus is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
@@ skipping 99 matching lines @@
   private:
     explicit HashContainer(const HashContainer& other);
     void operator=(const HashContainer& other);
 
   protected:
     static constexpr size_type MIN_BUCKETS = 1;
     static constexpr double LOAD_FACTOR = 0.8;
     std::unique_ptr<entry_type[]> mBuckets;
     size_type mBucketCount;
     size_type mEntryCount;
+    size_type mExpectedEntries;

[Wladimir Palant, 2017/10/09 08:39:46]

No point recalculating bucket count if the map is cleared, mInitialBucketCount
will do.

 
 #if defined(DEBUG)
     size_type mInsertCounter;
 #endif
 
     explicit HashContainer(size_type expectedEntries = 0)
-        : mEntryCount(0)
-    {
-      expectedEntries = ceil(expectedEntries / LOAD_FACTOR);
-      mBucketCount = MIN_BUCKETS;
-      while (mBucketCount < expectedEntries)
-        mBucketCount <<= 1;
-
-      mBuckets.reset(new entry_type[mBucketCount]);
-      // Working around https://github.com/waywardmonkeys/emscripten-trace-collector/issues/2 here
-      annotate_address(reinterpret_cast<size_type*>(mBuckets.get()) - 1, "Hash table buffer");
+        : mEntryCount(0), mExpectedEntries(expectedEntries)
+    {
+      init(expectedEntries);
     }
 
     static size_type hash(const String& str)
     {
       // FNV-1a hash function
       size_type result = 2166136261;
       for (String::size_type i = 0; i < str.length(); i++)
         result = (result ^ str[i]) * 16777619;
       return result;
     }
@@ skipping 87 matching lines @@
       return const_iterator(&mBuckets[mBucketCount], &mBuckets[mBucketCount]);
     }
 
     size_type size() const
     {
       return mEntryCount;
     }
 
     void clear()
     {
-      resize(0);

[sergei, 2017/10/02 12:02:33]

I'm not sure that `resize` expects a value, lower than the previous size, and
`find_bucket` does not expected `mBucketCount` to have zero value.

[sergei, 2017/10/04 08:54:31]

This is not addressed.

[sergei, 2017/10/06 09:33:56]

What about creation of an `init(size_type expectedEntries)` method and call it
from both the contructor and from `clear`?

[hub, 2017/10/06 13:49:16]

Done.

[hub, 2017/10/06 13:49:16]

that's the approach I took.

+      mEntryCount = 0;
+      init(mExpectedEntries);

[Wladimir Palant, 2017/10/09 08:39:45]

IMHO, the shared function between constructor and clear() shouldn't involve
calculating the bucket count - merely assigning mBucketCount and reallocation.
So here it should be:

  mEntryCount = 0;
  allocate(mInitialBucketCount);

Same function can be called in resize() then.

+    }
+
+  private:
+    void init(size_type expectedEntries)
+    {
+      expectedEntries = ceil(expectedEntries / LOAD_FACTOR);
+      mBucketCount = MIN_BUCKETS;
+      while (mBucketCount < expectedEntries)
+        mBucketCount <<= 1;
+
+      mBuckets.reset(new entry_type[mBucketCount]);
+      // Working around https://github.com/waywardmonkeys/emscripten-trace-collector/issues/2 here
+      annotate_address(reinterpret_cast<size_type*>(mBuckets.get()) - 1, "Hash table buffer");
     }
   };
 
   struct StringSetEntry
   {
     StringSetEntry() {}
     StringSetEntry(const String& key)
         : first(key)
     {
     }
@@ skipping 99 matching lines @@
   const_reference find(const String& key) const
   {
     return super::find(key);
   }
 
   reference find(const String& key)
   {
     return reference(this, key, super::find_bucket(key));
   }
 };