From 7baa18b5dd9d93c6eff9f956d3747388363779a2 Mon Sep 17 00:00:00 2001
From: Eliot Jones <elioty@hotmail.co.uk>
Date: Sat, 4 Apr 2020 18:31:55 +0100
Subject: [PATCH] add stringbuilder pool for tokenizers

we could replace these with spans in the next net core however for now our pools seem to increase performance by reducing gc load.
---
 .../NumericTokenizer.cs                       |  5 +-
 .../PlainTokenizer.cs                         |  6 +-
 .../StringBuilderPool.cs                      | 65 +++++++++++++++++++
 .../StringTokenizer.cs                        |  5 +-
 4 files changed, 77 insertions(+), 4 deletions(-)
 create mode 100644 src/UglyToad.PdfPig.Tokenization/StringBuilderPool.cs

diff --git a/src/UglyToad.PdfPig.Tokenization/NumericTokenizer.cs b/src/UglyToad.PdfPig.Tokenization/NumericTokenizer.cs
index c6b75a4f..e5c56fc8 100644
--- a/src/UglyToad.PdfPig.Tokenization/NumericTokenizer.cs
+++ b/src/UglyToad.PdfPig.Tokenization/NumericTokenizer.cs
@@ -8,6 +8,8 @@
 
     internal class NumericTokenizer : ITokenizer
     {
+        private static readonly StringBuilderPool StringBuilderPool = new StringBuilderPool(10);
+
         private const byte Zero = 48;
         private const byte Nine = 57;
 
@@ -21,7 +23,7 @@
 
             if ((currentByte >= Zero && currentByte <= Nine) || currentByte == '-' || currentByte == '+' || currentByte == '.')
             {
-                characters = new StringBuilder();
+                characters = StringBuilderPool.Borrow();
                 characters.Append((char)currentByte);
             }
             else
@@ -51,6 +53,7 @@
             try
             {
                 var str = characters.ToString();
+                StringBuilderPool.Return(characters);
 
                 switch (str)
                 {
diff --git a/src/UglyToad.PdfPig.Tokenization/PlainTokenizer.cs b/src/UglyToad.PdfPig.Tokenization/PlainTokenizer.cs
index 5d1f8486..8ef87512 100644
--- a/src/UglyToad.PdfPig.Tokenization/PlainTokenizer.cs
+++ b/src/UglyToad.PdfPig.Tokenization/PlainTokenizer.cs
@@ -1,11 +1,12 @@
 namespace UglyToad.PdfPig.Tokenization
 {
-    using System.Text;
     using Core;
     using Tokens;
 
     internal class PlainTokenizer : ITokenizer
     {
+        private static readonly StringBuilderPool StringBuilderPool = new StringBuilderPool(10);
+
         public bool ReadsNextByte { get; } = true;
 
         public bool TryTokenize(byte currentByte, IInputBytes inputBytes, out IToken token)
@@ -17,7 +18,7 @@
                 return false;
             }
 
-            var builder = new StringBuilder();
+            var builder = StringBuilderPool.Borrow();
             builder.Append((char)currentByte);
             while (inputBytes.MoveNext())
             {
@@ -38,6 +39,7 @@
             }
 
             var text = builder.ToString();
+            StringBuilderPool.Return(builder);
 
             switch (text)
             {
diff --git a/src/UglyToad.PdfPig.Tokenization/StringBuilderPool.cs b/src/UglyToad.PdfPig.Tokenization/StringBuilderPool.cs
new file mode 100644
index 00000000..7ff0ba92
--- /dev/null
+++ b/src/UglyToad.PdfPig.Tokenization/StringBuilderPool.cs
@@ -0,0 +1,65 @@
+namespace UglyToad.PdfPig.Tokenization
+{
+    using System.Collections.Generic;
+    using System.Text;
+
+    /// <summary>
+    /// A pool for <see cref="StringBuilder"/>s to reduce allocations during tokenization.
+    /// </summary>
+    public class StringBuilderPool
+    {
+        private readonly int capacity;
+        private readonly object locker = new object();
+        private readonly Stack<StringBuilder> pool = new Stack<StringBuilder>();
+
+        /// <summary>
+        /// Create a new <see cref="StringBuilderPool"/> holding the number of items specified by the capacity.
+        /// </summary>
+        public StringBuilderPool(int capacity = 5)
+        {
+            this.capacity = capacity;
+
+            for (var i = 0; i < capacity; i++)
+            {
+                pool.Push(new StringBuilder());
+            }
+        }
+
+        /// <summary>
+        /// Get an item from the pool, remember to return it using <see cref="Return"/> at the end.
+        /// </summary>
+        public StringBuilder Borrow()
+        {
+            lock (locker)
+            {
+                if (pool.Count == 0)
+                {
+                    return new StringBuilder();
+                }
+
+                return pool.Pop();
+            }
+        }
+
+        /// <summary>
+        /// Returns an item to the pool of available builders.
+        /// </summary>
+        public void Return(StringBuilder instance)
+        {
+            if (instance == null)
+            {
+                return;
+            }
+
+            instance.Clear();
+
+            lock (locker)
+            {
+                if (pool.Count < capacity)
+                {
+                    pool.Push(instance);
+                }
+            }
+        }
+    }
+}
diff --git a/src/UglyToad.PdfPig.Tokenization/StringTokenizer.cs b/src/UglyToad.PdfPig.Tokenization/StringTokenizer.cs
index d701df67..bc68e919 100644
--- a/src/UglyToad.PdfPig.Tokenization/StringTokenizer.cs
+++ b/src/UglyToad.PdfPig.Tokenization/StringTokenizer.cs
@@ -6,11 +6,11 @@
 
     internal class StringTokenizer : ITokenizer
     {
+        private static readonly StringBuilderPool StringBuilderPool = new StringBuilderPool(16);
         public bool ReadsNextByte { get; } = false;
 
         public bool TryTokenize(byte currentByte, IInputBytes inputBytes, out IToken token)
         {
-            var builder = new StringBuilder();
             token = null;
 
             if (inputBytes == null)
@@ -23,6 +23,7 @@
                 return false;
             }
 
+            var builder = StringBuilderPool.Borrow();
             var numberOfBrackets = 1;
             var isEscapeActive = false;
             var isLineBreaking = false;
@@ -177,6 +178,8 @@
                 encodedWith = StringToken.Encoding.Iso88591;
             }
 
+            StringBuilderPool.Return(builder);
+
             token = new StringToken(tokenStr, encodedWith);
 
             return true;