From d935e45b5b0f688ef707b8bb9f46ddfd7217e3d2 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jakub=20M=C3=AD=C5=A1ek?= <jmisek@gmail.com>
Date: Thu, 15 Aug 2024 19:09:39 +0200
Subject: [PATCH] GetChars(Encoding,byte[],StringBuilder) helper

- encode ReadOnlySpan<byte> without allocs into StringBuilder
+ tests
---
 .../Utilities/EncodingExtensions.cs           | 73 +++++++++++++++++++
 .../Peachpie.Runtime.Tests/UtilitiesTests.cs  | 31 +++++++-
 2 files changed, 103 insertions(+), 1 deletion(-)
 create mode 100644 src/Peachpie.Runtime/Utilities/EncodingExtensions.cs

diff --git a/src/Peachpie.Runtime/Utilities/EncodingExtensions.cs b/src/Peachpie.Runtime/Utilities/EncodingExtensions.cs
new file mode 100644
index 0000000000..ae4e79400b
--- /dev/null
+++ b/src/Peachpie.Runtime/Utilities/EncodingExtensions.cs
@@ -0,0 +1,73 @@
+using System;
+using System.Buffers;
+using System.Collections.Generic;
+using System.Linq;
+using System.Runtime.InteropServices;
+using System.Text;
+using System.Threading.Tasks;
+
+namespace Pchp.Core.Utilities
+{
+    public static class EncodingExtensions
+    {
+        const int MaxBytesAtOnce = 1 * 1024 * 1024; // 1M
+
+        /// <summary>
+        /// Encodes byte array using <paramref name="encoding"/> into given <paramref name="builder"/>.
+        /// </summary>
+        /// <returns>Number of characters encoded.</returns>
+        public static int GetChars(this Encoding encoding, ReadOnlySpan<byte> bytes, StringBuilder builder)
+        {
+            ArgumentNullException.ThrowIfNull(encoding, nameof(encoding));
+            ArgumentNullException.ThrowIfNull(builder, nameof(builder));
+
+            if (bytes.IsEmpty)
+            {
+                return 0;
+            }
+
+            if (bytes.Length <= MaxBytesAtOnce)
+            {
+
+                var maxCharCount = encoding.GetCharCount(bytes);
+
+                var span = ArrayPool<char>.Shared.Rent(maxCharCount);
+                var chars = encoding.GetChars(bytes, span.AsSpan());
+
+                //
+                builder.Append(span.AsSpan(0, chars));
+
+                //
+                ArrayPool<char>.Shared.Return(span);
+
+                return chars;
+            }
+
+            //
+
+            var decoder = encoding.GetDecoder(); // we need to encode in chunks, preserve state between chunks
+            int charsCount = 0;
+
+            while (bytes.Length > 0)
+            {
+                var segment = bytes.Slice(0, Math.Min(MaxBytesAtOnce, bytes.Length));
+
+                bytes = bytes.Slice(segment.Length);
+
+                var maxCharCount = decoder.GetCharCount(segment, flush: bytes.IsEmpty);
+                var span = ArrayPool<char>.Shared.Rent(maxCharCount);
+                var chars = decoder.GetChars(segment, span.AsSpan(), flush: bytes.IsEmpty);
+
+                //
+                builder.Append(span.AsSpan(0, chars));
+                charsCount += chars;
+
+                //
+                ArrayPool<char>.Shared.Return(span);
+            }
+
+            //
+            return charsCount;
+        }
+    }
+}
diff --git a/src/Tests/Peachpie.Runtime.Tests/UtilitiesTests.cs b/src/Tests/Peachpie.Runtime.Tests/UtilitiesTests.cs
index 56e57a698a..412511e45b 100644
--- a/src/Tests/Peachpie.Runtime.Tests/UtilitiesTests.cs
+++ b/src/Tests/Peachpie.Runtime.Tests/UtilitiesTests.cs
@@ -1,9 +1,10 @@
-using System;
+using System;
 using System.Text;
 using Microsoft.VisualStudio.TestTools.UnitTesting;
 using Pchp.Core;
 using Pchp.Core.Collections;
 using Pchp.Core.Text;
+using Pchp.Core.Utilities;
 
 namespace Peachpie.Runtime.Tests
 {
@@ -63,5 +64,33 @@ public void ValueListToBytesTest()
 
             Assert.AreEqual(Encoding.UTF8.GetString(list.ToArray()), "hello");
         }
+
+        [DataTestMethod]
+        [DataRow("")]
+        [DataRow("lorem ipsum")]
+        [DataRow("顧客は非常に重要です、顧客は顧客に続きます")]
+        public void GetCharsTest(string input)
+        {
+            string value = input;
+
+            for (int multiplier = 0; multiplier < 10; multiplier++)
+            {
+                var encoding = Encoding.UTF8;
+                var bytes = encoding.GetBytes(value);
+
+                var builder = new StringBuilder();
+                var count = Pchp.Core.Utilities.EncodingExtensions.GetChars(encoding, bytes, builder);
+
+                Assert.AreEqual(value.Length, count, "Length don't match");
+                Assert.AreEqual(value, builder.ToString(), "String don't match");
+
+                // add ~1M chars
+                for (int i = 0; i < 1_000_000 / (input.Length + 1); i++)
+                {
+                    builder.Append(input);
+                }
+                value = builder.ToString();
+            }
+        }
     }
 }