Merge pull request #6 from Cysharp/hadashiA/fix-multibyte-alignment

Make sure string alignment calculate as char count
Cysharp · Nov 16, 2023 · 1497912 · 1497912
2 parents 8f4b9f1 + 0b3c0e7
commit 1497912
Show file tree

Hide file tree

Showing 3 changed files with 35 additions and 3 deletions.
diff --git a/src/Utf8StringInterpolation/Utf8String.AppendFormatted.cs b/src/Utf8StringInterpolation/Utf8String.AppendFormatted.cs
@@ -1,6 +1,7 @@
 using System.Buffers;
 using System.Buffers.Text;
 using System.Runtime.CompilerServices;
+using System.Text;
 
 namespace Utf8StringInterpolation;
 

diff --git a/src/Utf8StringInterpolation/Utf8StringWriter.cs b/src/Utf8StringInterpolation/Utf8StringWriter.cs
@@ -179,7 +179,19 @@ public void AppendFormatted(string value, int alignment = 0, string? format = nu
             var buffer = rentArray.AsSpan();
             var bytesWritten = Encoding.UTF8.GetBytes(value.AsSpan(), buffer);
 
-            var space = alignment - bytesWritten;
+            int charCount;
+#if NETSTANDARD2_0
+            unsafe
+            {
+                fixed (byte* ptr = &buffer[0])
+                {
+                    charCount = Encoding.UTF8.GetCharCount(ptr, bytesWritten);
+                }
+            }
+#else            
+            charCount = Encoding.UTF8.GetCharCount(buffer.Slice(0, bytesWritten));
+#endif
+            var space = alignment - charCount;
             if (space > 0)
             {
                 AppendWhitespace(space);
@@ -190,18 +202,30 @@ public void AppendFormatted(string value, int alignment = 0, string? format = nu
             destination = destination.Slice(bytesWritten);
             currentWritten += bytesWritten;
             ArrayPool<byte>.Shared.Return(rentArray);
-            return;
         }
         else
         {
             // add right whitespace
             var max = GetStringByteCount(value.AsSpan());
             TryGrow(max);
             var bytesWritten = Encoding.UTF8.GetBytes(value.AsSpan(), destination);
+
+            int charCount;
+#if NETSTANDARD2_0
+            unsafe
+            {
+                fixed (byte* ptr = &destination[currentWritten])
+                {
+                    charCount = Encoding.UTF8.GetCharCount(ptr, bytesWritten);
+                }
+            }
+#else
+            charCount = Encoding.UTF8.GetCharCount(destination.Slice(0, bytesWritten));
+#endif
             destination = destination.Slice(bytesWritten);
             currentWritten += bytesWritten;
 
-            var space = bytesWritten + alignment;
+            var space = charCount + alignment;
             if (space < 0)
             {
                 AppendWhitespace(-space);

diff --git a/tests/Utf8StringInterpolation.Tests/FormatTest.cs b/tests/Utf8StringInterpolation.Tests/FormatTest.cs
@@ -86,5 +86,12 @@ public void Unicode()
             Utf8String.Format($"\u30cf\u30fc\u30c8: {"\u2764"}, \u5bb6\u65cf: {"\uD83D\uDC69\u200D\uD83D\uDC69\u200D\uD83D\uDC67\u200D\uD83D\uDC67"}(\u7d75\u6587\u5b57)")
                  .Should().Be($"\u30cf\u30fc\u30c8: {"\u2764"}, \u5bb6\u65cf: {"\uD83D\uDC69\u200D\uD83D\uDC69\u200D\uD83D\uDC67\u200D\uD83D\uDC67"}(\u7d75\u6587\u5b57)");
         }
+
+        [Fact]
+        public void MultibyteStringAlignment()
+        {
+            Utf8String.Format($"abc{"あいう",10}").Should().Be($"abc{"あいう",10}");
+            Utf8String.Format($"def{"えおか",-10}").Should().Be($"def{"えおか",-10}");
+        }
     }
 }