Do not treat combining marks as whitespace (#19498)

* Do not treat spacing combining marks as whitespace

* Do not treat combining marks as whitespace
Try to find a matching typeface on a grapheme boundary

* Add embedded font for testing

* Remove using

* Fix naming and remove redundant code

* Apply default until we reach the next base character

src/Avalonia.Base/Media/TextFormatting/TextCharacters.cs

@@ -115,21 +115,13 @@ namespace Avalonia.Media.TextFormatting
 
             var codepoint = Codepoint.ReplacementCodepoint;
 
-            var codepointEnumerator = new CodepointEnumerator(text.Slice(count).Span);
+            var graphemeEnumerator = new GraphemeEnumerator(text.Slice(count).Span);
 
-            while (codepointEnumerator.MoveNext(out var cp))
+            if (graphemeEnumerator.MoveNext(out var grapheme))
             {
-                if (cp.IsWhiteSpace)
-                {
-                    continue;
-                }
-
-                codepoint = cp;
-
-                break;
+                codepoint = grapheme.FirstCodepoint;
             }
 
-            //ToDo: Fix FontFamily fallback
             var matchFound =
                 fontManager.TryMatchCharacter(codepoint, defaultTypeface.Style, defaultTypeface.Weight,
                     defaultTypeface.Stretch, defaultTypeface.FontFamily, defaultProperties.CultureInfo,
@@ -151,7 +143,8 @@ namespace Avalonia.Media.TextFormatting
             // no fallback found
             var enumerator = new GraphemeEnumerator(textSpan);
 
-            while (enumerator.MoveNext(out var grapheme))
+            //Move forward until we reach the next base character
+            while (enumerator.MoveNext(out grapheme))
             {
                 if (!grapheme.FirstCodepoint.IsWhiteSpace && defaultGlyphTypeface.TryGetGlyph(grapheme.FirstCodepoint, out _))
                 {

src/Avalonia.Base/Media/TextFormatting/Unicode/Codepoint.cs

@@ -128,10 +128,8 @@ namespace Avalonia.Media.TextFormatting.Unicode
             {
                 const ulong whiteSpaceMask =
                     (1UL << (int)GeneralCategory.Control) |
-                    (1UL << (int)GeneralCategory.NonspacingMark) |
                     (1UL << (int)GeneralCategory.Format) |
-                    (1UL << (int)GeneralCategory.SpaceSeparator) |
-                    (1UL << (int)GeneralCategory.SpacingMark);
+                    (1UL << (int)GeneralCategory.SpaceSeparator);
 
                 return ((1UL << (int)GeneralCategory) & whiteSpaceMask) != 0UL;
             }

tests/Avalonia.Skia.UnitTests/Media/TextFormatting/TextFormatterTests.cs

@@ -1088,6 +1088,32 @@ namespace Avalonia.Skia.UnitTests.Media.TextFormatting
             }
         }
 
+        [Fact]
+        public void Should_MatchCharacter_For_Spacing_CombiningMark()
+        {
+            using (Start())
+            {
+                var text = "𖾇";
+
+                var typeface = new Typeface(new FontFamily(new Uri("resm:Avalonia.Skia.UnitTests.Fonts?assembly=Avalonia.Skia.UnitTests"), "Noto Mono"));
+                var defaultRunProperties = new GenericTextRunProperties(typeface);
+                var paragraphProperties = new GenericTextParagraphProperties(defaultRunProperties, textWrapping: TextWrapping.Wrap);
+                var textLine = TextFormatter.Current.FormatLine(new SimpleTextSource(text, defaultRunProperties), 0, 120, paragraphProperties);
+
+                Assert.NotNull(textLine);
+
+                var textRuns = textLine.TextRuns;
+
+                Assert.NotEmpty(textRuns);
+
+                var firstRun = textRuns[0];
+
+                Assert.NotNull(firstRun.Properties);
+
+                Assert.Equal("Noto Sans Miao", firstRun.Properties.Typeface.GlyphTypeface.FamilyName);
+            }
+        }
+
         protected readonly record struct SimpleTextSource : ITextSource
         {
             private readonly string _text;