Add early exit gap detection

Author: Genbox

AGENTS.md

@@ -27,6 +27,7 @@
 - Scripted build: `pwsh Scripts/Build.ps1`.
 - Test full solution: `dotnet test FastData.slnx -c Debug`.
 - Test a single project: `dotnet test Src/FastData.Tests/FastData.Tests.csproj -c Debug`.
+- Test harness runner (uses real generators): `dotnet test --project Src/FastData.TestHarness.Runner/FastData.TestHarness.Runner.csproj -c Debug`.
 - Test project list:
   - `Src/FastData.Tests/FastData.Tests.csproj`
   - `Src/FastData.Cli.Tests/FastData.Cli.Tests.csproj`

Src/FastData.Benchmarks/Benchmarks/GPerfAnalyzerBenchmarks.cs

@@ -19,8 +19,8 @@ public GPerfAnalyzerBenchmarks()
         Random rng = new Random(42);
         _data = Enumerable.Range(1, 100).Select(_ => TestHelper.GenerateRandomString(rng, 50)).ToArray();
 
-        StringKeyProperties props = KeyAnalyzer.GetStringProperties(_data, false, false, GeneratorEncoding.ASCII);
-        _analyzer = new GPerfAnalyzer(_data.Length, props, new GPerfAnalyzerConfig(), new Simulator(_data.Length, GeneratorEncoding.UTF16), NullLogger<GPerfAnalyzer>.Instance);
+        StringKeyProperties props = KeyAnalyzer.GetStringProperties(_data, false, false, GeneratorEncoding.Utf16CodeUnits);
+        _analyzer = new GPerfAnalyzer(_data.Length, props, new GPerfAnalyzerConfig(), new Simulator(_data.Length, GeneratorEncoding.Utf16CodeUnits), NullLogger<GPerfAnalyzer>.Instance);
     }
 
     [Benchmark]

Src/FastData.Benchmarks/Benchmarks/KeyAnalyzerBenchmarks.cs

@@ -16,5 +16,5 @@ public KeyAnalyzerBenchmarks()
     }
 
     [Benchmark]
-    public object GetStringProperties() => KeyAnalyzer.GetStringProperties(_data, true, false, GeneratorEncoding.UTF16);
+    public object GetStringProperties() => KeyAnalyzer.GetStringProperties(_data, true, false, GeneratorEncoding.Utf16CodeUnits);
 }

Src/FastData.Benchmarks/Benchmarks/SegmentGeneratorsBenchmarks.cs

@@ -15,7 +15,7 @@ public class SegmentGeneratorsBenchmarks
     private readonly OffsetGenerator _ofGen = new OffsetGenerator();
 
     //We start at 8 and go up to 100 to cover as many cases as possible
-    private readonly StringKeyProperties _props = KeyAnalyzer.GetStringProperties(Enumerable.Range(8, 100).Select(x => TestHelper.GenerateRandomString(Random.Shared, x)).ToArray(), false, false, GeneratorEncoding.UTF16);
+    private readonly StringKeyProperties _props = KeyAnalyzer.GetStringProperties(Enumerable.Range(8, 100).Select(x => TestHelper.GenerateRandomString(Random.Shared, x)).ToArray(), false, false, GeneratorEncoding.Utf16CodeUnits);
 
     [Benchmark]
     public object BruteForceGenerator() => _bfGen.Generate(_props).ToArray();

Src/FastData.Generator.CPlusPlus.TestHarness/CPlusPlusBootstrap.cs

@@ -11,7 +11,7 @@ public sealed class CPlusPlusBootstrap : BootstrapBase
     public CPlusPlusBootstrap(HarnessType type) : base("CPlusPlus", ".cpp", type, "silkeh/clang", GetCommandTemplate(type))
     {
         CPlusPlusLanguageDef langDef = new CPlusPlusLanguageDef();
-        Map = new TypeMap(langDef.TypeDefinitions, GeneratorEncoding.UTF8);
+        Map = new TypeMap(langDef.TypeDefinitions, GeneratorEncoding.Utf8Bytes);
     }
 
     public TypeMap Map { get; }

Src/FastData.Generator.CPlusPlus/CPlusPlusCodeGenerator.cs

@@ -5,7 +5,7 @@
 
 namespace Genbox.FastData.Generator.CPlusPlus;
 
-public sealed class CPlusPlusCodeGenerator(CPlusPlusCodeGeneratorConfig cppCfg) : TemplatedCodeGenerator(new CPlusPlusLanguageDef(), GeneratorEncoding.UTF8)
+public sealed class CPlusPlusCodeGenerator(CPlusPlusCodeGeneratorConfig cppCfg) : TemplatedCodeGenerator(new CPlusPlusLanguageDef(), GeneratorEncoding.Utf8Bytes)
 {
     protected override string GenerateTemplated<TKey, TValue>(GeneratorConfigBase genCfg, TemplateManager manager, Dictionary<string, object?> variables)
     {

Src/FastData.Generator.CPlusPlus/Internal/CPlusPlusLanguageDef.cs

@@ -25,12 +25,10 @@ internal class CPlusPlusLanguageDef : ILanguageDef
 
         new ObjectTypeDef(PrintDeclaration, PrintValue),
 
-        //Support reduction from UTF16 to ASCII
         new DynamicStringTypeDef(
-            new StringType(GeneratorEncoding.UTF32, "std::u32string_view", static x => $"U\"{x}\""),
-            new StringType(GeneratorEncoding.UTF16, "std::u16string_view", static x => $"u\"{x}\""),
-            new StringType(GeneratorEncoding.UTF8, "std::string_view", static x => $"u8\"{x}\""),
-            new StringType(GeneratorEncoding.ASCII, "std::string_view", static x => $"\"{x}\""))
+            new StringType(GeneratorEncoding.Utf16CodeUnits, "std::u16string_view", static x => $"u\"{x}\""),
+            new StringType(GeneratorEncoding.Utf8Bytes, "std::string_view", static x => $"u8\"{x}\""),
+            new StringType(GeneratorEncoding.AsciiBytes, "std::string_view", static x => $"\"{x}\""))
     };
 
     private static string PrintDeclaration(TypeMap map, Type type)

Src/FastData.Generator.CPlusPlus/Templates/Includes/Footer.ttinclude

@@ -4,8 +4,8 @@
     if (GeneratorConfig is StringGeneratorConfig strCfg)
     {
 #>
-    static constexpr size_t min_key_length = <#= strCfg.Constants.MinStringLength #>;
-    static constexpr size_t max_key_length = <#= strCfg.Constants.MaxStringLength #>;
+    static constexpr int32_t min_key_length = <#= strCfg.Constants.MinStringLength #>;
+    static constexpr int32_t max_key_length = <#= strCfg.Constants.MaxStringLength #>;
     static constexpr bool has_number = <#= strCfg.Constants.CharacterClasses.HasFlag(CharacterClass.Number) ? "true" : "false" #>;
     static constexpr bool has_uppercase = <#= strCfg.Constants.CharacterClasses.HasFlag(CharacterClass.Uppercase) ? "true" : "false" #>;
     static constexpr bool has_lowercase = <#= strCfg.Constants.CharacterClasses.HasFlag(CharacterClass.Lowercase) ? "true" : "false" #>;

Src/FastData.Generator.CPlusPlus/Templates/Includes/Functions.ttinclude

@@ -116,7 +116,7 @@
         {
             if (keyType == TypeCode.Single)
             {
-                return numCfg.HasZeroOrNaN
+                return numCfg.HasZero
                     ? """
                                 uint32_t bits;
                                 std::memcpy(&bits, &value, sizeof(bits));
@@ -133,7 +133,7 @@
 
             if (keyType == TypeCode.Double)
             {
-                return numCfg.HasZeroOrNaN
+                return numCfg.HasZero
                     ? """
                                 uint64_t bits;
                                 std::memcpy(&bits, &value, sizeof(bits));

Src/FastData.Generator.CPlusPlus/Templates/Includes/Header.ttinclude

@@ -9,6 +9,7 @@
 <#@ import namespace="Genbox.FastData.Generator.Extensions" #>
 <#@ import namespace="Genbox.FastData.Generator.Helpers" #>
 <#@ import namespace="Genbox.FastData.Generators" #>
+<#@ import namespace="Genbox.FastData.Generators.Helpers" #>
 <#@ import namespace="Genbox.FastData.Generators.Abstracts" #>
 <#@ import namespace="Genbox.FastData.Generators.EarlyExits" #>
 <#@ import namespace="Genbox.FastData.Generators.Enums" #>
@@ -100,7 +101,7 @@ class <#= CPlusPlusConfig.ClassName #> final {
             { StringFunction.GetLastChar, "constexpr char GetLastChar(std::string_view str) { return str[str.length() - 1]; }" },
             { StringFunction.GetLastCharLower, "constexpr char GetLastCharLower(std::string_view str) { return ToLowerAscii(GetLastChar(str)); }" },
 
-            { StringFunction.GetLength, "constexpr uint32_t GetLength(std::string_view str) { return static_cast<uint32_t>(str.length()); }" },
+            { StringFunction.GetLength, "constexpr int32_t GetLength(std::string_view str) { return static_cast<int32_t>(str.length()); }" },
 
             { StringFunction.StartsWith, """
     constexpr bool StartsWith(std::string_view prefix, std::string_view str) {