#78 - переписал SG на полностью автоматическую генерацию паттерна рег…

…улярки
Stepami · Sep 20, 2024 · c12be00 · c12be00
1 parent c1bd557
commit c12be00
Show file tree

Hide file tree

Showing 8 changed files with 73 additions and 111 deletions.
diff --git a/...ure/HydraScript.Infrastructure.LexerRegexGenerator/DefaultTokenTypesJsonStringProvider.cs b/...ure/HydraScript.Infrastructure.LexerRegexGenerator/DefaultTokenTypesJsonStringProvider.cs
@@ -0,0 +1,9 @@
+using HydraScript.Domain.FrontEnd.Lexer;
+
+namespace HydraScript.Infrastructure.LexerRegexGenerator;
+
+internal class DefaultTokenTypesJsonStringProvider :
+    ITokenTypesJsonStringProvider
+{
+    public string TokenTypesJsonString => TokenTypesJson.String;
+}
diff --git a/....Infrastructure.LexerRegexGenerator/HydraScript.Infrastructure.LexerRegexGenerator.csproj b/....Infrastructure.LexerRegexGenerator/HydraScript.Infrastructure.LexerRegexGenerator.csproj
@@ -14,4 +14,8 @@
         <PackageReference Include="Microsoft.CodeAnalysis.CSharp.Workspaces" />
     </ItemGroup>
 
+    <ItemGroup>
+      <ProjectReference Include="..\..\Domain\HydraScript.Domain.FrontEnd\HydraScript.Domain.FrontEnd.csproj" />
+    </ItemGroup>
+
 </Project>
diff --git a/...structure/HydraScript.Infrastructure.LexerRegexGenerator/ITokenTypesJsonStringProvider.cs b/...structure/HydraScript.Infrastructure.LexerRegexGenerator/ITokenTypesJsonStringProvider.cs
@@ -0,0 +1,6 @@
+namespace HydraScript.Infrastructure.LexerRegexGenerator;
+
+public interface ITokenTypesJsonStringProvider
+{
+    public string TokenTypesJsonString { get; }
+}
diff --git a/src/Infrastructure/HydraScript.Infrastructure.LexerRegexGenerator/PatternGenerator.cs b/src/Infrastructure/HydraScript.Infrastructure.LexerRegexGenerator/PatternGenerator.cs
@@ -1,96 +1,42 @@
-using System.Collections.Immutable;
 using System.Text;
 using System.Text.Json;
 using Microsoft.CodeAnalysis;
-using Microsoft.CodeAnalysis.CSharp;
-using Microsoft.CodeAnalysis.CSharp.Syntax;
 using Microsoft.CodeAnalysis.Text;
 
 namespace HydraScript.Infrastructure.LexerRegexGenerator;
 
 [Generator]
 public partial class PatternGenerator : IIncrementalGenerator
 {
-    private const string AttributeSourceCode = @"// <auto-generated/>
-
-namespace HydraScript.Infrastructure;
-
-[System.AttributeUsage(System.AttributeTargets.Class)]
-public class PatternContainerAttribute<T>(string json) : System.Attribute
-    where T : HydraScript.Domain.FrontEnd.Lexer.IGeneratedRegexContainer
-{
-    public string Json { get; } = json;
-}
-";
+    public ITokenTypesJsonStringProvider Provider { get; init; } = new DefaultTokenTypesJsonStringProvider();
 
     public void Initialize(IncrementalGeneratorInitializationContext context)
     {
-        context.RegisterPostInitializationOutput(ctx => ctx.AddSource(
-            "PatternContainerAttribute.g.cs",
-            SourceText.From(AttributeSourceCode, Encoding.UTF8)));
-
-        var provider = context.SyntaxProvider
-            .ForAttributeWithMetadataName(
-                "HydraScript.Infrastructure.PatternContainerAttribute`1",
-                static (s, _) => IsSyntaxTargetForGeneration(s),
-                static (ctx, _) => GetTypeDeclarationForSourceGen(ctx))
-            .Where(static x => x is not null)
-            .Select(static (x, _) => x!);
-
-        context.RegisterImplementationSourceOutput(provider.Collect(), GenerateCode);
-    }
-
-    private static bool IsSyntaxTargetForGeneration(SyntaxNode node) =>
-        node is ClassDeclarationSyntax candidate &&
-        candidate.Modifiers.Any(SyntaxKind.PartialKeyword) &&
-        candidate.Modifiers.Any(SyntaxKind.InternalKeyword);
-
-    private static RegexContainerInfo? GetTypeDeclarationForSourceGen(
-        GeneratorAttributeSyntaxContext context)
-    {
-        var attribute = context.Attributes.FirstOrDefault();
-        if (attribute is null)
-            return null;
-        var visitable = (ClassDeclarationSyntax)context.TargetNode;
-        var json = attribute.ConstructorArguments.First().Value!.ToString()!;
-        return new RegexContainerInfo(
-            ClassName: visitable.Identifier.Text,
-            json);
-    }
+        var tokenTypes = JsonSerializer.Deserialize(
+                Provider.TokenTypesJsonString,
+                PatternGeneratorContext.Default.IEnumerableTokenType)!
+            .OrderBy(x => x.Priority)
+            .Concat([new TokenType("ERROR", @"\S+", int.MaxValue)]);
+        var pattern = string.Join('|', tokenTypes.Select(t => t.GetNamedRegex()));
 
-    private static void GenerateCode(
-        SourceProductionContext context,
-        ImmutableArray<RegexContainerInfo> containerInfos)
-    {
-        foreach (var info in containerInfos)
-        {
-            var tokenTypes = JsonSerializer.Deserialize(
-                    info.Json,
-                    PatternGeneratorContext.Default.IEnumerableTokenType)!
-                .OrderBy(x => x.Priority)
-                .Concat([new TokenType("ERROR", @"\S+", int.MaxValue)]);
-            var pattern = string.Join('|', tokenTypes.Select(t => t.GetNamedRegex()));
-
-            var code = $@"// <auto-generated/>
+        var code = $@"// <auto-generated/>
 
 using System.Diagnostics.CodeAnalysis;
 
 namespace HydraScript.Infrastructure;
 
-internal partial class {info.ClassName}
+internal partial class PatternContainer
 {{
     [StringSyntax(StringSyntaxAttribute.Regex)]
-    public const string Pattern =
+    public const string Value =
         """"""
             {pattern}
         """""";
 }}
 ";
-            context.AddSource($"{info.ClassName}.g.cs", SourceText.From(code, Encoding.UTF8));
-        }
-    }
 
-    private record RegexContainerInfo(
-        string ClassName,
-        string Json);
+        context.RegisterPostInitializationOutput(ctx => ctx.AddSource(
+            "PatternContainer.g.cs",
+            SourceText.From(code, Encoding.UTF8)));
+    }
 }
diff --git a/src/Infrastructure/HydraScript.Infrastructure/GeneratedRegexContainer.cs b/src/Infrastructure/HydraScript.Infrastructure/GeneratedRegexContainer.cs
@@ -3,9 +3,8 @@
 
 namespace HydraScript.Infrastructure;
 
-[PatternContainer<GeneratedRegexContainer>(TokenTypesJson.String)]
 internal partial class GeneratedRegexContainer : IGeneratedRegexContainer
 {
-    [GeneratedRegex("""(?<Comment>[/]{2}.*)|(?<FloatLiteral>[0-9]+[.][0-9]+)|(?<IntegerLiteral>[0-9]+)|(?<NullLiteral>null)|(?<BooleanLiteral>true|false)|(?<StringLiteral>\"(\\.|[^"\\])*\")|(?<Keyword>let|const|function|if|else|while|break|continue|return|as|type)|(?<Operator>[+]{1,2}|[-]|[*]|[/]|[%]|([!]|[=])[=]|([<]|[>])[=]?|[!]|[|]{2}|[&]{2}|[~]|[:]{2})|(?<Ident>[a-zA-Z][a-zA-Z0-9]*)|(?<QuestionMark>[?])|(?<Colon>[:])|(?<SemiColon>[;])|(?<Assign>[=])|(?<Comma>[,])|(?<LeftCurl>[{])|(?<RightCurl>[}])|(?<LeftParen>[(])|(?<RightParen>[)])|(?<Dot>[.])|(?<LeftBracket>[[])|(?<RightBracket>[]])|(?<ERROR>\S+)""")]
+    [GeneratedRegex(PatternContainer.Value)]
     public static partial Regex GetRegex();
 }
diff --git a/...ure.LexerRegexGenerator.Tests/HydraScript.Infrastructure.LexerRegexGenerator.Tests.csproj b/...ure.LexerRegexGenerator.Tests/HydraScript.Infrastructure.LexerRegexGenerator.Tests.csproj
@@ -8,6 +8,7 @@
             <IncludeAssets>runtime; build; native; contentfiles; analyzers; buildtransitive</IncludeAssets>
             <PrivateAssets>all</PrivateAssets>
         </PackageReference>
+        <PackageReference Include="NSubstitute" />
     </ItemGroup>
 
     <ItemGroup>

diff --git a/tests/HydraScript.Infrastructure.LexerRegexGenerator.Tests/PatternGeneratorTests.cs b/tests/HydraScript.Infrastructure.LexerRegexGenerator.Tests/PatternGeneratorTests.cs
@@ -1,62 +1,72 @@
 using System.Diagnostics;
+using System.Diagnostics.CodeAnalysis;
 using System.Reflection;
 using Microsoft.CodeAnalysis;
 using Microsoft.CodeAnalysis.CSharp;
+using NSubstitute;
 using Xunit;
 
 namespace HydraScript.Infrastructure.LexerRegexGenerator.Tests;
 
 public class PatternGeneratorTests
 {
+    [StringSyntax(StringSyntaxAttribute.Json)]
+    private const string JsonStringReplacement =
+        """
+        [
+            {
+                "tag": "Test2",
+                "pattern": "test2",
+                "priority": 2
+            },
+            {
+                "tag": "Test1",
+                "pattern": "test1",
+                "priority": 1
+            }
+        ]
+        """;
+
     [Fact]
     public void Initialize_PatternContainerMarked_CorrectlyGenerated()
     {
-        var inputCompilation = CreateCompilation(
-            """
-            using System.Text.RegularExpressions;
-            using HydraScript.Domain.FrontEnd.Lexer;
+        var provider = Substitute.For<ITokenTypesJsonStringProvider>();
+        provider.TokenTypesJsonString.Returns(JsonStringReplacement);
+        var generator = new PatternGenerator
+        {
+            Provider = provider
+        };
+        GeneratorDriver driver = CSharpGeneratorDriver.Create(generator);
 
-            namespace HydraScript.Infrastructure;
+        driver = driver.RunGeneratorsAndUpdateCompilation(CreateCompilation(string.Empty), out var outputCompilation,
+            out var diagnostics);
+        Debug.Assert(diagnostics.IsEmpty);
+        Debug.Assert(outputCompilation.SyntaxTrees.Count() == 2);
 
-            [PatternContainer<TestPatternContainer>("[{ \"tag\": \"Number\", \"pattern\": \"[0-9]+\", \"priority\": 2 }, { \"tag\": \"Word\", \"pattern\": \"[a-zA-Z]+\", \"priority\": 1 }]")]
-            internal partial class TestPatternContainer : IGeneratedRegexContainer
-            {
-                public static Regex GetRegex() => throw new NotImplementedException();
-            }
-            """);
+        var runResult = driver.GetRunResult();
+
+        var generatedFileSyntax = runResult.GeneratedTrees
+            .Single(t => t.FilePath.EndsWith("PatternContainer.g.cs"));
 
-        const string expectedSource =
+        const string expectedSource = 
 """"
 // <auto-generated/>
 
 using System.Diagnostics.CodeAnalysis;
 
 namespace HydraScript.Infrastructure;
 
-internal partial class TestPatternContainer
+internal partial class PatternContainer
 {
     [StringSyntax(StringSyntaxAttribute.Regex)]
-    public const string Pattern =
+    public const string Value =
         """
-            (?<Word>[a-zA-Z]+)|(?<Number>[0-9]+)|(?<ERROR>\S+)
+            (?<Test1>test1)|(?<Test2>test2)|(?<ERROR>\S+)
         """;
 }
 
 """";
 
-        var generator = new PatternGenerator();
-        GeneratorDriver driver = CSharpGeneratorDriver.Create(generator);
-
-        driver = driver.RunGeneratorsAndUpdateCompilation(inputCompilation, out var outputCompilation,
-            out var diagnostics);
-        Debug.Assert(diagnostics.IsEmpty);
-        Debug.Assert(outputCompilation.SyntaxTrees.Count() == 3);
-
-        var runResult = driver.GetRunResult();
-
-        var generatedFileSyntax = runResult.GeneratedTrees
-            .Single(t => t.FilePath.EndsWith("TestPatternContainer.g.cs"));
-
         Assert.Equal(
             expectedSource,
             generatedFileSyntax.GetText().ToString(),

diff --git a/tests/HydraScript.Tests/Unit/Infrastructure/GeneratedRegexContainerTests.cs b/tests/HydraScript.Tests/Unit/Infrastructure/GeneratedRegexContainerTests.cs