Add support for checksum-based interceptors

docs/features/interceptors.md

@@ -12,9 +12,9 @@ using System;
 using System.Runtime.CompilerServices;
 
 var c = new C();
-c.InterceptableMethod(1); // (L1,C1): prints "interceptor 1"
-c.InterceptableMethod(1); // (L2,C2): prints "other interceptor 1"
-c.InterceptableMethod(2); // (L3,C3): prints "other interceptor 2"
+c.InterceptableMethod(1); // L1: prints "interceptor 1"
+c.InterceptableMethod(1); // L2: prints "other interceptor 1"
+c.InterceptableMethod(2); // L3: prints "other interceptor 2"
 c.InterceptableMethod(1); // prints "interceptable 1"
 
 class C
@@ -28,14 +28,14 @@ class C
 // generated code
 static class D
 {
-    [InterceptsLocation("Program.cs", line: /*L1*/, character: /*C1*/)] // refers to the call at (L1, C1)
+    [InterceptsLocation(version: 1, data: "...(refers to the call at L1")]
     public static void InterceptorMethod(this C c, int param)
     {
         Console.WriteLine($"interceptor {param}");
     }
 
-    [InterceptsLocation("Program.cs", line: /*L2*/, character: /*C2*/)] // refers to the call at (L2, C2)
-    [InterceptsLocation("Program.cs", line: /*L3*/, character: /*C3*/)] // refers to the call at (L3, C3)
+    [InterceptsLocation(version: 1, data: "...(refers to the call at L2")]
+    [InterceptsLocation(version: 1, data: "...(refers to the call at L3")]
     public static void OtherInterceptorMethod(this C c, int param)
     {
         Console.WriteLine($"other interceptor {param}");
@@ -54,7 +54,7 @@ A method indicates that it is an *interceptor* by adding one or more `[Intercept
 namespace System.Runtime.CompilerServices
 {
     [AttributeUsage(AttributeTargets.Method, AllowMultiple = true)]
-    public sealed class InterceptsLocationAttribute(string filePath, int line, int character) : Attribute
+    public sealed class InterceptsLocationAttribute(int version, string data) : Attribute
     {
     }
 }
@@ -66,29 +66,28 @@ Any "ordinary method" (i.e. with `MethodKind.Ordinary`) can have its calls inter
 
 File-local declarations of this type (`file class InterceptsLocationAttribute`) are valid and usages are recognized by the compiler when they are within the same file and compilation. A generator which needs to declare this attribute should use a file-local declaration to ensure it doesn't conflict with other generators that need to do the same thing.
 
-#### File paths
+In prior experimental releases of the feature, a well-known constructor signature `InterceptsLocation(string path, int line, int column)]` was also supported. Support for this constructor will be **dropped** prior to stable release of the feature.
 
-The *referenced syntax tree* of an `[InterceptsLocation]` is determined by normalizing the `filePath` argument value relative to the path of the containing syntax tree of the `[InterceptsLocation]` usage, similar to how paths in `#line` directives are normalized. Let this normalized path be called `normalizedInterceptorPath`. If exactly one syntax tree in the compilation has a normalized path which matches `normalizedInterceptorPath` by ordinal string comparison, that is the *referenced syntax tree*. Otherwise, an error occurs.
+#### Location encoding
 
-`#line` directives are not considered when determining the call referenced by an `[InterceptsLocation]` attribute. In other words, the file path, line and column numbers used in `[InterceptsLocation]` are expected to refer to *unmapped* source locations.
+The arguments to `[InterceptsLocation]` are:
+1. a version number. The compiler may introduce new encodings for the location in the future, with corresponding new version numbers.
+2. an opaque data string. This is not intended to be human-readable.
 
-Temporarily, for compatibility purposes, when the initial matching strategy outlined above fails to match any syntax trees, we will fall back to a "compat" matching strategy which works in the following way:
-- A *mapped path* of each syntax tree is determined by applying [`/pathmap`](https://learn.microsoft.com/en-us/dotnet/api/microsoft.codeanalysis.commandlinearguments.pathmap?view=roslyn-dotnet-4.7.0) substitution to `SyntaxTree.FilePath`.
-- For a given `[InterceptsLocation]` usage, the `filePath` argument value is compared to the *mapped path* of each syntax tree using ordinal string comparison. If exactly one syntax tree matches under this comparison, that is the *referenced syntax tree*. Otherwise, an error occurs.
-
-Support for the "compat" strategy will be dropped prior to stable release. Tracked by https://github.com/dotnet/roslyn/issues/72265.
+The "version 1" data encoding is a base64-encoded string consisting of the following data:
+- 16 byte xxHash128 content checksum of the file containing the intercepted call.
+- int32 in little-endian format for the position (i.e. `SyntaxNode.Position`) of the call in syntax.
+- utf-8 string data containing a display file name, used for error reporting.
 
 #### Position
 
-Line and column numbers in `[InterceptsLocation]` are 1-indexed to match existing places where source locations are displayed to the user. For example, in `Diagnostic.ToString`.
+Line and column numbers in the "version 1" data encoding are 1-indexed to match existing places where source locations are displayed to the user. For example, in `Diagnostic.ToString`.
 
 The location of the call is the location of the simple name syntax which denotes the interceptable method. For example, in `app.MapGet(...)`, the name syntax for `MapGet` would be considered the location of the call. For a static method call like `System.Console.WriteLine(...)`, the name syntax for `WriteLine` is the location of the call. If we allow intercepting calls to property accessors in the future (e.g `obj.Property`), we would also be able to use the name syntax in this way.
 
 #### Attribute creation
 
-The goal of the above decisions is to make it so that when source generators are filling in `[InterceptsLocation(...)]`, they simply need to read `nameSyntax.SyntaxTree.FilePath` and `nameSyntax.GetLineSpan().Span.Start` for the exact file path and position information they need to use.
-
-We should provide samples of recommended coding patterns for generator authors to show correct usage of these, including the "translation" from 0-indexed to 1-indexed positions.
+Roslyn provides a convenience API, `GetInterceptableLocation(this SemanticModel, InvocationExpressionSyntax, CancellationToken)` for inserting `[InterceptsLocation]` into generated source code. We recommend that source generators depend on this API in order to intercept calls.
 
 ### Non-invocation method usages
 
@@ -103,7 +102,7 @@ Interceptors cannot be declared in generic types at any level of nesting.
 Interceptors must either be non-generic, or have arity equal to the sum of the arity of the original method's arity and containing type arities. For example:
 
 ```cs
-Grandparent<int>.Parent<bool>.Original<string>(1, false, "a");
+Grandparent<int>.Parent<bool>.Original<string>(1, false, "a"); // L1
 
 class Grandparent<T1>
 {
@@ -115,7 +114,7 @@ class Grandparent<T1>
 
 class Interceptors
 {
-    [InterceptsLocation("Program.cs", 1, 33)]
+    [InterceptsLocation(1, "..(refers to call at L1)")]
     public static void Interceptor<T1, T2, T3>(T1 t1, T2 t2, T3 t3) { }
 }
 ```
@@ -136,13 +135,13 @@ static class Program
 {
     public static void M<T2>(T2 t)
     {
-        C.InterceptableMethod(t);
+        C.InterceptableMethod(t); // L1
     }
 }
 
 static class D
 {
-    [InterceptsLocation("Program.cs", 12, 11)]
+    [InterceptsLocation(1, "..(refers to call at L1)")]
     public static void Interceptor1<T2>(T2 t) => throw null!;
 }
 ```

src/Compilers/CSharp/Portable/CSharpExtensions.cs

@@ -1638,6 +1638,26 @@ public static Conversion ClassifyConversion(this SemanticModel? semanticModel, i
             var csModel = semanticModel as CSharpSemanticModel;
             return csModel?.GetInterceptorMethod(node, cancellationToken);
         }
+
+        /// <summary>
+        /// If 'node' cannot be intercepted syntactically, returns null.
+        /// Otherwise, returns an instance which can be used to intercept the call denoted by <paramref name="node"/>.
+        /// </summary>
+        [Experimental(RoslynExperiments.Interceptors, UrlFormat = RoslynExperiments.Interceptors_Url)]
+        public static InterceptableLocation? GetInterceptableLocation(this SemanticModel? semanticModel, InvocationExpressionSyntax node, CancellationToken cancellationToken = default)
+        {
+            var csModel = semanticModel as CSharpSemanticModel;
+            return csModel?.GetInterceptableLocation(node, cancellationToken);
+        }
+
+        /// <summary>
+        /// Gets an attribute list syntax consisting of an InterceptsLocationAttribute, which intercepts the call referenced by parameter <paramref name="location"/>.
+        /// </summary>
+        [Experimental(RoslynExperiments.Interceptors, UrlFormat = RoslynExperiments.Interceptors_Url)]
+        public static string GetInterceptsLocationAttributeSyntax(this InterceptableLocation location)
+        {
+            return $"""[global::System.Runtime.CompilerServices.InterceptsLocationAttribute({location.Version}, "{location.Data}")]""";
+        }
         #endregion
     }
 }

src/Compilers/CSharp/Portable/CSharpResources.resx

@@ -7905,4 +7905,19 @@ To remove the warning, you can use /reference instead (set the Embed Interop Typ
   <data name="ERR_NoModifiersOnUsing" xml:space="preserve">
     <value>Modifiers cannot be placed on using declarations</value>
   </data>
+  <data name="ERR_InterceptsLocationDataInvalidFormat" xml:space="preserve">
+    <value>The data argument to InterceptsLocationAttribute is not in the correct format.</value>
+  </data>
+  <data name="ERR_InterceptsLocationUnsupportedVersion" xml:space="preserve">
+    <value>Version '{0}' of the interceptors format is not supported. The latest supported version is '1'.</value>
+  </data>
+  <data name="ERR_InterceptsLocationDuplicateFile" xml:space="preserve">
+    <value>Cannot intercept a call in file '{0}' because it is duplicated elsewhere in the compilation.</value>
+  </data>
+  <data name="ERR_InterceptsLocationFileNotFound" xml:space="preserve">
+    <value>Cannot intercept a call in file '{0}' because a matching file was not found in the compilation.</value>
+  </data>
+  <data name="ERR_InterceptsLocationDataInvalidPosition" xml:space="preserve">
+    <value>The data argument to InterceptsLocationAttribute refers to an invalid position in file '{0}'.</value>
+  </data>
 </root>

src/Compilers/CSharp/Portable/Compilation/CSharpCompilation.cs

@@ -3,6 +3,7 @@
 // See the LICENSE file in the project root for more information.
 
 using System;
+using System.Buffers.Binary;
 using System.Collections.Concurrent;
 using System.Collections.Generic;
 using System.Collections.Immutable;
@@ -160,8 +161,12 @@ internal Conversions Conversions
         private ImmutableSegmentedDictionary<string, OneOrMany<SyntaxTree>> _mappedPathToSyntaxTree;
 
         /// <summary>Lazily caches SyntaxTrees by their path. Used to look up the syntax tree referenced by an interceptor.</summary>
+        /// <remarks>Must be removed prior to interceptors stable release.</remarks>
         private ImmutableSegmentedDictionary<string, OneOrMany<SyntaxTree>> _pathToSyntaxTree;
 
+        /// <summary>Lazily caches SyntaxTrees by their xxHash128 checksum. Used to look up the syntax tree referenced by an interceptor.</summary>
+        private ImmutableSegmentedDictionary<ReadOnlyMemory<byte>, OneOrMany<SyntaxTree>> _contentHashToSyntaxTree;
+
         public override string Language
         {
             get
@@ -1075,6 +1080,45 @@ ImmutableSegmentedDictionary<string, OneOrMany<SyntaxTree>> computeMappedPathToS
             }
         }
 
+        private class ContentHashComparer : IEqualityComparer<ReadOnlyMemory<byte>>
+        {
+            public bool Equals(ReadOnlyMemory<byte> x, ReadOnlyMemory<byte> y)
+            {
+                return x.Span.SequenceEqual(y.Span);
+            }
+
+            public int GetHashCode(ReadOnlyMemory<byte> obj)
+            {
+                return BinaryPrimitives.ReadInt32LittleEndian(obj.Span);
+            }
+
+            public static ContentHashComparer Instance { get; } = new ContentHashComparer();
+        }
+
+        internal OneOrMany<SyntaxTree> GetSyntaxTreesByContentHash(ReadOnlyMemory<byte> contentHash)
+        {
+            var contentHashToSyntaxTree = _contentHashToSyntaxTree;
+            if (contentHashToSyntaxTree.IsDefault)
+            {
+                RoslynImmutableInterlocked.InterlockedInitialize(ref _contentHashToSyntaxTree, computeHashToSyntaxTree());
+                contentHashToSyntaxTree = _contentHashToSyntaxTree;
+            }
+
+            return contentHashToSyntaxTree.TryGetValue(contentHash, out var value) ? value : OneOrMany<SyntaxTree>.Empty;
+
+            ImmutableSegmentedDictionary<ReadOnlyMemory<byte>, OneOrMany<SyntaxTree>> computeHashToSyntaxTree()
+            {
+                var builder = ImmutableSegmentedDictionary.CreateBuilder<ReadOnlyMemory<byte>, OneOrMany<SyntaxTree>>(ContentHashComparer.Instance);
+                foreach (var tree in SyntaxTrees)
+                {
+                    var text = tree.GetText();
+                    var hash = text.GetContentHash().AsMemory();
+                    builder[hash] = builder.TryGetValue(hash, out var existing) ? existing.Add(tree) : OneOrMany.Create(tree);
+                }
+                return builder.ToImmutable();
+            }
+        }
+
         internal OneOrMany<SyntaxTree> GetSyntaxTreesByPath(string path)
         {
             // We could consider storing this on SyntaxAndDeclarationManager instead, and updating it incrementally.
@@ -2399,15 +2443,15 @@ internal void AddModuleInitializerMethod(MethodSymbol method)
         internal bool InterceptorsDiscoveryComplete;
 
         // NB: the 'Many' case for these dictionary values means there are duplicates. An error is reported for this after binding.
-        private ConcurrentDictionary<(string FilePath, int Line, int Character), OneOrMany<(Location AttributeLocation, MethodSymbol Interceptor)>>? _interceptions;
+        private ConcurrentDictionary<(string FilePath, int Position), OneOrMany<(Location AttributeLocation, MethodSymbol Interceptor)>>? _interceptions;
 
-        internal void AddInterception(string filePath, int line, int character, Location attributeLocation, MethodSymbol interceptor)
+        internal void AddInterception(string filePath, int position, Location attributeLocation, MethodSymbol interceptor)
         {
             Debug.Assert(!_declarationDiagnosticsFrozen);
             Debug.Assert(!InterceptorsDiscoveryComplete);
 
             var dictionary = LazyInitializer.EnsureInitialized(ref _interceptions);
-            dictionary.AddOrUpdate((filePath, line, character),
+            dictionary.AddOrUpdate((filePath, position),
                 addValueFactory: static (key, newValue) => OneOrMany.Create(newValue),
                 updateValueFactory: static (key, existingValues, newValue) =>
                 {
@@ -2427,9 +2471,9 @@ internal void AddInterception(string filePath, int line, int character, Location
                 factoryArgument: (AttributeLocation: attributeLocation, Interceptor: interceptor));
         }
 
-        internal (Location AttributeLocation, MethodSymbol Interceptor)? TryGetInterceptor(Location? callLocation)
+        internal (Location AttributeLocation, MethodSymbol Interceptor)? TryGetInterceptor(SimpleNameSyntax? node)
         {
-            if (callLocation is null || !callLocation.IsInSource)
+            if (node is null)
             {
                 return null;
             }
@@ -2440,9 +2484,7 @@ internal void AddInterception(string filePath, int line, int character, Location
                 return null;
             }
 
-            var callLineColumn = callLocation.GetLineSpan().Span.Start;
-            var key = (callLocation.SourceTree.FilePath, callLineColumn.Line, callLineColumn.Character);
-
+            var key = (node.SyntaxTree.FilePath, node.Position);
             if (_interceptions.TryGetValue(key, out var interceptionsAtAGivenLocation) && interceptionsAtAGivenLocation is [var oneInterception])
             {
                 return oneInterception;

src/Compilers/CSharp/Portable/Compilation/CSharpSemanticModel.cs

@@ -5207,13 +5207,34 @@ protected sealed override ISymbol GetDeclaredSymbolCore(SyntaxNode node, Cancell
 
             CheckSyntaxNode(node);
 
-            if (node.GetInterceptableNameSyntax() is { } nameSyntax && Compilation.TryGetInterceptor(nameSyntax.GetLocation()) is (_, MethodSymbol interceptor))
+            if (node.GetInterceptableNameSyntax() is { } nameSyntax && Compilation.TryGetInterceptor(nameSyntax) is (_, MethodSymbol interceptor))
             {
                 return interceptor.GetPublicSymbol();
             }
 
             return null;
         }
+
+#pragma warning disable RSEXPERIMENTAL002 // Internal usage of experimental API
+        public InterceptableLocation? GetInterceptableLocation(InvocationExpressionSyntax node, CancellationToken cancellationToken)
+        {
+            CheckSyntaxNode(node);
+            if (node.GetInterceptableNameSyntax() is not { } nameSyntax)
+            {
+                return null;
+            }
+
+            var tree = node.SyntaxTree;
+            var text = tree.GetText(cancellationToken);
+            var path = tree.FilePath;
+            var checksum = text.GetContentHash();
+
+            var lineSpan = nameSyntax.Location.GetLineSpan().Span.Start;
+            var lineNumberOneIndexed = lineSpan.Line + 1;
+            var characterNumberOneIndexed = lineSpan.Character + 1;
+
+            return new InterceptableLocation1(checksum, path, nameSyntax.Position, lineNumberOneIndexed, characterNumberOneIndexed);
+        }
 #nullable disable
 
         protected static SynthesizedPrimaryConstructor TryGetSynthesizedPrimaryConstructor(TypeDeclarationSyntax node, NamedTypeSymbol type)

src/Compilers/CSharp/Portable/Errors/ErrorCode.cs

@@ -2302,6 +2302,12 @@ internal enum ErrorCode
 
         ERR_NoModifiersOnUsing = 9229,
 
+        ERR_InterceptsLocationDataInvalidFormat = 9230,
+        ERR_InterceptsLocationUnsupportedVersion = 9231,
+        ERR_InterceptsLocationDuplicateFile = 9232,
+        ERR_InterceptsLocationFileNotFound = 9233,
+        ERR_InterceptsLocationDataInvalidPosition = 9234,
+
         #endregion
 
         // Note: you will need to do the following after adding warnings:

src/Compilers/CSharp/Portable/Errors/ErrorFacts.cs

@@ -2431,6 +2431,11 @@ internal static bool IsBuildOnlyDiagnostic(ErrorCode code)
                 case ErrorCode.ERR_ParamsCollectionExtensionAddMethod:
                 case ErrorCode.ERR_ParamsCollectionMissingConstructor:
                 case ErrorCode.ERR_NoModifiersOnUsing:
+                case ErrorCode.ERR_InterceptsLocationDataInvalidFormat:
+                case ErrorCode.ERR_InterceptsLocationUnsupportedVersion:
+                case ErrorCode.ERR_InterceptsLocationDuplicateFile:
+                case ErrorCode.ERR_InterceptsLocationFileNotFound:
+                case ErrorCode.ERR_InterceptsLocationDataInvalidPosition:
                     return false;
                 default:
                     // NOTE: All error codes must be explicitly handled in this switch statement

src/Compilers/CSharp/Portable/Lowering/LocalRewriter/LocalRewriter.cs

@@ -247,7 +247,7 @@ private PEModuleBuilder? EmitModule
             {
                 Debug.Assert(!nameofOperator.WasCompilerGenerated);
                 var nameofIdentiferSyntax = (IdentifierNameSyntax)((InvocationExpressionSyntax)nameofOperator.Syntax).Expression;
-                if (this._compilation.TryGetInterceptor(nameofIdentiferSyntax.Location) is not null)
+                if (this._compilation.TryGetInterceptor(nameofIdentiferSyntax) is not null)
                 {
                     this._diagnostics.Add(ErrorCode.ERR_InterceptorCannotInterceptNameof, nameofIdentiferSyntax.Location);
                 }