Add debug-only use of new AssemblyBuilder.Save in Regex.CompileToAsse…

…mbly

To aid in debugging RegexCompiler issues and to help vet the new AssemblyBuilder.Save support.

src/libraries/System.Text.RegularExpressions/src/System.Text.RegularExpressions.csproj

@@ -28,6 +28,7 @@
     <Compile Include="System\Text\RegularExpressions\Regex.Split.cs" />
     <Compile Include="System\Text\RegularExpressions\Regex.EnumerateMatches.cs" />
     <Compile Include="System\Text\RegularExpressions\Regex.Timeout.cs" />
+    <Compile Include="System\Text\RegularExpressions\RegexAssemblyCompiler.cs" />
     <Compile Include="System\Text\RegularExpressions\RegexCaseBehavior.cs" />
     <Compile Include="System\Text\RegularExpressions\RegexCaseEquivalences.Data.cs" />
     <Compile Include="System\Text\RegularExpressions\RegexCaseEquivalences.cs" />

src/libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/Regex.cs

@@ -227,8 +227,41 @@ public static void CompileToAssembly(RegexCompilationInfo[] regexinfos, Assembly
             CompileToAssembly(regexinfos, assemblyname, attributes, null);
 
         [Obsolete(Obsoletions.RegexCompileToAssemblyMessage, DiagnosticId = Obsoletions.RegexCompileToAssemblyDiagId, UrlFormat = Obsoletions.SharedUrlFormat)]
-        public static void CompileToAssembly(RegexCompilationInfo[] regexinfos, AssemblyName assemblyname, CustomAttributeBuilder[]? attributes, string? resourceFile) =>
+        public static void CompileToAssembly(RegexCompilationInfo[] regexinfos, AssemblyName assemblyname, CustomAttributeBuilder[]? attributes, string? resourceFile)
+        {
+#if DEBUG
+            // This code exists only to help with the development of the RegexCompiler.
+            // .NET no longer supports CompileToAssembly; the source generator should be used instead.
+#pragma warning disable IL3050
+            ArgumentNullException.ThrowIfNull(assemblyname);
+            ArgumentNullException.ThrowIfNull(regexinfos);
+
+            var c = new RegexAssemblyCompiler(assemblyname, attributes, resourceFile);
+
+            for (int i = 0; i < regexinfos.Length; i++)
+            {
+                ArgumentNullException.ThrowIfNull(regexinfos[i]);
+
+                string pattern = regexinfos[i].Pattern;
+
+                RegexOptions options = regexinfos[i].Options | RegexOptions.Compiled; // ensure compiled is set; it enables more optimization specific to compilation
+
+                string fullname = regexinfos[i].Namespace.Length == 0 ?
+                    regexinfos[i].Name :
+                    regexinfos[i].Namespace + "." + regexinfos[i].Name;
+
+                RegexTree tree = RegexParser.Parse(pattern, options, (options & RegexOptions.CultureInvariant) != 0 ? CultureInfo.InvariantCulture : CultureInfo.CurrentCulture);
+                RegexInterpreterCode code = RegexWriter.Write(tree);
+
+                c.GenerateRegexType(pattern, options, fullname, regexinfos[i].IsPublic, tree, code, regexinfos[i].MatchTimeout);
+            }
+
+            c.Save(assemblyname.Name ?? "RegexCompileToAssembly");
+#pragma warning restore IL3050
+#else
             throw new PlatformNotSupportedException(SR.PlatformNotSupported_CompileToAssembly);
+#endif
+        }
 
         /// <summary>
         /// Escapes a minimal set of metacharacters (\, *, +, ?, |, {, [, (, ), ^, $, ., #, and

src/libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexAssemblyCompiler.cs

@@ -0,0 +1,273 @@
+// Licensed to the .NET Foundation under one or more agreements.
+// The .NET Foundation licenses this file to you under the MIT license.
+
+using System.Collections;
+using System.Collections.Generic;
+using System.Diagnostics.CodeAnalysis;
+using System.Reflection;
+using System.Reflection.Emit;
+using System.Threading;
+
+// WARNING:
+// The code generated by RegexCompiler is not intended to be saved to disk and loaded back later.
+// It accesses internal and private members of System.Text.RegularExpressions, which may change
+// at any point in the future, and relies on details about the current machine/process, e.g. is
+// it 32-bit or 64-bit. The generated surface area has also not been vetted. This code exists
+// only for debugging purposes, to make it easier to examine the IL that RegexCompiler emits.
+
+#if DEBUG
+namespace System.Text.RegularExpressions
+{
+    /// <summary>Compiles a Regex to an assembly that can be saved to disk.</summary>
+    [RequiresDynamicCode("The RegexAssemblyCompiler type requires dynamic code to be enabled.")]
+    internal sealed class RegexAssemblyCompiler : RegexCompiler
+    {
+        /// <summary>Type count used to augment generated type names to create unique names.</summary>
+        private static int s_typeCount;
+
+        private readonly AssemblyBuilder _assembly;
+        private readonly ModuleBuilder _module;
+        private readonly MethodInfo _save;
+
+        internal RegexAssemblyCompiler(AssemblyName an, CustomAttributeBuilder[]? attribs, string? resourceFile)
+        {
+            if (resourceFile != null)
+            {
+                // Unmanaged resources are not supported: _assembly.DefineUnmanagedResource(resourceFile);
+                throw new PlatformNotSupportedException();
+            }
+
+            // TODO: Use public API when it's available: https://github.com/dotnet/runtime/issues/15704
+            Type abType = Type.GetType("System.Reflection.Emit.AssemblyBuilderImpl, System.Reflection.Emit", throwOnError: true)!;
+            MethodInfo defineDynamicAssembly = abType.GetMethod("DefinePersistedAssembly",
+                BindingFlags.NonPublic | BindingFlags.Static,
+                [typeof(AssemblyName), typeof(Assembly), typeof(List<CustomAttributeBuilder>)]) ??
+                throw new InvalidOperationException("Could not find method AssemblyBuilderImpl.DefinePersistedAssembly");
+            _assembly = (AssemblyBuilder?)defineDynamicAssembly.Invoke(null, [an, typeof(object).Assembly, attribs is not null ? new List<CustomAttributeBuilder>(attribs) : null]) ??
+                throw new InvalidOperationException("DefinePersistedAssembly returned null");
+            _save = abType.GetMethod("Save", BindingFlags.NonPublic | BindingFlags.Instance, [typeof(string)]) ??
+                throw new InvalidOperationException("Could not find method AssemblyBuilderImpl.Save");
+
+            _module = _assembly.DefineDynamicModule(an.Name + ".dll");
+        }
+
+        internal void GenerateRegexType(string pattern, RegexOptions options, string name, bool isPublic, RegexTree tree, RegexInterpreterCode code, TimeSpan matchTimeout)
+        {
+            // Store arguments into the base type's fields
+            _options = options;
+            _regexTree = tree;
+
+            // Pick a name for the class.
+            string typenumString = ((uint)Interlocked.Increment(ref s_typeCount)).ToString();
+
+            // Generate the RegexRunner-derived type.
+            TypeBuilder regexRunnerTypeBuilder = DefineType(_module, $"{name}Runner{typenumString}", isPublic: false, isSealed: true, typeof(RegexRunner));
+
+            _ilg = DefineMethod(regexRunnerTypeBuilder, "TryFindNextPossibleStartingPosition", [typeof(ReadOnlySpan<char>)], typeof(bool), out MethodBuilder tryFindNextPossibleStartingPositionMethod);
+            EmitTryFindNextPossibleStartingPosition();
+
+            _ilg = DefineMethod(regexRunnerTypeBuilder, "TryMatchAtCurrentPosition", [typeof(ReadOnlySpan<char>)], typeof(bool), out MethodBuilder tryMatchAtCurrentPositionMethod);
+            EmitTryMatchAtCurrentPosition();
+
+            _ilg = DefineMethod(regexRunnerTypeBuilder, "Scan", [typeof(ReadOnlySpan<char>)], null, out _);
+            EmitScan(options, tryFindNextPossibleStartingPositionMethod, tryMatchAtCurrentPositionMethod);
+
+            Type runnerType = regexRunnerTypeBuilder.CreateType()!;
+
+            // Generate the RegexRunnerFactory-derived type.
+            TypeBuilder regexRunnerFactoryTypeBuilder = DefineType(_module, $"{name}Factory{typenumString}", isPublic: false, isSealed: true, typeof(RegexRunnerFactory));
+            _ilg = DefineMethod(regexRunnerFactoryTypeBuilder, "CreateInstance", null, typeof(RegexRunner), out _);
+            GenerateCreateInstance(runnerType);
+            Type regexRunnerFactoryType = regexRunnerFactoryTypeBuilder.CreateType()!;
+
+            // Generate the Regex-derived type.
+            TypeBuilder regexTypeBuilder = DefineType(_module, name, isPublic, isSealed: false, typeof(Regex));
+            ConstructorBuilder defaultCtorBuilder = regexTypeBuilder.DefineConstructor(MethodAttributes.Public, CallingConventions.Standard, Type.EmptyTypes);
+            _ilg = defaultCtorBuilder.GetILGenerator();
+            GenerateRegexDefaultCtor(pattern, options, regexRunnerFactoryType, tree, code, matchTimeout);
+            if (matchTimeout != Regex.InfiniteMatchTimeout)
+            {
+                // We only generate a constructor with a timeout parameter if the regex information supplied has a non-infinite timeout.
+                // If it has an infinite timeout, then the generated code is not going to respect the timeout. This is a difference from netfx,
+                // due to the fact that we now special-case an infinite timeout in the code generator to avoid spitting unnecessary code
+                // and paying for the checks at run time.
+                _ilg = regexTypeBuilder.DefineConstructor(MethodAttributes.Public, CallingConventions.Standard, new Type[] { typeof(TimeSpan) }).GetILGenerator();
+                GenerateRegexTimeoutCtor(defaultCtorBuilder, regexTypeBuilder);
+            }
+            regexTypeBuilder.CreateType();
+        }
+
+        /// <summary>Generates a very simple factory method.</summary>
+        private void GenerateCreateInstance([DynamicallyAccessedMembers(DynamicallyAccessedMemberTypes.PublicParameterlessConstructor)] Type type)
+        {
+            // return new Type();
+            _ilg!.Emit(OpCodes.Newobj, type.GetConstructor(Type.EmptyTypes)!);
+            Ret();
+        }
+
+        private void GenerateRegexDefaultCtor(
+            string pattern,
+            RegexOptions options,
+            [DynamicallyAccessedMembers(DynamicallyAccessedMemberTypes.PublicParameterlessConstructor)] Type regexRunnerFactoryType,
+            RegexTree tree,
+            RegexInterpreterCode code,
+            TimeSpan matchTimeout)
+        {
+            // Call the base ctor and store pattern, options, and factory.
+            // base.ctor();
+            // base.pattern = pattern;
+            // base.options = options;
+            // base.factory = new DerivedRegexRunnerFactory();
+            Ldthis();
+            _ilg!.Emit(OpCodes.Call, typeof(Regex).GetConstructor(BindingFlags.Public | BindingFlags.NonPublic | BindingFlags.Instance, null, Type.EmptyTypes, Array.Empty<ParameterModifier>())!);
+            Ldthis();
+            Ldstr(pattern);
+            Stfld(RegexField(nameof(Regex.pattern)));
+            Ldthis();
+            Ldc((int)options);
+            Stfld(RegexField(nameof(Regex.roptions)));
+            Ldthis();
+            _ilg!.Emit(OpCodes.Newobj, regexRunnerFactoryType.GetConstructor(Type.EmptyTypes)!);
+            Stfld(RegexField(nameof(Regex.factory)));
+
+            // Store the timeout (no need to validate as it should have happened in RegexCompilationInfo)
+            Ldthis();
+            if (matchTimeout == Regex.InfiniteMatchTimeout)
+            {
+                // base.internalMatchTimeout = Regex.InfiniteMatchTimeout;
+                _ilg.Emit(OpCodes.Ldsfld, RegexField(nameof(Regex.InfiniteMatchTimeout)));
+            }
+            else
+            {
+                // base.internalMatchTimeout = TimeSpan.FromTick(matchTimeout.Ticks);
+                LdcI8(matchTimeout.Ticks);
+                Call(typeof(TimeSpan).GetMethod(nameof(TimeSpan.FromTicks), BindingFlags.Public | BindingFlags.Static)!);
+            }
+            Stfld(RegexField(nameof(Regex.internalMatchTimeout)));
+
+            // Set capsize, caps, capnames, capslist.
+            Ldthis();
+            Ldc(tree.CaptureCount);
+            Stfld(RegexField(nameof(Regex.capsize)));
+            if (tree.CaptureNumberSparseMapping != null)
+            {
+                // Caps = new Hashtable {{0, 0}, {1, 1}, ... };
+                GenerateCreateHashtable(RegexField(nameof(Regex.caps)), tree.CaptureNumberSparseMapping);
+            }
+            if (tree.CaptureNameToNumberMapping != null)
+            {
+                // CapNames = new Hashtable {{"0", 0}, {"1", 1}, ...};
+                GenerateCreateHashtable(RegexField(nameof(Regex.capnames)), tree.CaptureNameToNumberMapping);
+            }
+            if (tree.CaptureNames != null)
+            {
+                // capslist = new string[...];
+                // capslist[0] = "0";
+                // capslist[1] = "1";
+                // ...
+                Ldthis();
+                Ldc(tree.CaptureNames.Length);
+                _ilg.Emit(OpCodes.Newarr, typeof(string));  // create new string array
+                FieldInfo capslistField = RegexField(nameof(Regex.capslist));
+                Stfld(capslistField);
+                for (int i = 0; i < tree.CaptureNames.Length; i++)
+                {
+                    Ldthisfld(capslistField);
+                    Ldc(i);
+                    Ldstr(tree.CaptureNames[i]);
+                    _ilg.Emit(OpCodes.Stelem_Ref);
+                }
+            }
+
+            // return;
+            Ret();
+        }
+
+        private void GenerateRegexTimeoutCtor(ConstructorBuilder defaultCtorBuilder, TypeBuilder regexTypeBuilder)
+        {
+            // base.ctor();
+            // ValidateMatchTimeout(timeSpan);
+            // base.internalMatchTimeout = timeSpan;
+            Ldthis();
+            _ilg!.Emit(OpCodes.Call, defaultCtorBuilder);
+            _ilg.Emit(OpCodes.Ldarg_1);
+            Call(typeof(Regex).GetMethod(nameof(Regex.ValidateMatchTimeout), BindingFlags.Public | BindingFlags.NonPublic | BindingFlags.Static)!);
+            Ldthis();
+            _ilg.Emit(OpCodes.Ldarg_1);
+            Stfld(RegexField(nameof(Regex.internalMatchTimeout)));
+            Ret();
+        }
+
+        internal void GenerateCreateHashtable(FieldInfo field, Hashtable ht)
+        {
+            // hashtable = new Hashtable();
+            Ldthis();
+            _ilg!.Emit(OpCodes.Newobj, typeof(Hashtable).GetConstructor(Type.EmptyTypes)!);
+            Stfld(field);
+
+            // hashtable.Add(key1, value1);
+            // hashtable.Add(key2, value2);
+            // ...
+            MethodInfo addMethod = typeof(Hashtable).GetMethod(nameof(Hashtable.Add), BindingFlags.Public | BindingFlags.Instance)!;
+            IDictionaryEnumerator en = ht.GetEnumerator();
+            while (en.MoveNext())
+            {
+                Ldthisfld(field);
+
+                if (en.Key is int key)
+                {
+                    Ldc(key);
+                    _ilg!.Emit(OpCodes.Box, typeof(int));
+                }
+                else
+                {
+                    Ldstr((string)en.Key);
+                }
+
+                Ldc((int)en.Value!);
+                _ilg!.Emit(OpCodes.Box, typeof(int));
+                Callvirt(addMethod);
+            }
+        }
+
+        /// <summary>Gets the named instance field from the Regex type.</summary>
+        private static FieldInfo RegexField(string fieldname) =>
+            typeof(Regex).GetField(fieldname, BindingFlags.Public | BindingFlags.NonPublic | BindingFlags.Instance | BindingFlags.Static)!;
+
+        /// <summary>Saves the assembly to a file in the current directory based on the assembly's name.</summary>
+        internal void Save(string fileName)
+        {
+            if (!fileName.EndsWith(".dll", StringComparison.Ordinal))
+            {
+                fileName += ".dll";
+            }
+
+            _save.Invoke(_assembly, [fileName]); // TODO: Use public API when it's available: https://github.com/dotnet/runtime/issues/15704
+        }
+
+        /// <summary>Begins the definition of a new type with a specified base class</summary>
+        private static TypeBuilder DefineType(
+            ModuleBuilder moduleBuilder,
+            string typeName,
+            bool isPublic,
+            bool isSealed,
+            [DynamicallyAccessedMembers(DynamicallyAccessedMemberTypes.All)] Type inheritFromClass)
+        {
+            TypeAttributes attrs = TypeAttributes.Class | TypeAttributes.BeforeFieldInit | (isPublic ? TypeAttributes.Public : TypeAttributes.NotPublic);
+            if (isSealed)
+            {
+                attrs |= TypeAttributes.Sealed;
+            }
+
+            return moduleBuilder.DefineType(typeName, attrs, inheritFromClass);
+        }
+
+        /// <summary>Begins the definition of a new method (no args) with a specified return value.</summary>
+        private static ILGenerator DefineMethod(TypeBuilder typeBuilder, string methname, Type[]? parameterTypes, Type? returnType, out MethodBuilder builder)
+        {
+            builder = typeBuilder.DefineMethod(methname, MethodAttributes.Family | MethodAttributes.Virtual, returnType, parameterTypes);
+            return builder.GetILGenerator();
+        }
+    }
+}
+#endif

src/libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexCompiler.cs

@@ -5305,7 +5305,7 @@ void EmitStackPop()
             }
         }
 
-        protected void EmitScan(RegexOptions options, DynamicMethod tryFindNextStartingPositionMethod, DynamicMethod tryMatchAtCurrentPositionMethod)
+        protected void EmitScan(RegexOptions options, MethodInfo tryFindNextStartingPositionMethod, MethodInfo tryMatchAtCurrentPositionMethod)
         {
             // As with the source generator, we can emit special code for common circumstances rather than always emitting
             // the most general purpose scan loop.  Unlike the source generator, however, code appearance isn't important

src/libraries/System.Text.RegularExpressions/src/System/Text/RegularExpressions/RegexLWCGCompiler.cs

@@ -24,7 +24,7 @@ internal sealed class RegexLWCGCompiler : RegexCompiler
         /// </remarks>
         private static readonly bool s_includePatternInName = Environment.GetEnvironmentVariable(IncludePatternInNamesEnvVar) == "1";
 
-        /// <summary>Parameter types for the generated Go and FindFirstChar methods.</summary>
+        /// <summary>Parameter types for the generated TryFindNextPossibleStartingPosition and TryMatchAtCurrentPosition methods.</summary>
         private static readonly Type[] s_paramTypes = [typeof(RegexRunner), typeof(ReadOnlySpan<char>)];
 
         /// <summary>Id number to use for the next compiled regex.</summary>