Convert Perl scripts to C# in coreclr (#104867)

* Convert Perl scripts to C# in coreclr

* Delete obsolete IL script

* Add license header and regenerate

* Add a README

* Mention generator in ilasm readme and cleanups

src/coreclr/ilasm/GrammarExtractor/GrammarExtractor.csproj

@@ -0,0 +1,9 @@
+<Project Sdk="Microsoft.NET.Sdk">
+
+  <PropertyGroup>
+    <OutputType>Exe</OutputType>
+    <TargetFramework>$(NetCoreAppToolCurrent)</TargetFramework>
+    <Nullable>enable</Nullable>
+  </PropertyGroup>
+
+</Project>

src/coreclr/ilasm/GrammarExtractor/Program.cs

@@ -0,0 +1,98 @@
+// Licensed to the .NET Foundation under one or more agreements.
+// The .NET Foundation licenses this file to you under the MIT license.
+
+using System;
+using System.IO;
+using System.Text.RegularExpressions;
+using static Patterns;
+
+if (args.Length != 1)
+{
+    Console.Error.WriteLine("Usage: <file>");
+    return 1;
+}
+
+string filePath = args[0];
+
+string fileContent = File.ReadAllText(filePath);
+
+var match = GetRegexExtractMarkers().Match(fileContent);
+if (!match.Success)
+{
+    Console.Error.WriteLine("Could not find %% markers");
+    return 1;
+}
+
+//string prefix = match.Groups[1].Value;
+string grammar = match.Groups[2].Value;
+
+// Remove any text in {}
+var regexRemoveTextInBraces = GetRegexRemoveTextInBraces();
+string previousGrammar;
+
+do
+{
+    previousGrammar = grammar;
+    grammar = regexRemoveTextInBraces.Replace(grammar, "$1");
+} while (grammar != previousGrammar);
+
+// Change keyword identifiers into the string they represent (lowercase)
+grammar = GetRegexKeywordIdentifiers().Replace(grammar, m => $"'{m.Groups[1].Value.ToLowerInvariant()}'");
+
+// Change assembler directives into their string (lowercase with a period)
+grammar = GetRegexAssemblerDirectives().Replace(grammar, m => $"'.{m.Groups[1].Value.ToLowerInvariant()}'");
+
+// Handle special punctuation
+grammar = GetRegexEllipsis().Replace(grammar, "'...'");
+grammar = GetRegexDcolon().Replace(grammar, "'::'");
+
+// Print the output header
+Console.Write(@"// Licensed to the .NET Foundation under one or more agreements.
+// The .NET Foundation licenses this file to you under the MIT license.
+
+Lexical tokens
+    ID - C style alphaNumeric identifier (e.g. Hello_There2)
+    DOTTEDNAME - Sequence of dot-separated IDs (e.g. System.Object)
+    QSTRING  - C style quoted string (e.g.  ""hi\n"")
+    SQSTRING - C style singlely quoted string(e.g.  'hi')
+    INT32    - C style 32 bit integer (e.g.  235,  03423, 0x34FFF)
+    INT64    - C style 64 bit integer (e.g.  -2353453636235234,  0x34FFFFFFFFFF)
+    FLOAT64  - C style floating point number (e.g.  -0.2323, 354.3423, 3435.34E-5)
+    INSTR_*  - IL instructions of a particular class (see opcode.def).
+    HEXBYTE  - 1- or 2-digit hexadecimal number (e.g., A2, F0).
+Auxiliary lexical tokens
+    TYPEDEF_T - Aliased class (TypeDef or TypeRef).
+    TYPEDEF_M - Aliased method.
+    TYPEDEF_F - Aliased field.
+    TYPEDEF_TS - Aliased type specification (TypeSpec).
+    TYPEDEF_MR - Aliased field/method reference (MemberRef).
+    TYPEDEF_CA - Aliased Custom Attribute.
+----------------------------------------------------------------------------------
+START           : decls
+                ;");
+
+// Print the output
+Console.Write(grammar);
+
+return 0;
+
+internal static partial class Patterns
+{
+    [GeneratedRegex(@"^(.*)%%(.*)%%", RegexOptions.Singleline)]
+    internal static partial Regex GetRegexExtractMarkers();
+
+    [GeneratedRegex(@"\s*([^'])\{[^{}]*\}", RegexOptions.Singleline)]
+    internal static partial Regex GetRegexRemoveTextInBraces();
+
+    [GeneratedRegex(@"\b([A-Z0-9_]+)_\b", RegexOptions.Singleline)]
+    internal static partial Regex GetRegexKeywordIdentifiers();
+
+    [GeneratedRegex(@"\b_([A-Z0-9]+)\b", RegexOptions.Singleline)]
+    internal static partial Regex GetRegexAssemblerDirectives();
+
+    [GeneratedRegex(@"\bELLIPSIS\b", RegexOptions.Singleline)]
+    internal static partial Regex GetRegexEllipsis();
+
+    [GeneratedRegex(@"\bDCOLON\b", RegexOptions.Singleline)]
+    internal static partial Regex GetRegexDcolon();
+}

src/coreclr/ilasm/GrammarExtractor/README.md

@@ -0,0 +1,10 @@
+# Grammar extractor for IL tools
+
+Tool to extract IL grammar in `Backus-Naur Form (BNF)` from `Yet Another Compiler-Compiler (Yacc)`.
+
+Usage:
+
+```sh
+cd runtime
+./dotnet.sh run --project src/coreclr/ilasm/GrammarExtractor src/coreclr/ilasm/asmparse.y > src/coreclr/ilasm/prebuilt/asmparse.grammar
+```

src/coreclr/ilasm/README.md

@@ -16,3 +16,4 @@ $ docker run --rm -v$(pwd):/runtime -w /runtime/src/coreclr/ilasm alpine \
     sh -c 'apk add bison && yacc asmparse.y -o prebuilt/asmparse.cpp'
 ```
 
+To generate grammar, see [GrammarExtractor README](GrammarExtractor/README.md).

src/coreclr/ilasm/prebuilt/asmparse.grammar

@@ -76,11 +76,11 @@ dottedName              : id
                         | dottedName '.' dottedName 
                         ;
 
-int32                   : INT32 
+int32                   : INT32_V 
                         ;
 
-int64                   : INT64 
-                        | INT32 
+int64                   : INT64_V 
+                        | INT32_V 
                         ;
 
 float64                 : FLOAT64