[lldb] Use Python script to generate SBLanguages.h (#90753)

Use a Python script to generate SBLanguages.h instead of piggybacking on
LLDB TableGen. This addresses Nico Weber's post-commit feedback.

lldb/scripts/generate-sbapi-dwarf-enum.py

@@ -0,0 +1,67 @@
+#!/usr/bin/env python3
+
+import argparse
+import re
+
+HEADER = """\
+//===-- SBLanguages.h -----------------------------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLDB_API_SBLANGUAGE_H
+#define LLDB_API_SBLANGUAGE_H
+/// Used by \\ref SBExpressionOptions.
+/// These enumerations use the same language enumerations as the DWARF
+/// specification for ease of use and consistency.
+enum SBSourceLanguageName : uint16_t {
+"""
+
+FOOTER = """\
+};
+
+#endif
+"""
+
+REGEX = re.compile(
+    r'^ *HANDLE_DW_LNAME *\( *(?P<value>[^,]+), (?P<comment>[^"]+), "(?P<name>.*)",.*\)'
+)
+
+
+def emit_enum(input, output):
+    # Read the input and break it up by lines.
+    lines = []
+    with open(input, "r") as f:
+        lines = f.readlines()
+
+    # Write the output.
+    with open(output, "w") as f:
+        # Emit the header.
+        f.write(HEADER)
+
+        # Emit the enum values.
+        for line in lines:
+            match = REGEX.match(line)
+            if not match:
+                continue
+            f.write(f"  /// {match.group('comment')}.\n")
+            f.write(f"  eLanguageName{match.group('name')} = {match.group('value')},\n")
+
+        # Emit the footer
+        f.write(FOOTER)
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument("--output", "-o")
+    parser.add_argument("input")
+    args = parser.parse_args()
+
+    emit_enum(args.input, args.output)
+
+
+if __name__ == "__main__":
+    main()

lldb/source/API/CMakeLists.txt

@@ -20,9 +20,19 @@ if(LLDB_ENABLE_LUA)
   set(lldb_lua_wrapper ${lua_bindings_dir}/LLDBWrapLua.cpp)
 endif()
 
-lldb_tablegen(../../include/lldb/API/SBLanguages.h -gen-lldb-sbapi-dwarf-enum
-  SOURCE ${LLVM_MAIN_INCLUDE_DIR}/llvm/BinaryFormat/Dwarf.def
-  TARGET lldb-sbapi-dwarf-enums)
+# Target to generate SBLanguages.h from Dwarf.def.
+set(sb_languages_file
+  ${CMAKE_CURRENT_BINARY_DIR}/../../include/lldb/API/SBLanguages.h)
+add_custom_target(
+  lldb-sbapi-dwarf-enums
+  "${Python3_EXECUTABLE}"
+      ${LLDB_SOURCE_DIR}/scripts/generate-sbapi-dwarf-enum.py
+      ${LLVM_MAIN_INCLUDE_DIR}/llvm/BinaryFormat/Dwarf.def
+      -o ${sb_languages_file}
+  BYPRODUCTS ${sb_languages_file}
+  DEPENDS ${LLVM_MAIN_INCLUDE_DIR}/llvm/BinaryFormat/Dwarf.def
+  WORKING_DIRECTORY ${LLVM_LIBRARY_OUTPUT_INTDIR}
+)
 
 add_lldb_library(liblldb SHARED ${option_framework}
   SBAddress.cpp
@@ -106,7 +116,7 @@ add_lldb_library(liblldb SHARED ${option_framework}
 
   DEPENDS
     lldb-sbapi-dwarf-enums
-  
+
   LINK_LIBS
     lldbBreakpoint
     lldbCore

lldb/utils/TableGen/CMakeLists.txt

@@ -10,7 +10,6 @@ if (NOT DEFINED LLDB_TABLEGEN_EXE)
     add_tablegen(lldb-tblgen LLDB
       LLDBOptionDefEmitter.cpp
       LLDBPropertyDefEmitter.cpp
-      LLDBSBAPIDWARFEnum.cpp
       LLDBTableGen.cpp
       LLDBTableGenUtils.cpp
       )

lldb/utils/TableGen/LLDBSBAPIDWARFEnum.cpp

@@ -1,67 +0,0 @@
-//===- LLDBPropertyDefEmitter.cpp -----------------------------------------===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-//
-// Produce the list of source languages header file fragment for the SBAPI.
-//
-//===----------------------------------------------------------------------===//
-
-#include <fstream>
-#include <llvm/ADT/StringRef.h>
-#include <regex>
-
-namespace lldb_private {
-int EmitSBAPIDWARFEnum(int argc, char **argv) {
-  std::string InputFilename;
-  std::string OutputFilename;
-  std::string DepFilename;
-  // This command line option parser is as robust as the worst shell script.
-  for (int i = 0; i < argc; ++i) {
-    if (llvm::StringRef(argv[i]).ends_with("Dwarf.def"))
-      InputFilename = std::string(argv[i]);
-    if (llvm::StringRef(argv[i]) == "-o" && i + 1 < argc)
-      OutputFilename = std::string(argv[i + 1]);
-    if (llvm::StringRef(argv[i]) == "-d" && i + 1 < argc)
-      DepFilename = std::string(argv[i + 1]);
-  }
-  std::ifstream input(InputFilename);
-  std::ofstream output(OutputFilename);
-  output
-      << R"(//===-- SBLanguages.h -----------------------------------------*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef LLDB_API_SBLANGUAGE_H
-#define LLDB_API_SBLANGUAGE_H
-/// Used by \ref SBExpressionOptions.
-/// These enumerations use the same language enumerations as the DWARF
-/// specification for ease of use and consistency.
-enum SBSourceLanguageName : uint16_t {
-)";
-  std::string line;
-  std::regex macro_regex(R"(^ *HANDLE_DW_LNAME *\( *([^,]+), ([^,]+), )"
-                         "\"(.*)\",.*\\).*",
-                         std::regex::extended);
-  while (std::getline(input, line)) {
-    std::smatch match;
-    if (!std::regex_match(line, match, macro_regex))
-      continue;
-
-    output << "  /// " << match[3] << ".\n";
-    output << "  eLanguageName" << match[2] << " = " << match[1] << ",\n";
-  }
-  output << "};\n\n";
-  output << "#endif\n";
-  // Emit the dependencies file.
-  std::ofstream(DepFilename) << OutputFilename << ": " << InputFilename << '\n';
-  return 0;
-}
-} // namespace lldb_private

lldb/utils/TableGen/LLDBTableGen.cpp

@@ -27,7 +27,6 @@ enum ActionType {
   GenOptionDefs,
   GenPropertyDefs,
   GenPropertyEnumDefs,
-  GenSBAPIDWARFEnum
 };
 
 static cl::opt<ActionType> Action(
@@ -41,8 +40,6 @@ static cl::opt<ActionType> Action(
                clEnumValN(GenPropertyDefs, "gen-lldb-property-defs",
                           "Generate lldb property definitions"),
                clEnumValN(GenPropertyEnumDefs, "gen-lldb-property-enum-defs",
-                          "Generate lldb property enum definitions"),
-               clEnumValN(GenSBAPIDWARFEnum, "gen-lldb-sbapi-dwarf-enum",
                           "Generate lldb property enum definitions")));
 
 static bool LLDBTableGenMain(raw_ostream &OS, RecordKeeper &Records) {
@@ -62,8 +59,6 @@ static bool LLDBTableGenMain(raw_ostream &OS, RecordKeeper &Records) {
   case GenPropertyEnumDefs:
     EmitPropertyEnumDefs(Records, OS);
     break;
-  case GenSBAPIDWARFEnum:
-    llvm_unreachable("already handled");
   }
   return false;
 }
@@ -74,9 +69,6 @@ int main(int argc, char **argv) {
   cl::ParseCommandLineOptions(argc, argv);
   llvm_shutdown_obj Y;
 
-  if (Action == GenSBAPIDWARFEnum)
-    return EmitSBAPIDWARFEnum(argc, argv);
-
   return TableGenMain(argv[0], &LLDBTableGenMain);
 }