不要な関数・変数・コメントを削除するやつ

cleanup.py

import subprocess
import sys
from pathlib import Path
from clang.cindex import CursorKind, Index

# argv チェック
if len(sys.argv) == 1:
    print(f"Usage: python3 {sys.argv[0]} <filename> [compile options]")
    sys.exit(1)

filename = Path(sys.argv[1]).absolute()
user_options = sys.argv[2:]

# コンパイルできることを確認
subprocess.check_call(["clang++", "-fsyntax-only", "-std=gnu++2b"] + user_options + [str(filename)])

# ファイルを読み込む
with open(filename, "rb") as f:
    program = f.read().splitlines(keepends=True)

#line を削除
program = [line for line in program if not line.startswith(b"#line")]

#include をコメントアウト
prefix = b"//cleanup_Sq9XpeQ6"
for i, line in enumerate(program):
    if line.startswith(b"#include"):
        program[i] = prefix + line

# clang でマクロを展開
p = subprocess.run(["clang++", "-std=gnu++23", "-w", "-E", "-C", "-P", "-xc++", "-"] + user_options, input=b"".join(program), stdout=subprocess.PIPE)
program = p.stdout.splitlines(keepends=True)

#include を元に戻す
for i, line in enumerate(program):
    if line.startswith(prefix):
        program[i] = line.removeprefix(prefix)

# GCC でコメントを削除
p = subprocess.run(["gcc", "-std=gnu++23", "-w", "-fpreprocessed", "-E", "-P", "-xc++", "-"] + user_options, input=b"".join(program), stdout=subprocess.PIPE)
program = p.stdout.splitlines(keepends=True)

# bytearray に変換
program = [bytearray(line) for line in program]

# 指定された範囲を空白に変換
def erase(extent):
    l1 = extent.start.line - 1
    c1 = extent.start.column - 1
    l2 = extent.end.line - 1
    c2 = extent.end.column - 1
    while l1 < l2 or (l1 == l2 and c1 < c2):
        if program[l1][c1] == ord('\n'):
            c1 = 0
            l1 += 1
            continue
        program[l1][c1] = ord(' ')
        c1 += 1

options = (
    f"""
    -resource-dir /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/lib/clang/16
    -isysroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk
    -I /usr/local/include
    -I /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include/c++/v1
    -I /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/local/include
    -I /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX.sdk/usr/include
    -I /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/lib/clang/16/include
    -I /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/include
    -I {filename.parent}
    -fcxx-exceptions
    -fsyntax-only
    -std=gnu++2b
    -Wno-everything
    -Wunused
    -Wunused-template
    -Wunused-macros
    -Wunused-member-function
    """.split() + user_options
)
program_fix = b"".join(program)
def cleanup():
    # ファイルをパースして構文木を取得
    index = Index.create()
    tu = index.parse(path="main.cpp", args=options, unsaved_files=[("main.cpp", "".join(str(line, 'utf-8') for line in program))])
    warnings = list(tu.diagnostics)
    if not warnings:
        return False

    # 未使用のマクロ，変数，関数を列挙する
    unused_macros = set()
    unused_variables = set()
    unused_functions = set()

    for w in warnings:
        match w.option:
            case "-Wunused-macros":
                unused_macros.add(w.location.line)
            case "-Wunused-variable" | "-Wunused-const-variable" | "-Wunused-local-typedefs" | "-Wunused-private-field":  # VAR_DECL | TYPE_ALIAS_DECL | FIELD_DECL
                unused_variables.add((w.location.line, w.location.column))
            case "-Wunused-function" | "-Wunneeded-internal-declaration" | "-Wunused-template" | "-Wunused-member-function" | "-Wunneeded-member-function":  # FUNCTION_DECL | FUNCTION_TEMPLATE | CXX_METHOD
                unused_functions.add((w.location.line, w.location.column))

    # 構文木を DFS, 未使用の関数・変数を削除する
    def dfs(node, depth=0):
        if node.location.file is not None and node.location.file.name != "main.cpp":
            return
        # print(f"{node.location.file} {node.extent.start.line:03}:{node.extent.start.column:02}-{node.extent.end.line:03}:{node.extent.end.column:02}\tdep={depth}\t{str(node.kind)[11:]:20}\t{node.displayname}", file=sys.stderr)

        match node.kind:
            case CursorKind.FUNCTION_DECL | CursorKind.FUNCTION_TEMPLATE | CursorKind.CXX_METHOD | CursorKind.CONSTRUCTOR | CursorKind.DESTRUCTOR:
                loc = node.location
                if (loc.line, loc.column) in unused_functions:
                    # print(f"[UNUSED FUNCTION] {node.extent.start.line:>3}:{node.extent.start.column:<3}-{node.extent.end.line:>4}:{node.extent.end.column:<3}\t{node.displayname}", file=sys.stderr)
                    erase(node.extent)
                    return
            case CursorKind.VAR_DECL | CursorKind.TYPE_ALIAS_DECL | CursorKind.FIELD_DECL:
                loc = node.location
                if (loc.line, loc.column) in unused_variables:
                    # print(f"[UNUSED VARIABLE] {node.extent.start.line:>3}:{node.extent.start.column:<3}-{node.extent.end.line:>4}:{node.extent.end.column:<3}\t{node.displayname}", file=sys.stderr)
                    erase(node.extent)
                    return
        for child in node.get_children():
            dfs(child, depth + 1)

    dfs(tu.cursor)

    # 未使用のマクロを削除する
    for l in unused_macros:
        l -= 1
        line = program[l]
        while len(line) >= 2 and line[-2] == ord('\\'):
            line[:] = b"\n"
            l += 1
            if l >= len(program):
                break
            line = program[l]
        line[:] = b"\n"
    
    # 標準入力から変更後のプログラムを入れ，コンパイルできるか確認
    program_ = b"".join(program)
    p = subprocess.run(["clang++", "-fsyntax-only", "-std=gnu++2b", "-x", "c++", "-"] + user_options, input=program_, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
    print(p.stderr.decode(), file=sys.stderr)
    if p.returncode:
        return False
    global program_fix
    if program_fix == program_:
        return False
    program_fix = program_
    return True

# 5 回まで cleanup を試行
for _ in range(5):
    if not cleanup():
        break

# clang-format で整形する
p = subprocess.Popen(["clang-format"], stdin=subprocess.PIPE, stdout=subprocess.PIPE)
stdout, _ = p.communicate(input=b"".join(program))

# ";" のみの行を削除
program = stdout.splitlines(keepends=True)
program = [line for line in program if line != b";\n"]

# 結果を出力
sys.stdout.buffer.write(b"".join(program))

使い方.md

注: 完全に自分用コードになっています

1. #include を一番上に
2. #include と main 関数以外を無名 namespace で囲む (unused の warning を出すため, global にあると他のコードとリンクして使われる可能性がある)
3. python cleanup.py <code>

依存関係

• clang (unused に関する Warning が豊富)
• gcc (コメントだけ削除するために -preprocessed を使っている)
• libclang (構文木を作ってくれる) : pip install libclang
• clang-format (最後にフォーマット)

options には clang++ -v -xc++ - で出てくるオプションから適当に取り出しています