ykxmz/_tests/check_encoding.py

# -*- coding: utf-8 -*-
r'''检查 .java / .py 文件的编码和换行符约定。

项目约定：
    .java / .py   UTF-8 **无 BOM**、**LF** 换行
    .ps1          UTF-8 **带 BOM**（PowerShell 5 读中文需要）
    .bat          GBK + CRLF

用法:
    python _tests/check_encoding.py [文件或目录 ...]   # 不给参数就查全仓库
'''
from __future__ import annotations

import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
BOM = b"\xef\xbb\xbf"

# 默认检查范围：局域网这条线上**我们负责的文件**。
#
# 为什么不全仓库默认：
#   仓库里有十几处历史 CRLF 欠账（_build/ 的一次性脚本、
#   _tests/java/*.java、几个早期测试）。它们和局域网无关，
#   一次性归一会产生"整文件改写"的巨型 diff，反而淹没真正的改动。
#   所以默认只盯自己这条线，**新引入**的 CRLF 一律算失败；
#   想看全仓库用 --all（历史欠账只提醒，不判失败）。
SCOPE = (
    "windows/lanremote/discovery.py",
    "windows/lanremote/instance.py",
    "windows/lanremote/server.py",
    "android/src/com/lanremote/watch",
    "_tests/test_lan_detect.py",
    "_tests/test_lan_path.py",
    "_tests/test_lan_optimize.py",
    "_tests/test_watch_lan_switch.py",
    "_tests/test_watch_relayonly.py",
    "_tests/test_watch_skeleton.py",
    "_tests/bench_lan_vs_relay.py",
    "_tests/check_encoding.py",
    "_tests/verify_all.ps1",
)

# 期望的编码/行尾规则（见文件头说明）
RULES = {
    ".java": (False, False, "utf-8"),
    ".py": (False, False, "utf-8"),
    ".ps1": (True, False, "utf-8"),
    ".bat": (False, True, "gbk"),
}

SKIP_DIRS = {".git", "dist", "dist_exe", "build", "toolchain", "__pycache__",
             "site_shots", "downloads", "_apk_build", "out_java",
             "out_java_bench"}

# 已经存在的行尾欠账（**不是**这次改出来的）。
# 这些文件早就是 CRLF 了，而且不在局域网这条线上。
# 一次性把它们归一会产生"整文件改写"的巨型 diff，有害无益；
# 但也不能假装看不见 —— 所以单独列出来报成提醒，
# 而**新引入**的 CRLF 一律算失败。
PREEXISTING_CRLF = {
    "windows/lanremote/actions.py",
    "windows/lanremote/diagnostics.py",
    "windows/lanremote/headless.py",
    "windows/lanremote/protocol.py",
    "_tests/test_features.py",
    "_tests/test_indicator_visible.py",
    "_tests/test_selfcheck.py",
    "_tests/test_undo.py",
}


def check_file(path: Path) -> tuple:
    """返回 (致命问题列表, 提醒列表)。"""
    rule = RULES.get(path.suffix.lower())
    if rule is None:
        return [], []
    want_bom, want_crlf, encoding = rule
    try:
        raw = path.read_bytes()
    except OSError as exc:
        return ["读不到: %s" % exc], []

    fatal = []
    note = []
    try:
        rel = path.relative_to(ROOT).as_posix()
    except ValueError:
        rel = path.as_posix()
    known = rel in PREEXISTING_CRLF

    has_bom = raw.startswith(BOM)
    if has_bom != want_bom:
        message = ("BOM 应为%s，实际%s"
                   % ("有" if want_bom else "无", "有" if has_bom else "无"))
        (note if known else fatal).append(message)

    body = raw[len(BOM):] if has_bom else raw
    crlf = body.count(b"\r\n")
    lone_cr = body.count(b"\r") - crlf
    if want_crlf and crlf == 0:
        (note if known else fatal).append("应为 CRLF 换行，实际是 LF")
    if not want_crlf and crlf:
        message = "有 %d 处 CRLF（应为 LF）" % crlf
        (note if known else fatal).append(message)
    if lone_cr:
        fatal.append("有 %d 处单独的 CR" % lone_cr)

    # 编码必须是声明的那种
    try:
        body.decode(encoding)
    except (UnicodeDecodeError, LookupError) as exc:
        fatal.append("不是合法 %s: %s" % (encoding.upper(), exc))

    return fatal, note


def iter_targets(args: list) -> list:
    if not args:
        # 默认：只查 SCOPE 里的文件
        return [p for p in _resolve(SCOPE) if p.is_file()]
    if args == ["--all"]:
        found = []
        for path in ROOT.rglob("*"):
            if not path.is_file():
                continue
            if any(part in SKIP_DIRS for part in path.parts):
                continue
            if path.suffix.lower() in RULES:
                found.append(path)
        return found
    return _resolve(args)


def _resolve(items: list) -> list:
    out = []
    for arg in items:
        path = Path(arg)
        if not path.is_absolute():
            path = ROOT / path
        if path.is_dir():
            out.extend(p for p in path.rglob("*")
                       if p.is_file() and p.suffix.lower() in RULES)
        elif path.is_file():
            out.append(path)
    return out


def main() -> int:
    targets = iter_targets(sys.argv[1:])
    bad = 0
    noted = 0
    for path in sorted(targets):
        fatal, note = check_file(path)
        try:
            shown = path.relative_to(ROOT)
        except ValueError:
            shown = path
        if fatal:
            bad += 1
            print("  [FAIL] %s  %s" % (shown, "; ".join(fatal)))
        if note:
            noted += 1
            print("  [提醒] %s  %s（早就这样，不属于本次改动）"
                  % (shown, "; ".join(note)))
    print()
    if noted:
        print("  （%d 个文件有历史行尾欠账，已单独列出，不算失败）" % noted)
    if bad:
        print("  检查 %d 个文件，%d 个不符合约定" % (len(targets), bad))
        return 1
    print("  检查 %d 个文件，编码/换行全部符合约定" % len(targets))
    return 0


if __name__ == "__main__":
    raise SystemExit(main())