mirror/mirror-shim/src/utils.rs

use mlua::LuaString;
use std::ffi::OsString;

/// 将 Lua 字符串转为 Rust 字符串；非 UTF-8 按系统默认的 ANSI/OEM (如 GBK) 进行安全解码
pub fn lua_string_2_os_string(s: &LuaString) -> mlua::Result<OsString> {
    let raw_bytes = &s.as_bytes();
    // 1. Unix 平台：直接零拷贝透传原始字节（无损支持任意编码）
    #[cfg(unix)]
    {
        use std::os::unix::ffi::OsStrExt;
        Ok(OsStr::from_bytes(raw_bytes).to_os_string())
    }
    // 2. Windows 平台：优先按 UTF-8 解码，失败则按当前系统本地代码页 (ANSI/GBK) 转换
    #[cfg(windows)]
    {
        // 先尝试标准的 UTF-8
        if let Ok(utf8_str) = std::str::from_utf8(raw_bytes) {
            return Ok(OsString::from(utf8_str));
        }

        unsafe {
            use std::os::windows::ffi::OsStringExt;
            use windows_sys::Win32::Globalization::{
                CP_ACP, MB_ERR_INVALID_CHARS, MultiByteToWideChar,
            };

            if raw_bytes.is_empty() {
                return Ok(OsString::new());
            }

            let len = MultiByteToWideChar(
                CP_ACP,
                MB_ERR_INVALID_CHARS,
                raw_bytes.as_ptr(),
                raw_bytes.len() as i32,
                std::ptr::null_mut(),
                0,
            );

            if len <= 0 {
                // 构造一个带上下文的 FromLua 转换错误，便于定位配置问题
                return Err(mlua::Error::FromLuaConversionError {
                    from: "LuaString",
                    to: "OsString".to_string(),
                    message: Some(
                        format!("字符串{:?}包含无效或当前系统无法识别的编码字节", s).to_string(),
                    ),
                });
            }

            let mut buf = vec![0u16; len as usize];
            MultiByteToWideChar(
                CP_ACP,
                MB_ERR_INVALID_CHARS,
                raw_bytes.as_ptr(),
                raw_bytes.len() as i32,
                buf.as_mut_ptr(),
                len,
            );

            Ok(OsString::from_wide(&buf))
        }
    }
}

/// 将输入的字符串按 Shell 规则切分为独立的 CLI 参数 Token
/// - 自动过滤连续空格
/// - 支持单引号 `'...'` 和双引号 `"..."` 包裹包含空格的参数
pub fn parse_tokens(input: &str) -> Vec<OsString> {
    let mut tokens = Vec::new();
    let mut current_token = String::new();
    let mut in_quote: Option<char> = None;
    let mut chars = input.chars().peekable();

    while let Some(ch) = chars.next() {
        match (ch, in_quote) {
            // 处理转义字符 (例如 \")
            ('\\', _quote) => {
                if let Some(&next_ch) = chars.peek() {
                    let should_escape = if cfg!(windows) {
                        // Windows 策略：只有在转义引号、反斜杠本身时才剥离 \
                        // （如果在双引号内部，空格也不应该被 \ 转义）
                        next_ch == '"' || next_ch == '\'' || next_ch == '\\'
                    } else {
                        // Unix 策略：标准 Shell 转义（引号、反斜杠、空格等）
                        next_ch == '"'
                            || next_ch == '\''
                            || next_ch == '\\'
                            || next_ch.is_whitespace()
                    };

                    if should_escape {
                        chars.next(); // 消耗掉下一个字符
                        current_token.push(next_ch);
                    } else {
                        // 保留 Windows 路径分隔符或未知转义中的 \
                        current_token.push('\\');
                    }
                } else {
                    // 结尾孤立的 \
                    current_token.push('\\');
                }
            }
            // 遇到引号：开启或关闭引号包裹
            ('"' | '\'', None) => {
                in_quote = Some(ch);
            }
            ('"' | '\'', Some(q)) if q == ch => {
                in_quote = None;
            }
            // 引号外部遇到空白字符：切分出一个完整的 Token
            (ch, None) if ch.is_whitespace() => {
                if !current_token.is_empty() {
                    tokens.push(OsString::from(std::mem::take(&mut current_token)));
                }
            }
            // 其他字符或引号内部字符：直接追加
            (ch, _) => {
                current_token.push(ch);
            }
        }
    }

    // 收尾最后一个 Token
    if !current_token.is_empty() {
        tokens.push(OsString::from(current_token));
    }

    tokens
}

#[cfg(test)]
mod tests {
    use super::*;
    use std::ffi::OsString;
    /// 辅助宏：简化声明与断言对比
    macro_rules! assert_tokens {
        ($input:expr, $expected:expr) => {
            let actual = parse_tokens($input);
            let expected_os: Vec<OsString> = $expected.into_iter().map(OsString::from).collect();
            assert_eq!(
                actual, expected_os,
                "\n测试输入: {:?}\n期望输出: {:?}\n实际输出: {:?}",
                $input, expected_os, actual
            );
        };
    }
    #[test]
    fn test_parse_tokens_basic_split() {
        // 场景 1：基础多参数拆分（空格分隔）
        assert_tokens!("cargo run --verbose", vec!["cargo", "run", "--verbose"]);
        assert_tokens!("git status", vec!["git", "status"]);
    }

    #[test]
    fn test_parse_tokens_multi_alias_ref() {
        // 场景 2：多别名混合与组合引用
        assert_tokens!(
            "mr:run --bin mr:base_flags",
            vec!["mr:run", "--bin", "mr:base_flags"]
        );
        assert_tokens!(
            "mr:app1 mr:app2 --flag",
            vec!["mr:app1", "mr:app2", "--flag"]
        );
    }

    #[test]
    fn test_parse_tokens_continuous_whitespaces() {
        // 场景 3：连续多空格与制表符过滤
        assert_tokens!(
            "mr:run    --bin \t  my_app",
            vec!["mr:run", "--bin", "my_app"]
        );
        assert_tokens!("   cargo   build   ", vec!["cargo", "build"]);
    }

    #[test]
    fn test_parse_tokens_double_quotes() {
        // 场景 4：双引号包裹包含空格的参数
        assert_tokens!(
            "git commit -m \"fix a bug\"",
            vec!["git", "commit", "-m", "fix a bug"]
        );
        assert_tokens!("echo \"hello world\"", vec!["echo", "hello world"]);
    }

    #[test]
    fn test_parse_tokens_single_quotes() {
        // 场景 5：单引号包裹包含空格的参数
        assert_tokens!(
            "gcc -O2 'my file.c' -o app",
            vec!["gcc", "-O2", "my file.c", "-o", "app"]
        );
        assert_tokens!(
            "python 'script with space.py'",
            vec!["python", "script with space.py"]
        );
    }

    #[test]
    fn test_parse_tokens_escaped_characters() {
        // 场景 6：反斜杠转义字符
        assert_tokens!("echo hello\\ world", vec!["echo", "hello\\", "world"]);
        assert_tokens!(
            "echo \"hello \\\"world\\\"\"",
            vec!["echo", "hello \"world\""]
        );
    }

    #[test]
    fn test_parse_tokens_single_scalar_and_edge_cases() {
        // 场景 7：单标量参数与边界情况
        assert_tokens!("git", vec!["git"]);
        assert_tokens!("8000", vec!["8000"]);
        assert_tokens!("", Vec::<&str>::new());
        assert_tokens!("    ", Vec::<&str>::new());
    }

    #[test]
    fn test_parse_tokens_unclosed_quotes() {
        // 场景 8：未闭合引号的容错处理（会尽量追加到当前 Token 中）
        assert_tokens!("echo \"hello world", vec!["echo", "hello world"]);
    }
}