Windows下Perl脚本传递UTF-8参数至系统命令的问题

1. 无需systemW()的跨平台UTF-8系统调用方案 #

核心是针对Windows和Linux分别处理编码逻辑，绕过Perl默认system()的编码转换问题：

• Linux环境：直接用原生system()即可，多数Linux发行版默认UTF-8环境，传递UTF-8字符串不会有编码冲突。
• Windows环境：
  • 方案一：用Win32::API直接调用系统原生的CreateProcessW API，手动处理宽字符命令行。先把Perl中的UTF-8字符串转成字节流，再传递给宽字符API，避免双重编码。代码示例：

    use strict;
    use warnings;
    use utf8;
    use if $^O eq 'MSWin32', 'Win32::API';
    
    sub cross_platform_system {
        my ($cmd) = @_;
        return system($cmd) unless $^O eq 'MSWin32';
    
        # 转UTF-8字符串为字节流，适配宽字符API
        utf8::encode($cmd);
        my $CreateProcessW = Win32::API->new(
            'kernel32', 'CreateProcessW',
            ['P', 'P', 'P', 'P', 'I', 'I', 'P', 'P', 'P', 'P'], 'N'
        );
        my ($si, $pi) = ("\x00" x 28, "\x00" x 16);
        $CreateProcessW->Call(undef, $cmd, undef, undef, 0, 0, undef, undef, $si, $pi);
    
        # 等待进程结束并获取退出码
        my $WaitForSingleObject = Win32::API->new('kernel32', 'WaitForSingleObject', ['N', 'N'], 'N');
        $WaitForSingleObject->Call(unpack('N', substr($pi, 8, 4)), 0xFFFFFFFF);
        my $GetExitCodeProcess = Win32::API->new('kernel32', 'GetExitCodeProcess', ['N', 'P'], 'I');
        my $exit_code = "\x00" x 4;
        $GetExitCodeProcess->Call(unpack('N', substr($pi, 8, 4)), $exit_code);
        return unpack('N', $exit_code);
    }
    

  • 方案二：用IPC::Run3模块，它会自动处理Windows下的宽字符命令行，跨平台兼容性更强，代码更简洁：

    use strict;
    use warnings;
    use utf8;
    use IPC::Run3;
    
    sub cross_platform_system {
        my ($cmd) = @_;
        run3($cmd, \undef, \undef, \undef);
        return $? >> 8;
    }
    

2. 实现类似systemW()的反引号输出捕获 #

要解决Windows下反引号捕获UTF-8输出的问题，同样需要针对平台处理：

• Linux环境：直接用原生反引号或qx//即可，输出默认是UTF-8，直接处理就行。
• Windows环境：
  • 方案一：调用CreateProcessW并重定向标准输出到管道，读取宽字符输出后转成UTF-8。代码示例：

    use strict;
    use warnings;
    use utf8;
    use if $^O eq 'MSWin32', 'Win32::API';
    use Encode qw(decode_utf16le);
    
    sub cross_platform_backtick {
        my ($cmd) = @_;
        return `$cmd` unless $^O eq 'MSWin32';
    
        utf8::encode($cmd);
        # 创建匿名管道用于捕获输出
        my ($hRead, $hWrite);
        my $CreatePipe = Win32::API->new('kernel32', 'CreatePipe', ['P', 'P', 'P', 'N'], 'I');
        my $sa = "\x00" x 12;
        $CreatePipe->Call(\$hRead, \$hWrite, $sa, 0);
    
        # 设置进程重定向stdout
        my $si = "\x00" x 28;
        substr($si, 20, 4) = pack('L', $hWrite);
        substr($si, 8, 4) = pack('L', 1); # 启用标准句柄重定向
        my $pi = "\x00" x 16;
        my $CreateProcessW = Win32::API->new(
            'kernel32', 'CreateProcessW',
            ['P', 'P', 'P', 'P', 'I', 'I', 'P', 'P', 'P', 'P'], 'N'
        );
        $CreateProcessW->Call(undef, $cmd, undef, undef, 1, 0, undef, undef, $si, $pi);
    
        # 关闭管道写端，读取输出
        my $CloseHandle = Win32::API->new('kernel32', 'CloseHandle', ['N'], 'I');
        $CloseHandle->Call($hWrite);
        my $buffer = '';
        my $bytes_read = "\x00" x 4;
        my $ReadFile = Win32::API->new('kernel32', 'ReadFile', ['N', 'P', 'N', 'P', 'P'], 'I');
        while ($ReadFile->Call($hRead, my $chunk, 4096, $bytes_read, undef)) {
            my $len = unpack('L', $bytes_read);
            last if $len == 0;
            $buffer .= substr($chunk, 0, $len);
        }
    
        # 把宽字节输出转成UTF-8
        my $utf8_output = decode_utf16le($buffer);
        utf8::decode($utf8_output);
    
        # 清理资源
        $CloseHandle->Call($hRead);
        $CloseHandle->Call(unpack('L', substr($pi, 8, 4)));
        $CloseHandle->Call(unpack('L', substr($pi, 0, 4)));
        return $utf8_output;
    }
    

  • 方案二：用IPC::Run模块，自动处理跨平台的输出捕获和编码转换，代码更简洁：

    use strict;
    use warnings;
    use utf8;
    use IPC::Run qw(run);
    
    sub cross_platform_backtick {
        my ($cmd) = @_;
        my $output;
        run($cmd, \undef, \$output);
        return $output;
    }
    

内容的提问来源于stack exchange，提问作者Paolo