Perl glob未匹配到预期文件问题排查求助

问题说明 #

使用Perl脚本读取new.txt中的每行路径，通过glob匹配对应路径下的wint_nightly_nfarm_*.main.20241216.4262507文件，再提取文件名中的bankName，但new.txt第二行始终无法正确匹配，输出为空，需解决该问题以得到预期结果。

现有代码 #

my $bankNameTest;
my $bankName;
my @lines;

my $document = do {
    local $/ = undef;
    open my $fh, "<", "new.txt"
        or die "could not open $file: $!";
    <$fh>;
};

chomp ($document);
print "$document is doc\n";
@lines = split (/
/,$document);


foreach my $test (@lines) {
    $bankNameTest=glob ("${test}/wint_nightly_nfarm_*.main.20241216.4262507") ;
    chomp ($bankNameTest);
    print "$bankNameTest is testbankname\n";
    $bankName=`ls -d $bankNameTest | sed -r "s?${var}/??g" | sed -r "s/wint_nightly_nfarm_//g" | sed -r "s/.main.20241216.4262507//g"`;
    chomp ($bankName);
    print "$bankName is the bankName\n";
}

new.txt内容 #

_qa/SERVFARM/CAPTURE_DB/CaptureViewer/2487964_brd_launch/cap_allegro_view
_qa/SERVFARM/CAPTURE_DB/CaptureViewer/FIND/enh_find/DRC_MARKERS/hierdsn_incomp_entry
_qa/SERVFARM/CAPTURE_DB/CaptureViewer/crossprobe/03_between_cap_allegroviewer/dsn_schm_page

当前输出 #

_qa/SERVFARM/CAPTURE_DB/CaptureViewer/2487964_brd_launch/cap_allegro_view/wint_nightly_nfarm_capture_viewer.main.20241216.4262507 is testbankname
capture_viewer is the bankName
 is testbankname
 is the bankName
_qa/SERVFARM/CAPTURE_DB/CaptureViewer/crossprobe/03_between_cap_allegroviewer/dsn_schm_page_and_part_name_with_special_chars/wint_nightly_nfarm_capture_viewer.main.20241216.4262507 is testbankname
capture_viewer is the bankName

预期输出 #

_qa/SERVFARM/CAPTURE_DB/CaptureViewer/2487964_brd_launch/cap_allegro_view/wint_nightly_nfarm_capture_viewer.main.20241216.4262507 is testbankname
capture_viewer is the bankName
_qa/SERVFARM/CAPTURE_DB/CaptureViewer/FIND/enh_find/DRC_MARKERS/hierdsn_incomp_entry/wint_nightly_nfarm_capture_viewer.main.20241216.4262507 is testbankname
capture_viewer is the bankName
_qa/SERVFARM/CAPTURE_DB/CaptureViewer/crossprobe/03_between_cap_allegroviewer/dsn_schm_page_and_part_name_with_special_chars/wint_nightly_nfarm_capture_viewer.main.20241216.4262507 is testbankname
capture_viewer is the bankName

问题根源与解决方案 #

问题1：换行符处理错误 #

原脚本一次性读取整个文件后用硬编码换行符拆分，且仅对整个文档做一次chomp，若文件包含Windows风格的\r\n换行，chomp只会去掉\n，导致部分行末尾残留\r，glob无法匹配带\r的路径。

问题2：未定义变量$var #

脚本中使用了未定义的$var变量，导致sed命令执行异常，影响bankName提取。

问题3：不必要的系统调用 #

使用ls和sed等系统命令处理字符串，既低效又存在shell注入风险，完全可以用Perl内置字符串操作替代。

修改后的代码 #

use strict;
use warnings;

my $target_pattern = 'wint_nightly_nfarm_*.main.20241216.4262507';

open my $fh, '<', 'new.txt' or die "无法打开new.txt: $!";
while (my $test = <$fh>) {
    chomp $test;  # 去掉每行末尾的所有换行符（包括\r\n或\n）
    next if $test =~ /^\s*$/;  # 跳过空行

    my ($bankNameTest) = glob("$test/$target_pattern");
    if ($bankNameTest) {
        print "$bankNameTest is testbankname\n";
        # 用Perl正则提取bankName，替代系统调用
        if ($bankNameTest =~ m{wint_nightly_nfarm_(.*)\.main\.20241216\.4262507$}) {
            my $bankName = $1;
            print "$bankName is the bankName\n";
        } else {
            print "无法提取bankName: $bankNameTest\n";
        }
    } else {
        print "未找到匹配文件: $test/$target_pattern\n";
    }
}
close $fh;

代码说明 #

1. 启用strict和warnings，提前发现语法和变量问题。
2. 逐行读取文件，每行单独chomp，彻底处理不同系统的换行符。
3. 用Perl正则直接从匹配到的文件路径中提取bankName，避免系统调用。
4. 增加空行判断和错误提示，提升脚本健壮性。

内容的提问来源于stack exchange，提问作者AskQuestionsP