Perl脚本A替换脚本B数组时匹配异常及数组标识未修改问题求助

Let's break down why your current approach isn't working and fix it properly:

What's Wrong with Your Current Code? #

You're hitting three key issues that are causing unintended behavior:

1. Unresolved Variable in Shell Command
   Your perl -pi -e command is wrapped in single quotes, which means the $list variable from Script A never gets expanded by your main Perl process—it’s treated as a literal, undefined variable in the one-off Perl instance you’re spawning. That’s why your new array values aren’t showing up, and the replacement logic goes haywire.

2. Overly Greedy & Unescaped Regex
   The ^my \@old_array.* regex uses a greedy .* that can match beyond the end of the array line. Worse, the @ symbol is a special character in shells—if your escaping isn’t perfect, it might get interpreted as a shell variable, turning your regex into something that matches almost any line starting with my.

3. Trailing Comma Syntax Error
   Your loop to build $list adds an extra trailing comma (e.g., value1, value2, ), which will break Perl syntax in Script B once replaced.

Solution 1: Fix the Shell Command (Quick Fix) #

Adjust the quoting and regex to properly expand $list and target only the exact array line:

# Script A code
my @new_values = ('new_val1', 'new_val2', 'new_val3');
# Use join to avoid trailing commas (cleaner than looping)
my $list = join(', ', @new_values);

# Use double-quoted qq{} to expand $list, with proper escaping for Perl/shell
my $replace_cmd = qq{perl -pi -e 's/^my \\@old_array\\s*=\\s*\\(.*?\\);/my \\@new_array = ( $list );/' script_B.pl};
system($replace_cmd);

Key Fixes: #

• join(', ', @new_values) builds a valid list without extra trailing commas.
• qq{} lets $list expand while keeping Perl regex characters (like @ and ()) properly escaped.
• The regex uses \s* to allow whitespace around = and .*? (non-greedy match) to only capture content inside the array parentheses, preventing over-matching.

Solution 2: Modify Script B Directly (More Reliable) #

Calling an external Perl process can be fragile due to shell escaping. A better approach is to have Script A directly edit Script B:

# Script A code
my @new_values = ('new_val1', 'new_val2', 'new_val3');
my $new_array_line = 'my @new_array = (' . join(', ', @new_values) . ');';

# Open Script B for reading and a temp file for writing
open my $in_fh, '<', 'script_B.pl' or die "Can't open script_B.pl: $!";
open my $out_fh, '>', 'script_B.pl.tmp' or die "Can't create temp file: $!";

while (my $line = <$in_fh>) {
    # Replace only the exact old array line
    $line =~ s/^my \@old_array\s*=\s*\(.*?\);/$new_array_line/;
    print $out_fh $line;
}

# Swap the temp file with the original
close $in_fh;
close $out_fh;
rename 'script_B.pl.tmp', 'script_B.pl' or die "Can't update script_B.pl: $!";

Why This Is Better: #

• No shell escaping headaches—everything stays within Perl.
• Total control over which lines get replaced (no accidental matches).
• Easier to debug: you can check the temp file before overwriting the original.

内容的提问来源于stack exchange，提问作者Zero