Perl中local修改$/后FH3为何读取整文件？如何实现逐行读取？

给定文件/serrs/env.txt内容：

abc
def
ghi jkl mno
good is the conet

使用以下Perl代码读取文件：

use strict;
use warnings;

my $ModeRefFile;
open(FH2, '<', "/home/vivi/ee/ppp.pl") or die $!;   
local $/ = undef;
$ModeRefFile = <FH2>;
close(FH2);

open(FH3, '<', "/serrs/env.txt") or die $!;
while (<FH3>)   {
     chomp ($_);
     print "$_ is the line which is being read\n";
}
close(FH3);

实际输出：

abc
def
ghi jkl mno
good is the conet is the line which is being read

预期输出：

abc is the line which is being read
def is the line which is being read
ghi jkl mno is the line which is being read
good is the conet is the line which is being read

疑问：为什么第二个文件句柄FH3读取时，$_存储了整个文件内容？明明第一个FH2读取时用local修改的$/（输入记录分隔符），按理解local会恢复原有值，FH3应该能逐行读取，$_每次存一行。请问如何实现预期的逐行输出效果？

原因 #

你对local $/的作用域理解有误。local的作用域是当前代码块，但你修改$/的代码并未放在单独代码块中，而是直接在主程序流执行。这意味着local $/ = undef的作用范围会持续到程序末尾，后续FH3读取时$/仍为undef，导致一次性读取整个文件。

解决方法 #

方法一：用代码块限制local $/的作用域

把修改$/和读取FH2的代码包裹在单独代码块内，离开代码块后$/会自动恢复为默认换行符\n，这是更Perl化的写法：

use strict;
use warnings;

my $ModeRefFile;
# 代码块限制local作用域
{
    open(FH2, '<', "/home/vivi/ee/ppp.pl") or die $!;   
    local $/ = undef;
    $ModeRefFile = <FH2>;
    close(FH2);
}

open(FH3, '<', "/serrs/env.txt") or die $!;
while (<FH3>)   {
     chomp ($_);
     print "$_ is the line which is being read\n";
}
close(FH3);

方法二：手动恢复$/的值

在FH3读取前手动将$/设回默认换行符：

use strict;
use warnings;

my $ModeRefFile;
open(FH2, '<', "/home/vivi/ee/ppp.pl") or die $!;   
local $/ = undef;
$ModeRefFile = <FH2>;
close(FH2);

# 手动恢复输入记录分隔符
$/ = "\n";

open(FH3, '<', "/serrs/env.txt") or die $!;
while (<FH3>)   {
     chomp ($_);
     print "$_ is the line which is being read\n";
}
close(FH3);

内容的提问来源于stack exchange，提问作者PPP