How to output all lines except the last one

I got the idea for this article when I saw in one of the blogs that this problem is proposed to be solved with the following one-line code in Perl:

 echo -e '1\n2\n3' | perl -nalE 'push(@lines, $_); }{ say $lines[$_] foreach 0..$#lines -1'

What about the head utility, you ask me? Yes, this is the best option. But as it turned out, in some systems, you can't use this utility with a negative value for the n parameter (if you don't know how many lines are in the file). What was said in the original article before offering this solution.

But I think there are two problems with this code snippet:

Two passes are made through the data. The first is when we store each row in the @lines array. And the second, when we output the necessary lines
The entire file is stored in memory

So I decided to offer my own solution, which has no such problems:

 echo -e '1\n2\n3' | perl -nale 'push(@l, $_); print shift(@l) if @l> 1;'

The code is quite simple, but let's look at it in more detail. First, we do not store the entire file in memory, but only the number of lines that we do not want to output - a kind of buffer. Secondly, we immediately decide whether to output a string or not. The number of lines that we don't output from the end is set as a number. In this example, this is a single line.

I tested the memory consumption measurements on a 72m file

 ls -lah big.file
 -rw-r--r-- 1 madskill madskill 72M ноя  1 03:56 big.file

For the first option, got the following values

 PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 23400 madskill  20   0  205844 188952   3960 S   0,0  4,8   0:07.44 perl

And for my option the following

 PID USER      PR  NI    VIRT    RES    SHR S  %CPU %MEM     TIME+ COMMAND
 23413 madskill  20   0   21056   3492   3248 S   0,0  0,1   0:10.71 perl

From the above measurements, you can see that my version uses memory more economically. But I didn't stop there and then roughly compared both variations in performance. It turned out that there is almost no difference. With repeated launches, one or the other option was faster. And the difference between them was very small.

 time cat big.file | perl -nalE 'push(@lines, $_); }{ say $lines[$_] foreach 0..$#lines -1;'> /dev/null
 real	0m1,789s
 user	0m1,662s
 sys	0m0,188s
 time cat big.file | perl -nale 'push(@l, $_); print shift(@l) if @l> 1;'> /dev/null
 real	0m1,732s
 user	0m1,688s
 sys	0m0,116s