It's not really your unpacks that are messy.

• The comments that parrot the obvious.
• Concatenation is used when interpolation or printf would be clearer.
• Some var names are needlessly long. ("load" doesn't add value.)
• Each record's size and crc are printed, but that's not data.

But your unpacks could be improved by treating the records as follows the following format:

header        body       footer
------------  ---------  ------
id body_size  size path  crc32
[download]

How I'd write it:

#!/usr/bin/perl

use strict;
use warnings;

my $mem = pack('H*', '54424c004100001300000BAD2F62696E2F7465737432302E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');

(my $section, $mem) = unpack('Z* a*', $mem);

print("$section\n");
print( ( "-" x length($section) ), "\n");

while (length($mem)) {
    (my $id, my $body, my $crc, $mem) = unpack('n n/a N a*', $mem);
    my ($size, $path) = unpack('N a*', $body);

    # Check CRC here.

    print("Load ID   : $id\n");
    print("Load Size : $size\n");
    print("Load Path : $path\n");
    print("\n");
}
[download]

TBL
---
Load ID   : 16640
Load Size : 2989
Load Path : /bin/test20.dfl

Load ID   : 16896
Load Size : 4
Load Path : /bin/test01.dfl

Load ID   : 17152
Load Size : 4
Load Path : /bin/test02.dfl

...
[download]

Update: Added enumeration at the top.

Thank you very much for your great answer. I think it is a good way of learning to try it first time by myself and then looking at a better solution.

Very good idea from you to use 'a*' to read the rest of the memory and shorten it by this way instead of using substr. Also interesting that you were just interpreting the offset field as the length of a body. So the length is directly in front of the body. And another small thing I already knew but did not use. The 'x' operator to underline the string.

So I learnt a lot from your post. Thank you.

Hello,

It's me again. Now I wrote a new version of the script. The first unpack is getting the length of the path. The seoncd unpack is reading one repetition of data in one swoop. I like this solution because it matches the original record the most. Although I have in mind that I can use the length/string technique in the post of ikegami if the length is directly before the string.

#!/usr/bin/perl

use strict;
use warnings;

my $mem = pack('H*', '54424c004100001300000BAD2F62696E2F7465737432302E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');


# NOTE:
# Section name is a null terminated string. 
# Z* is setting the memory pointer after the 
# terminating null byte, but in the corresponding 
# variable the string is stored without the null
(my $section, $mem) = unpack('Z* a*', $mem);
print("$section\n");
print("-" x length($section), "\n\n");

# unpack and print 
# load id, offset, size, 
# path (length of path is "offset - 4") 
# and crc32 of each load entry
while (length($mem))
{
    # NOTE:
    # The path is a variable string. And the length 
    # of the string is not directly before the string.
    # If the length of the string would be directly 
    # before the string then I could use the 
    # length/string technique. But in this case I have 
    # to use two unpacks. The first is to get the length 
    # of the string (offset_to_next_load - 4). The second 
    # unpack then has all information to read one 
    # repetition of data in one swoop.

    # get length of path
    my $length_of_load_path = unpack('x2 n', $mem) - 4;

    # get data
    (my $load_id, my $offset_to_next_load, 
     my $load_size, my $load_path, my $crc32_load_path, $mem) =
        unpack("n n N A$length_of_load_path N a*", $mem);

    print("Load ID             : $load_id\n");
    print("Offset to next load : $offset_to_next_load\n");
    print("Load Size           : $load_size\n");
    print("Load Path           : $load_path\n");
    print("CRC32 of Load Path  : $crc32_load_path\n\n");
    
}
[download]