sub dec_bin2s {
        return unpack "B8", pack "c", shift;
}
[download]

This one is correct (as I'm sure I0 knows).

From perldoc -f pack:

c   A signed char value.
...
i   A signed integer value.
    (This 'integer' is at least 32 bits wide.  Its exact
    size depends on what a local C compiler calls 'int',
    and may even be larger than the 'long' described in
    the next item.)

The question asks specifically for signed 8-bit, which is "c", a signed char value (8 bits). "At least 32 bits" is not 8 bits.

print "$_\t", dec_bin2s($_), "\n" for -42 .. 42;

sub dec_bin2s {
   $dec = shift;
   return unpack "B8", (pack "i", $dec);
}
[download]

#!/usr/bin/perl
my $decimal=-33;
#How to generate:
#$decimal - your Perl variable. MUST be: -127<$decimal<127
my $packed=pack("N",$decimal); # Little-endian 4 byte.
my $YOUR8BITVALUE=substr($packed,-1); # last byte from $packed

open F,">your_test_file";
print F $YOUR8BITVALUE;
close F; 
#   Now you have file with your $decimal 
#   as one byte in "two's complement";


# How to get from file:
open F, "<your_test_file";
read(F,$packed,1); #read byte;
my $value; # result
if (unpack("N","\x00\x00\x00".$packed) & (1<<7)) { 
# sign bit is set. It's negative value
  $packed=("\xFF\xFF\xFF").$packed; # expand to int with 3 byte of 0xF
+F
  $value=-(unpack("N",~$packed)+1); # inverse all bits and unpack,incr
+ement,reverse sign.
} else {
  $packed=("\x00\x00\x00").$packed; # expand to int
  $value=unpack("N",$packed); unpack
}
print "$value\n"; # That's all!
[download]