I have a question about encoding, string comparison, and read() that is best expressed through an example. Here goes.

Suppose I am reading from a "binary" file. For this example let's use an IFF file since it's pretty familiar. Part of the process of decoding IFF involves checking the "Type ID" - a 4-byte indicator of chunk type, similar to a FourCC. Example type IDs might be "FORM", "LIST" etc. (but being 32 bits long, also quickly comparable to an int32 as 0x464f524d). Reading and doing something with the Type ID might look like this:

open(my $fp, '<:raw', $filename) or die "Couldn't open $filename: $!";

read($fp, my $type_id, 4);
if ($type_id eq 'FORM') {
  ...
}
[download]

Now my question is: is this string comparison always "safe"? In Python I think this would be an error, because it's comparing a "string" to what is technically a "bytearray". That language would force you to make an explicit conversion. In Perl, this is allowed, but I don't know if there are dangers. One option would be to "unpack" like so:

read($fp, my $buffer, 4);
my $type_id = unpack('A4', $buffer);
if ($type_id eq 'FORM') {
  ...
}
[download]

Now I've guaranteed that it is an ASCII string, but, have I really gained anything? Or is this overkill?

What's the encoding of a "string" read from a filehandle opened with :raw or changed with binmode()? What about the encoding of literal strings within my Perl script?