Re: Fast way to check if a sort was done (using stringification is *broken*)

Stringifying the array can lead to incorrect results in some cases. Consider the following:

use strict;
use warnings;
use Digest::MD5 qw( md5 );

my @a1 = ( 'aa', 'a', '' );
print '[', join( '|', @a1 ), "]\n";
my $d1 = md5( @a1 );

my @sorted_a1 = sort @a1;
print '[', join( '|', @sorted_a1 ), "]\n";
my $sorted_d1 = md5( @sorted_a1 );

printf( "Digests are %s\n",
        ( $d1 eq $sorted_d1 ) ? 'equal' : 'unequal' );
[download]

Prints:

[aa|a|]
[|a|aa]
Digests are equal
[download]

In this case, the elements in the array were completely reversed by the sort but the md5 digest was identical.

This is (partly) documented in the docs for Digest::MD5:

The result of md5("a", "b", "c") will be exactly the same as the result of md5("abc").

Comment on Re: Fast way to check if a sort was done (using stringification is broken) Select or Download Code

Replies are listed 'Best First'.
Re^2: Fast way to check if a sort was done (using stringification is broken) by ikegami (Patriarch) on Jun 29, 2007 at 03:32 UTC
`my $signature = md5(pack('(J/a)', @array));` [download] should do the trick quickly. Update: Of course, `md5` really doesn't add anything. You could just use `my $signature = pack('(J/a)', @array);` [download]	[reply] [d/l] [select]
Re^3: Fast way to check if a sort was done (using stringification is broken) by bobf (Monsignor) on Jun 29, 2007 at 04:13 UTC
Nice solution! I didn't know about '/' in pack. It's not quite perfect, though. It treats the empty string and undef as equal (there may be more ways to break it - this is just the first one I thought of): `my $signature_1 = pack( '(J/a)', '', undef ); my $signature_2 = pack( '(J/a)', undef, '' ); printf( "Signatures are %s\n", ( $signature_1 eq $signature_2 ) ? 'equal' : 'unequal' ); # eq +ual!` [download] These cases were not specified in the OP. I apologize for being pedantic. :-)	[reply] [d/l]

Replies are listed 'Best First'.
Re^2: Fast way to check if a sort was done (using stringification is broken) by ikegami (Patriarch) on Jun 29, 2007 at 03:32 UTC
`my $signature = md5(pack('(J/a)', @array));` [download] should do the trick quickly. Update: Of course, `md5` really doesn't add anything. You could just use `my $signature = pack('(J/a)', @array);` [download]	[reply] [d/l] [select]
Re^3: Fast way to check if a sort was done (using stringification is broken) by bobf (Monsignor) on Jun 29, 2007 at 04:13 UTC
Nice solution! I didn't know about '/' in pack. It's not quite perfect, though. It treats the empty string and undef as equal (there may be more ways to break it - this is just the first one I thought of): `my $signature_1 = pack( '(J/a)', '', undef ); my $signature_2 = pack( '(J/a)', undef, '' ); printf( "Signatures are %s\n", ( $signature_1 eq $signature_2 ) ? 'equal' : 'unequal' ); # eq +ual!` [download] These cases were not specified in the OP. I apologize for being pedantic. :-)	[reply] [d/l]