Re^4: Gratuitous use of Perl Prototypes

[dragonchild]

*snickers* It wasn't the name I was complaining about - it was your complete lack of variable names. :-)

And, now that I'm thinking about it, your version isn't prone to memory leaks because you don't have any lexical variables (which may be a good reason for writing it the way you did). It was the use of lexicals that, potentially, could create memory leaks in Error.

[BrowserUk]

Ah! I see what you mean. (I should have realised that, given you were advocating mapcar() (What the...?) :).

The main reason for my avoiding lexicals in a utility sub of this type is that I want to retain the aliasing nature of @_. If you look closely at the output from the second and third print statements, you'll see that they have side-effects. (Which will probably be considered as undesirable, but again, it is sometimes useful.)

Perhaps these are more palatable?

#! perl -slw
use strict;
use List::Util qw[ min ];

local $, = ' | ';

use constant {
    CODEBLOCK    => 0,
    ARRAY_ONE    => 1,
    ARRAY_TWO    => 2,
};

sub zip (&\@\@) {
    map {
        $_[CODEBLOCK]->( 
            $_[ARRAY_ONE][ $_ ], 
            $_[ARRAY_TWO][ $_ ] 
        ) 
    } 0 .. ( 
            @{ $_[ARRAY_ONE]} < @{ $_[ARRAY_TWO] } 
            ? $#{ $_[ARRAY_ONE] } 
            : $#{ $_[ARRAY_TWO] }
    );
}

my @x = 'a' .. 'z';
my @n = 1 .. 10;

print zip{ "@_" } @x, @n;

print zip{ $_[ 0 ] .= $_[ 1 ] } @x, @n;
print zip{ $_[ 0 ] .= $_[ 1 ] } @x, @n;

sub zipn (&\@\@;\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@\@) {
    map {
        my $i = $_;
        $_[CODEBLOCK]->( map{ $_[ $_ ][ $i ] } 1 .. $#_ ) 
    } 0 .. min map{ $#{ $_ } } @_[ 1 .. $#_ ];
}

my @a = split'','ARC';
my @b = 1 .. 3;
my @c = split'','ADP';
my @d = ( 1, 2, 0 );

print "\n---\n";

print for zipn{ join '', @_ } @a, @b, @c, @d;

__END__
[17:21:15.53] P:\test>406231
a 1 | b 2 | c 3 | d 4 | e 5 | f 6 | g 7 | h 8 | i 9 | j 10
a1 | b2 | c3 | d4 | e5 | f6 | g7 | h8 | i9 | j10
a11 | b22 | c33 | d44 | e55 | f66 | g77 | h88 | i99 | j1010

---

A1A1
R2D2
C3P0

Now, if anyone can tell me how to make zipn truely generic in the number of arrays it can deal with I'd be very interested. 20 is probably more than enough for any purpose I am likely to encounter, but I hate arbitrary limits.

Also, if anyone can see how to retain the alias status of the parameters to zipn, that would be really cool.