"Tom Phoenix" schreef:
> &silly_1($fred, $barney, $betty, $wilma); # four parameters
>
> my $params = [$fred, $barney, $betty, $wilma];
> &silly_2($params); # one parameter
>
> It looks promising. Is the second subroutine call more efficient than
> the first one? Yes, it probably is. Does this style make the program
> run any faster? Probably not; it takes about as long to build $params
> as to call the sub with a list of parameters.
Internally the parameters of the first variant are *aliased* into @_.
$ perl -wle'
sub doit { $_[0] = "dirty" }
my $foo = my $bar = "clean";
print "$foo, $bar";
doit $foo, $bar;
print "$foo, $bar";
'
clean, clean
dirty, clean
In the second variant, the coder puts *copies* of the variables in an
array, to be able to pass an array reference.
$ perl -MData::Dumper -wle'
my $foo = my $bar = "clean";
my $aref = [$foo, $bar];
$foo = "dirty";
print Dumper $aref;
'
$VAR1 = [
'clean',
'clean'
];
Each of these copied variables could hold multi-megabyte strings. But
hey, it was a silly example, because only scalars are involved. The
profit is in passing references of non-scalars.
slightly_less_silly( [EMAIL PROTECTED], [EMAIL PROTECTED] );
ANFSCD: Tom, I keep wondering why do you often write the &-sigil where
it is not needed?
--
Affijn, Ruud
"Gewoon is een tijger."
--
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
http://learn.perl.org/
