Front page | perl.perl5.porters |
Postings from August 2009
On Tue, Aug 11, 2009 at 2:31 PM, Eirik Berg Hanssen<Eirik-
>> As an r-value, the list assignment operator passes its own
>> r-value through.
>
> What is "its own r-value"? If you mean to say the "elements
> produced by the expression on the right hand side of the assignment",
> it is incorrect.
>
> (And if you meant something else, I found it less clear than what
> the patch had.)
Yes, that's what I meant. Thanks for straightening out my misconception.
>
> use 5.010;
> use strict;
> use warnings;
> my @x = my ($x, undef, $y) = 1..10;
> say $_||'undef' for @x;
> __END__
> 1
> 2
> 3
>
> 3 elements, not 10. Specifically, the three elements assigned to.
wow.
/tmp: $ perl -le '@a=qw/24 57 32 hike!/; $count = () = @a; print $count'
4
/tmp: $ perl -le '@a=qw/24 57 32 hike!/; @copy = () = @a; print @copy'
/tmp: $
So list assignment in scalar context returns the count of the RHS, but
list assignment in list context returns the LHS.
>
> Aside:
>
> And none of them undef either. In fact, this is one ... rather
> _special_ ... corner case:
>
> use 5.010;
> use strict;
> use warnings;
> for my $u (undef) {
> my @x = ($u) = 1..10;
> say $_||'undef' for @x, $u;
> }
> __END__
> 1
> undef
>
> The assignment to ($u) evaluates to a list of one element, the value
> of which differs from the value of $u, which remains the same both
> before and after the assignment. Ah, the magic of undef. :)
Is that not a bug, that an assignment to a constant ($u is an alias to
undef) did not cause an exception? It looks like undef is special
that way.
/tmp: $ perl -le ' $count = undef = 3; print $count'
Modification of a read-only value attempted at -e line 1.
/tmp: $ perl -le ' ($count) = (undef) = (3); print $count'
3
/tmp: $ perl -le ' ($count) = (22) = (3); print $count'
Can't modify constant item in list assignment at -e line 1, near ");"
Execution of -e aborted due to compilation errors.
/tmp: $
Thread Previous
|
Thread Next