Front page | perl.perl5.porters |
Postings from September 2003
On Wed, 2003-09-10 at 22:11, Enache Adrian wrote:
> > What else?
>
> This :-)
>
> $ perl -le 'sub a { print "a" } sub b { print "b" } (a)[b]'
> b
> a
>
> Regards,
> Adi
so we're up to, in general left-to-right, except the RHS is always
before of the LVALUE, except for self-assignment, when the LVALUE
is part of the RHS and does not get reevaluated.
Also the index of an array is evaluated before a dynamic
array expression, as Adi points out. Curiously this does not
occur with hashes, only arrays.
perl -le 'sub a { print "a" } sub b { print "b" } @{[4 => a]}[b()]'
b
a
perl -le 'sub a { print "a" } sub b { print "b" } ${{4 => a}}{b()}'
a
b
More specificly, only array slices
perl -le 'sub a { print "a" } sub b { print "b" } ${[4 => a]}[b()]'
a
b
Hash slices evaluate the slice target first too
perl -le 'sub a { print "a" } sub b { print "b" } @{{4 => a}}{b()}'
b
a
so
(a)[b] # a slice, b first
[a]->[b] # not a slice, a first
Wow.
Is there any sense to this? The lookup of a sliced container is
not skipped when the key expression returns a null list:
perl -le 'sub a { print "a" } sub b { print "b";() } (a)[b()]'
b
a
was
that an optimization that was planned but never realized? It would
break some very odd code that considers fetching null slices out of
tied containers significant, it would also break designs that rely
on autovivification in such cases.
--
David Nicol / kernel 2.6.0 is pretty whippy
Thread Previous
|
Thread Next