At 01:08 PM 7/11/2002 -0700, Ashley Winters wrote:
>On Thursday 11 July 2002 11:47 am, Chip Salzenberg wrote:
> > According to Dan Sugalski:
> > > At 9:50 PM -0400 7/9/02, Chip Salzenberg wrote:
> > > > 3a. If so, how can one distinguish among the e.g. many C<my $foo>
> > > > variables declared within the current function?
> > >
> > > One pad per block, rather than per sub.
> >
> > I just remembered why I thought that woundn't work: BEGIN is a block.
> >
> > my $x = 1;
> > BEGIN { %MY::{'$y'} = \$x }
> > print $y;
>
>Even worse, you should be able to modify lexicals even outside your scope.
>
>sub violate_me {
> caller(1).MY{'$y'} := caller(1).MY{'$x'}; # hypothetical syntax
>}
>
>{
> my $x = 1;
> my $y; # Might be able to BEGIN { violate_me() } instead
> violate_me();
> print $y;
>}
This reminds me why I don't use Perl4 'local' anymore.
And now we have even uglier ways to write poor code. :)
The only real use I can see of %MY is debugging. If people are going
to take handles to pads and modify lexicals in closures, continuations
and routines from the outside, it probably means that the item needs to
be a class.
And side effects like "I call you, you modify me invisibly...." seems
more like taking dangerous drugs than programming.
Yep, I warned you about calling that routine, now look what it did to
your brains.
-Melvin
