Front page | perl.perl5.porters |
Postings from May 2009
On Fri, 8 May 2009 20:56:47 -0700
chromatic <chromatic@wgz.org> wrote:
> On Friday 08 May 2009 20:00:25 David Nicol wrote:
>
> > >> You're new here. "Nobody relies on it" is code for "somebody will now
> > >> show me code which relies on it". :)
>
> > I had hoped this thread would die off and I wouldn't have to write
> > this but here's a reasonable scenario for code relying on a for loop
> > returning undef. It is hypothetical, of course, but this is the kind
> > of entirely innocent thing that would break.
> >
> >
> > sub brake_for_safety { ... }
>
> There's a point at which a reasonable person should be able to say that anyone
> blindly upgrading the dependencies of a braking control system without running
> any tests whatsoever is so incompetent that he should give up programming
> altogether, forever, and should possibly spend the rest of his life stacking
> rocks.
I originally made my point as a half-serious, half "reductio ad absurdum"
demonstration on why we should think about it more carefully...
I don't think it would be safe to wholesale change the behaviour of half
of all the perl code ever written; that would be ... unwise.
Instead, if we want such an ability, why not protect it by a 'feature'
bit? We now have this lovely "use feature" introduced at 5.10 for just
such a purpose. If people want to enable it, then let them:
use feature qw( control_as_expression ); # someone find a better name
# ?:
my $name = if( $person eq "Fred" ) { "Bob" } else { "Jane" };
# map
my @squares = foreach my $number ( @nums ) { $x * $x }
# chain of ?:, like the OP requested
my $colour = given( $colname ) {
0 when "black";
1 when "red";
2 when "green";
3 when "yellow";
...
};
--
Paul "LeoNerd" Evans
leonerd@leonerd.org.uk
ICQ# 4135350 | Registered Linux# 179460
http://www.leonerd.org.uk/
Thread Previous
|
Thread Next