On 30/09/2010 10:49, Damian Conway wrote:
As long as C<.perl> works the way it does, there can be no real
privacy.
Sigh. That is indeed badly broken. Surely it ought to default to C<die>,
and require class architects to override .perl explicitly if they wish to
break encapsulation.
I see/use .perl as a debugging aid. When debugging, I tend to want to
see what is *really* there, including its private bits, since those may
well hold the key to the bug. I'd say .perl is currently optimized for
debugging, but of course anything that helps from that angle is open to
being (ab)used for encapsulation breaking too. And maybe that's not what
.perl should be optimized for.
And thus C<.get_value> and C<.set_value> are just convenient
access points for the same behaviour.
Yes. People are going to shoot themselves in the foot anyway,
so let's legalize semi-automatic weapons as well.
If we don't provide some way then I suspect folks will either start
parsing the result from .perl (or whatever the debuggy-dumper is
called), and/or pretty quickly a CPAN module (or several) will appear
that do some low-level, per-implementation trickery to break the
encapsulation.
I'd agree we don't want to make encapsulation breaking so convenient
that folks just mindlessly, effortlessly do it. I'm all for finding ways
to declare up front that evil is happening. But I think trying to make
it outright impossible may backfire and lead to there being multiple
different ways to behave badly rather than one standard, easy to spot one.
Plus, we probably do need *some* way for folks to write serializers in
standard Perl 6.
I'm still undecided on whether or not I think C<use MONKEY_TYPING;> is
the right way to enable this kind of privacy breakage. Maybe nothing
is needed, since the C<^> in C<$obj.^attributes> (or the C<HOW>)
already says "warning! meta!".
Except that, if the Attribute object is passed around, it may be saying
"warning! meta!" in a completely different scope, in a completely
different file.
Sounds like the encapsulation breaking thingy probably wants to be
looking for some pragma to have been used in the lexical scope of the
caller, maybe. I'd rather that we called it something other than
MONKEY_TYPING though. Different evil, different pragma. :-)
Thanks,
/jnthn