On Tue May 27 12:12:35 2008, MKUTTER wrote:
Show quoted text> This is a slanderous misinterpretation of Storables abilities, which
> clearly transcede the imagination of this miserable package's author.
Who happens to be one of the current maintainers of Storable, and one of
the people who made the portability fixes that got it into core. (Remote
debugging via e-mail on 64 bit AIX systems, IIRC, being one of the
highlights)
Show quoted text> The bottom-of-the-line thingy required to allow Storable serialization
> is implementing the methods STORABLE_freeze and STORABLE_thaw in
> packages using Acme::Class::Std.
>
> It is really offensive to see Storable so blatantly defamed.
No, you're missing the point. I made no such specific claim.
My specific claim was that the serialised think that they can serialise
something because they can see inside it.
You're right that if you define those two you *can* serialise it safely.
Because at that point the serialiser is not poking around in the guts.
Yes, I could be clearer on that one.
The point of the module was to note that the interaction between
inside-out objects and serialisation systems could result in potential
silent data loss, a danger that (at the time) was true of all
serialisers and inside-out object modules I was aware of.
Also, I should note that we fixed a core bug as a result of me writing
this module - you can't now dereference FORMAT references with $, which
you used to be able to.
