Over the weekend I tossed together an Archiver class and a root Object
class to provide default behavior. It is really quick and dirty and
the output files are rather wordy, and it doesn't handle a number
of Perl types, but it does allow objects with scalars, pointers to
object, arrays and hashes to be written to disk by one program
and read in by another. Of course if the other program doesn't
impliment the code your object won't bless, but hey... it's only
a weekend's work :-)
I don't know how much further this could be pushed in Perl code though.
You can't handle general archiving without a means of node marking,
which could be done at the high level with some probably slow hacks.
Otherwise you can get infinite loops. Well, not infinite. You'll run
out of space on the swap disk first :-)
Anyone working on this area? I'd be happy to chat about it from the
perspective of an Objective C / NeXT programmer.
class to provide default behavior. It is really quick and dirty and
the output files are rather wordy, and it doesn't handle a number
of Perl types, but it does allow objects with scalars, pointers to
object, arrays and hashes to be written to disk by one program
and read in by another. Of course if the other program doesn't
impliment the code your object won't bless, but hey... it's only
a weekend's work :-)
I don't know how much further this could be pushed in Perl code though.
You can't handle general archiving without a means of node marking,
which could be done at the high level with some probably slow hacks.
Otherwise you can get infinite loops. Well, not infinite. You'll run
out of space on the swap disk first :-)
Anyone working on this area? I'd be happy to chat about it from the
perspective of an Objective C / NeXT programmer.