Dan Sugalski <[EMAIL PROTECTED]> wrote: >>Not currently. There used to be a C<destroy> opcode, but I've deleted >>it, because I thought it's too dangerous.
> We really need to put it back in -- I knew it was dangerous, but it > was necessary. Yeah. > ... We should probably make it 'safe' by forcing the > destroyed PMC to be an Undef after destruction, in case something was > still referring to it. That sounds sane. Or maybe be: convert to an Undef and put a Null PMC vtable into it which would catch any further access to that PMC. BTW shouldn't we really separate C<destroy> and C<finalize>? The latter would be overridable by user code, the former frees allocate memory. leo
