John Siracusa wrote: > Okay, this part has me confused. And rightly so: it was a screw-up. I lost track of whether I was keeping the property on the value or on the node reference and ended up doing both. Interestingly, the code would still have *worked* since the (originally unset) property on the node reference would have returned C<undef> which would undergo the usual boolean conversion in the C<if>, and the usual promotion to zero in the numerical context of the increment. Anyway, a patch is winging it's way to Simon even as I type. Thanks for pointing out the snafu, John. DamianThread Previous | Thread Next