Less than 48 hours left to get any PPI issues sorted out.
It's related to the super-cool tree mechanism PPI uses. It looks like a normal tree, doesn't need circulars, never ever leaks, and looks just like what you would expect.
It's just that when you clone or thaw you have to batch-relink the child nodes to the parents in the new tree.
So I've added STORABLE_freeze/thaw hooks to PPI and 17_storable.t to the test dir and all is well again. So we're back to "zarro boogs".
Went completely over the main POD and rewrote about half of it for 1.000, removing all the temp and status stuff, rewriting stupid bits and generally reorganising it into a more logical order. And about a dozen people who contributed during the development should receive a happy release present.
Lastly but most significantly I did the actual conversion of the $VERSIONs to 1.000. So it's now sitting pristinely there in CVS ready to roll with a single shell command to the magic package assembler/roller/uploader script all that remains.
Unless you object and have a bug report. You have 47 hours and 52 minutes, and counting...