On Mon, July 13, 2009 6:13 pm, Bill Moseley wrote: > (in cleanup) Item #1 returned by STORABLE_freeze for DateTime > is not a reference at ../../lib/Storable.pm (autosplit into > ../../lib/auto/Storable/_freeze.al) line 339 during global > destruction, at <....> line 327
> It's expected that this nfreeze is happening in a DESTROY() (the data > gets serialized and stored upon destroy). Is the destroy a red herring? Order of destruction during global destruction isn't guaranteed; it's entirely possible there are important bits of Storable gone by the time your DESTROY is called. Arrange your scopes such that any important DESTROY happens before global destruction.
