Hmm, well PAR_CLEAN=1 only partially helps. It moves the problem from DBI to Storable, but it's a little different. I don't have any problems using Storable as I did with DBI, however when run the application that's using the PAR I now have the endless-loop stat64()'ing Storable.pm.
Really I'm not convinced the problem is with Storable so much as that's where it finally gives up the goat. Unfortunately I haven't yet been able to produce an isolated test-case. One of the things that seems to trigger the problem are circular dependencies. The code I'm dealing with has a few places where A depends on B which in-turn depends on A. Breaking the circle fixes the problem but I believe there must be something more to it as making test packages A -> B -> A doesn't demonstrate the problem. Regardless I'm past the point where I can spend any more time looking at PAR, which is a shame. Hopefully I'll have time in the future and someone will have fixed the DBI problems fully. Matt
