[elephant-devel] Re: Updated version of last Postmodern patchbundle

[Alex Mizrahi]

 IE> may be less of an issue in the past.  Deadlock is certainly likely to
 IE> happen if the set of operations to update an indexed slot is not
 IE> wrapped in a transaction.

can you explain why?
as i understand deadlocks can happen when there are two or less locks in 
both concurrent transactions and they hold each other in criss-cross way.
so isolated statements should not provoke deadlocks as far as i understand 
deadlock causes.

 IE> I'd love to see a reproducible deadlock test, even if it had to run for 
awhile.

Leslie's tests do not reproduce them for you?
(if you cannot run them, call (open-store ..) rather than with-open-store).