Unfortunately I can't say alot about schema evolution and
synchronization in the old code. I don't believe there are any
provisions for locking between threads, let alone processes, for class
evolution. I think there may be some index caching going on a per-
process level.
How do you synchronize schema changes between processes? After
updating the schema you don't have problems seeing the new index,
just seeing new instances in an index that was updated by the other
process, correct?
I'll try to look into this tomorrow also.
Ian
On Jul 17, 2008, at 5:04 AM, Leslie P. Polzer wrote:
LPP> Sometimes[1] when image/process B adds persistent objects,
LPP> I have to call SYNCHRONIZE-CLASS-TO-STORE in process A
LPP> to get them picked up.
does it have something to do with process of defining/redefining
classes,
or your schema is rock-solid and defined before processes A and B
start?
I regularly make small modifications to those classes (add/remove
slots,
add/remove slot indices) and activate them asynchronously in A and B
(usually process A gets the modifications first).
what do you mean by "picked up" -- seen via indices, or something
else?
Seen via indices.
Leslie
_______________________________________________
elephant-devel site list
elephant-devel@common-lisp.net
http://common-lisp.net/mailman/listinfo/elephant-devel
_______________________________________________
elephant-devel site list
elephant-devel@common-lisp.net
http://common-lisp.net/mailman/listinfo/elephant-devel
