Re: [rules-dev] workingmemory and dispose()
I am OK with that as long as examples and doco show the newStatelessSession
generously.
I never liked the weakhashmap - it kind of got rid of the leak on paper, but
put a load on the GC at the worst possible time.
Stateless and Statefull are great ideas I think. WorkingMemory is then
either just an interface, or mainly a "concept" that people talk about.
Everything else out there likes to use the word session (eg hibernate) so
the semantics of it are pretty well understood.
On 4/23/07, Mark Proctor <mproctor(a)codehaus.org> wrote:
currently working memories are auto garbage collected once you null
them, by using a weak hashmap in the RuleBase. I'm currently adding
async methods to the working memory, which means that the weak hashmap
no longer works. So I'm thinking of using a normal hashmap, which means
people MUST call workingMemory.dispose() or they will get memory leaks,
are people ok with that? I now have ruleBase.newStatelessSession which
won't adding the session to the Map, to avoid having to call dispose()
for stateless sessions.
Mark
_______________________________________________
rules-dev mailing list
rules-dev(a)lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-dev
Attachments:
- attachment.html (text/html — 1.6 KB)