[zeromq-dev] Java API is not notifed of C++ assert failures.

Martin Sustrik sustrik at fastmq.com
Fri Mar 13 13:55:14 CET 2009


> Perhaps this is a stupid idea, but is there benefit in implementing a
> HA solution in 0MQ as a standard component?  The work we did for
> OpenAMQ is worth reusing IMO, and this should be a generic solution
> for 0MQ applications.

Sure, it would be worth of implementing. However it's a matter of 
priorities: We should implement high-level constructs (pre-built shared 
queues, pre-built HA pairs etc.) after implementing missing low-level 
functionality (message routing, management). These features in turn are 
scheduled after the work done for clients (on-disk offload, multicast 
issues, AMQP issues)... so it'll take us some time to get there. Of 
course everybody has an option of becoming a customer and thus have the 
work on features he needs prioritised.


More information about the zeromq-dev mailing list