On 23/04/2020 10:44, Anna Kukliansky wrote:
You are right, this is exactly our goal. It's just that the new
NioAcceptor we implemented is a pretty simple extension over the
existing NioSocketAcceptor (which I'm changing here to support
inheritance), and it's a shame to duplicate all of the code.
A bit more context: we have long-running tcp connections, which we
don't want to disconnect. When we deploy new code to our servers we
start it with a new class loader, and pass the sockets from the old
code to the new one. What we did is fork Mina (with the changes I'm
suggesting here), so we can extend NioSocketAcceptor to support this
functionality. This is working well for us, but we'd like to get rid
of this fork.
Ah, now it makes more sense to me.
Ok, I think it's a no brainer. I may take a bit of time this week-end to
check some other MINA patches, in order to cut a release in the coming
days. That may take a week or two, I'm currently pretty much under the
water with many other tasks.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@mina.apache.org
For additional commands, e-mail: dev-h...@mina.apache.org
