On 4/29/20 8:56 AM, Rémy Maucherat wrote:
On Tue, Apr 28, 2020 at 7:18 PM Mark Thomas <ma...@apache.org
<mailto:ma...@apache.org>> wrote:
On 28/04/2020 17:30, Rémy Maucherat wrote:
> Hi,
>
> I'm still looking at things to remove or refactor in 10
following the
> rearchitecting failure for the Connector. One candidate could be the
> Tribes transport, since NIO is the default and BIO is probably
never used.
>
> Can it be removed ?
I don't see why not.
> There are a few classes here and there that could go too, for
example
> that BufferPool15Impl class. Given the name, I would say it could be
> merged into the superclass.
+1
Done. I wonder if this global static cache is actually that useful
anymore.
ByteBuffers when first introduced, promised a lot, and delivered very
little. The performance implications of creating new ones were pretty
bad, imho.
There is a good chance that much of this has changed as the JVMs have
evolved, and that a lot of code could be removed for the benefit of
simplicity without sacrificing performance
Rémy
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org
<mailto:dev-unsubscr...@tomcat.apache.org>
For additional commands, e-mail: dev-h...@tomcat.apache.org
<mailto:dev-h...@tomcat.apache.org>
