It would be cool to give some information avout the used JVM, the system
you are running the test on, the number of procs, etc...
( We experienced some strange problem lately with JRockit JVM on linux )
Thanks !
Emmanuel
Frédéric Brégier a écrit :
I have made some tests with removing the backport
dependencies with no problem using native java5.
However I did not get deep in the tests so far, so I cannot
say that there is no problem.
For instance, I didn't run the program test you run.
I agree with you the native java5 should be better
(at least in theory). I need to advance a bit in my
port to Mina 1.0 to say if there are other problems
like the one you get.
Also I plan to look inside Mina code and see where we can
put "java5" functions, using both backport and native,
and propose this to Mina. But it will take some time...
and of course, as others, I have a job to do every day. ;-)
Frederic
Mark wrote:
On 10/2/06, Raman Gupta <[EMAIL PROTECTED]> wrote:
I am building Mina for java5 -- I have removed all references to the
concurrent backport package and have tried to rebuild. I ran "mvn
install". The build proceeds normally, but when:
I build MINA using Java5, but use the backport libs. Not sure if this
is correct, but it works.
Yeah, I know that works but I'm trying to avoid including a dependency
that I don't need. Also, the native java5 concurrent utilities are, in
many cases, faster than the backport versions (though I doubt the
difference is significant).
Cheers,
Raman
