Ok, got it. Normally I work with Eclipse, but for this purpose I
installed a version of Netbeans too.
The first thing I noticed was, that while using ByteBuffer.getString(..)
a char[] array increases slightly over the time (after garbage
processing it decreases, but overall I noticed permanent increment).
I removed the TextLineDecoder and profiled again. The next thing I found
out was, that the DirectByteBuffer.DeAllocator / Cleaner (number of
living objects) increases over a longer period in a form like an
increasing sawtooth. Seems, that not all objects won't be garbaged? Has
Java itself a problem at this point?
Sorry, but more I can't say at the moment :(
The problem is, that I used in this test none special code from me. So
what could be the problem?
I want to use Mina as part of a bigger project, but at the moment I
don't know, how to decide.
Regards
Michael
[EMAIL PROTECTED] wrote:
I recommend trying the Sun JFluid profiler (which is now part of
Netbeans - profiler.netbeans.org) and is free. It is excellent for both
code profiling for performance and for tracking down memory leaks.
Robert
