With the amount of memory consumed growing by approx 4K per request, I'm beginning to suspect that this may be caused by some reference cycle that is preventing garbage collection, perhaps in the IO layer.
To detect something like this will require running the example code under some form of memory profiler. Which I don't have time for now, unfortunately.
_______________________________________
Jython tracker <report@bugs.jython.org>
<http://bugs.jython.org/issue1660>
_______________________________________