The large amount of new code in 0.6 means that we really need much
more sophisticated benchmarks. The simple pong benchmark was
sufficient to get a rough approximation of the performance of our web
server, but now that we have the higher-level functionality of
snaplets, a useful benchmark needs a larger site that will exercise
more of those code paths. Right now we've got a lot on our plate, and
this isn't exactly at the top of the list. We do want to do more
benchmarks at some point, but I'm not sure when that will actually
happen.
If you want to run some, we'd love to hear about the results. Bryan
O'Sullivan has been working on a new benchmarking tool called pronk
(https://github.com/mailrank/pronk) that might be helpful.
Doug Beardsley
2011/10/28 Krzysztof Skrzętnicki <gte...@gmail.com>: