renshaw

I'm with you, in favor of strongly typed headers. Working with strings can be very error-prone; you might not catch typo bugs until your system is in production and handling some rare case for which failure is expensive.

It seems to me that, in the worst case, the downside of your current scheme is an extra field dereference, as in `request.headers.extensions.get("crazyNewField")` instead of `request.headers.get("crazyNewField")`. I think that cost is easily outweighed by the benefits.

Update: the compilation error was actually due to some missing includes. https://github.com/kentonv/capnproto/pull/47

Switching compilers on the C++ side doesn't seem to make a significant difference in performance.

I should note that the Rust version currently omits some features of the C++ version, such as read-limiting and the actual counting of the throughput. These things are cheap, but they may explain why Rust is faster in that one case.

Author here. The makefiles I'm using are on github:

https://github.com/dwrensha/capnproto-rust/blob/master/Makef...

https://github.com/dwrensha/capnproto/blob/benchmark/c%2B%2B...

I compiled libcapnp with the latest Clang g++ that ships with XCode. It perhaps would be more fair to also compile the C++ benchmarks with Clang, instead of the Macports gcc4.8 that I'm using, but unfortunately Clang barfs on some template hackery in the benchmark driver.

I often use SML for writing games. The Functioning library (http://github.com/robsimmons/functioning) provides support for SDL and includes an SML port of Box2D.

Don't forget Pittsburgh.

Pittsburgh!

"The Soul of a New Machine" by Tracy Kidder.