And other reasons to avoid overusing reflection
On this fine day I have published the project that I use to compare relative performance of different approaches to algorithms. It aims to get a few things right with this:
- Tested scripts are re-shuffled between runs to lower execution order/CPU cache bias.
- Script call time can be excluded if you are comparing approaches rather than actual scripts.
- Easy to use - you add scripts that you want tested to the project, specify how you want them to be called, and that's it.
The project can be found on GitHub.