set up benchmarking suite.

[KRRT7]

Hi Folks, since beeware is familiar with pytest, I'd recommend to start setting up benchmarks and use https://pytest-benchmark.readthedocs.io/en/latest/ in order to do it.

[freakboy3742]

So... yes... and then what?

To date, I haven't been especially worried about performance, because a working "something" is better than an optimised high-performance "nothing". But, as Toga matures, I can see how benchmarking could be a useful tool to make sure that new code doesn't introduce unacceptable regressions in performance.

Picking a tool is essentially the last part of the process. The questions I have are:

1. What do we benchmark?
2. How do we establish what "acceptable" means for a benchmark?
3. How do we run those benchmarks in a way that feeds back into the development process?

If you want to propose adding a benchmark suite to Toga, those are the questions we need answered first, before we pick a tool - if only because the answers to those questions could well influence tool choice.