Get asv benchmark suite running again.

[danielballan]

Tiled has a nascent (nascent!) suite of benchmarks using asv. They have languished; the last update was in October 2021.

This PR gets them back into working order. Here is an example, running on this PR branch:

$ asv run --environment existing:same
· Discovering benchmarks
· Running 3 total benchmarks (1 commits * 1 environments * 3 benchmarks)
[ 0.00%] ·· Benchmarking existing-py_home_dallan_micromamba_envs_py311_bin_python3.11
[16.67%] ··· Running (benchmarks.TimeSuite.time_list_tree--)...
[66.67%] ··· benchmarks.TimeSuite.time_list_tree                                                                                                                                                 8.92±0.5ms
[83.33%] ··· benchmarks.TimeSuite.time_lookup                                                                                                                                                    4.63±0.2ms
[100.00%] ··· benchmarks.TimeSuite.time_lookup_and_read                                                                                                                                           12.9±0.4ms

Note: asv is designed to create dedicated conda environments to run its benchmarks in. This facilitates testing against different versions of Python, numpy, etc. if you configure it so. It is designed to run across ranged of commits, tracking a project's history, and publishing the results as nice plots in a static web site.

For quick tests though, I like the above incantation, which just runs the current commit in the current Python environment.

[danielballan]

Notice that instead of starting a server in a subprocess, this PR runs the server and client in the same process. Messages are passed directly via Python function calls, the same as it works in unit tests. This takes networking out of the equation and will give more stable results, focused on actual Tiled performance.

[padraic-shafer]

👍