Switch to lorenzwalthert/touchstone

[lorenzwalthert]

This PR replaces the continuous benchmarking infrastructure provided by r-lib/bench with the touchstone package. Key insights were:

• Does not make sense to only benchmark the feature branch, but benchmark the base branch as well and look at relative difference because two CI/CD runs are not exactly comparable in terms of available compute power. This matters in particular when running one iteration of a benchmark takes long (>> a few seconds) and speed implications are small.
• Maintaining a timeline such as the implementation of r-lib/bench is of limited use because of the first bullet. Speed implications are to be checked between two revisions.
• R and package versions must be fixed via RSPM to allow as much continuation as possible anyways. Changing the timestamp of RSPM can happen in PRs that are only dedicated to dependency updates.

From https://github.com/lorenzwalthert/touchstone.

First results with 20 iterations are promising and take about 30 minutes. They show 0% difference in median for this PR, which was expected as no code that is benchmarked changes. Second run: 0.9%, third run 3%, fourth -0.1%. Results can be found under the GitHub Action -> Artifacts.

Here is the comparison of 20 iterations with 20 times style_pkg() for the here package for master and touchstone branch.

[codecov-io]

Codecov Report

Merging #698 (933e176) into master (a2a2e2e) will decrease coverage by 0.00%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #698      +/-   ##
==========================================
- Coverage   90.42%   90.41%   -0.01%     
==========================================
  Files          47       47              
  Lines        2225     2223       -2     
==========================================
- Hits         2012     2010       -2     
  Misses        213      213              

Impacted Files	Coverage Δ
R/set-assert-args.R	92.30% <0.00%> (-0.38%)	⬇️
R/rules-line-break.R	100.00% <0.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6d6131b...933e176. Read the comment docs.