Stronger Testing Procedures

This is not a specific issue, but more of a general feedback. I am a big fan of this repository and have been since the very early days. However, it is my impression that changes to the code base, especially pull requests submitted by the community, are accepted without applying strong testing procedures. I'd like to make the case for this by explaining my situation:

- Version 3.2.1 was a stable version for me (I am only using line and scatter charts in a relatively simple manner)
- With version 3.2.2 the chart renderers were refactored ([3164](https://github.com/danielgindi/Charts/pull/3164)). However, the changes caused the first circle on line charts to disappear ([3915](https://github.com/danielgindi/Charts/issues/3915)).
- Version 3.3.0 meant to fix the issue with the circles, by using a new iterator ([3891](https://github.com/danielgindi/Charts/pull/3891)). Yet, this caused a much greater issue, leading to many reports of app crashes (compare [4049](https://github.com/danielgindi/Charts/issues/4049)). These app crashes are also rather frequent, with more than 5% of users of my app experiencing it.

As a result, version 3.2.1 is the last stable release which is suitable for use in my app.

Review PRs 3164 and 3891 alongside issues 3915 and 4049; the issue names no files or tests to run. Define the requested testing procedure and its acceptance criteria before implementation, since the issue provides no concrete change or measurable done condition.

Written by the indexing model from the issue text.