MNT: update github actions with logical sequence of tests - #500

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (f1367e3) 71.30% compared to head (38e8e46) 71.11%.
Report is 1 commits behind head on develop.

❗ Current head 38e8e46 differs from pull request most recent head e1c6d6f. Consider uploading reports for the commit e1c6d6f to get more accurate results

@@             Coverage Diff             @@
##           develop     #500      +/-   ##
===========================================
- Coverage    71.30%   71.11%   -0.19%     
===========================================
  Files           56       55       -1     
  Lines         9374     9279      -95     
===========================================
- Hits          6684     6599      -85     
+ Misses        2690     2680      -10     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Kinda doesn't work as expected. Gonna close it for a while so nobody reviews it, but we can take a deeper look in the short-future.

Please don't delete branch

After almost 4 hours spent on figuring out how the GitHub actions works, I believe I finally finished this one!!
It's open for a review @RocketPy-Team/code-owners !!

• We need to urgently reduce the time spent on "integration" tests, they are consuming too much of our time. This not only drains the limited quota that we have on github actions but also let us waiting for about 15-20 minutes each time we want to validate the checks.
• I tried to run concurrent jobs to open more runners at once on github in order to reduce the waiting time, but this was tricky since the amount of time speant on each runner will count for the final usage of github actions.
• I also think that it should not be necessary to run pytest in all the 3 operational system every time. Maybe in the future we could trigger the 3 of them only wen targeting to the master branch. But I left as it is for now.
• After approving, the 32 commits can be squashed into a single one, without any issues.

Final comment: This was the PR #500 :)

Also, I wonder why the tests on python 3.12 are much slower than the 3.8, it should be the opposite.