chore(docker): improve image generation and testing

[jansegre]

A trial run is running here: https://github.com/jansegre/hathor-core/runs/5760065789?check_suite_focus=true

This PR has three separate improvements:

• add a subcommand quick_test that will exit with success (exit_code==0) when a transaction is added to the database, or exit with failure (exit_code!=0) in any other case (including timing out if it take too long)
• add a step on the CI to test the built image before pushing the image to the repositories, this makes sure that an image which for any reason does not pass the quick_test check will not be uploaded
• use debian-slim instead of alpine as the base python image, this is mostly because in alpine rocksdb is not "stable", so we're subject to a lot more of instability, for example, the latest build of rocksdb depends on a version of stdlibc++ that is not on the main repository of the release we use (alpine 3.15) but instead has to be installed from the edge version, this time the update is safe, but it may be that an update in the future causes dependency installation problems much harder to solve, just switching to debian-slim seems much easier maintain, besides that's what the pypy image is based off anyway, so that would also be a little easier to maintain.

[codecov]

Codecov Report

Merging #390 (5eff11d) into dev (2820060) will decrease coverage by 0.01%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##              dev     #390      +/-   ##
==========================================
- Coverage   82.85%   82.84%   -0.02%     
==========================================
  Files         171      171              
  Lines       15542    15542              
  Branches     2457     2457              
==========================================
- Hits        12878    12876       -2     
- Misses       2220     2221       +1     
- Partials      444      445       +1     

Impacted Files	Coverage Δ
hathor/p2p/node_sync.py	84.02% <0.00%> (-0.50%)	⬇️

---

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 2820060...5eff11d. Read the comment docs.