Use Jest

[gillesdemey]

Inspired by #926, fixes #703 and partially addresses #559

As I understand the consensus seems to be:

• Use Jest
• Prefer assert
• Keep using supertest@3

I used Promises / async where it made most sense and where it simplified the tests.

I'm using --forceExit because some tests do not properly close app.listen, I don't think it's that big of a deal and the current test suite has the same behaviour.

Haven't focused on making the benchmarks work cross-platform, might send a PR in the future to use autocannon instead of wrk (although wrk is fine for me, it seems some folks are concerned about the Windows platform).

[codecov]

Codecov Report

Merging #981 into master will not change coverage.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #981    +/-   ##
======================================
  Coverage     100%   100%            
======================================
  Files           4      5     +1     
  Lines         431    356    -75     
  Branches      102      0   -102     
======================================
- Hits          431    356    -75

Impacted Files	Coverage Δ
lib/context.js	100% <0%> (ø)	⬆️
lib/application.js	100% <0%> (ø)	⬆️
lib/response.js	100% <0%> (ø)	⬆️
lib/request.js	100% <0%> (ø)	⬆️
test/helpers/context.js	100% <0%> (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 13c7ca6...91ed4ce. Read the comment docs.

[dead-horse]

great work!

[robinpokorny]

@gillesdemey, I was just wondering, what are the arguments for assert? In the issues you linked, I could not find it.

[gillesdemey]

@robinpokorny It's an alias of assert.ok

https://nodejs.org/api/assert.html#assert_assert_ok_value_message

[robinpokorny]

@gillesdemey Oh, I meant in comparison to expect that comes with Jest.

Full disclosure, I really like the new .resolves/.rejects matchers for async code. I was about to check if it is used here and create a PR otherwise. But found out that we use assert everywhere. It this PR, it says ‘Prefer assert’ and I could not find the reasons for it written anywhere.

[gillesdemey]

assert has a much simpler interface, contrasted to expect's more readable style.

I believe @jonathanong has stated his preference for assert a few time before but it ultimately comes down to personal preference :)

#926 (review)
#926 (review)

[robinpokorny]

@gillesdemey OK, thanks. That's what I wanted to find out!