Add not() matcher operator

[bendykowski]

Co-authored-by: cakeinpanic [email protected]

[codecov-io]

Codecov Report

Merging #84 into master will decrease coverage by 0.12%.
The diff coverage is 95.18%.

@@            Coverage Diff             @@
##           master      #84      +/-   ##
==========================================
- Coverage   94.58%   94.45%   -0.13%     
==========================================
  Files          35       36       +1     
  Lines         609      667      +58     
  Branches       69       72       +3     
==========================================
+ Hits          576      630      +54     
- Misses         24       25       +1     
- Partials        9       12       +3

Impacted Files	Coverage Δ
src/matcher/type/AnyOfClassMatcher.ts	100% <100%> (ø)	⬆️
src/ts-mockito.ts	96.82% <100%> (+0.1%)	⬆️
src/matcher/type/ObjectContainingMatcher.ts	90% <100%> (ø)	⬆️
src/matcher/ArgsToMatchersValidator.ts	100% <100%> (ø)	⬆️
src/matcher/type/DeepEqualMatcher.ts	80% <100%> (ø)	⬆️
src/matcher/type/AnyStringMatcher.ts	88.88% <100%> (ø)	⬆️
src/matcher/type/Matcher.ts	66.66% <100%> (ø)	⬆️
src/matcher/type/AnythingMatcher.ts	87.5% <100%> (ø)	⬆️
src/matcher/type/AnyFunctionMatcher.ts	100% <100%> (ø)	⬆️
src/matcher/type/BetweenMatcher.ts	90.9% <100%> (ø)	⬆️
... and 8 more

---

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 ba5997f...66da980. Read the comment docs.

[bendykowski]

Hi @cakeinpanic, I've created another PR with my changes. I thought it will be the fastest way. You are added as a co-author of those changes of course. I hope @NagRock will be able to check that soon.

[cakeinpanic]

Are you sure package.lock is needed to be in repo?

[cakeinpanic]

What is the difference between my PR and your except of changes in Mock.ts?

[bendykowski]

Hi @cakeinpanic, package-lock.json indeed should be a part of a repo (see: https://docs.npmjs.com/files/package-lock.json). The difference is that I've changed the not mechanism. I've renamed it to NotOperator which better describes what it is. Also removed all unnecessary code from all matchers. Instead of changing matchers themselves, adding everywhere handling functionality I've moved it to NotOperator. For now NotOperator, when ex. anyFunction() is called on it, creates AnyFunctionMatcher, saves it and returns itself (not AnyFunctionMatcher). When the stubbing mechanism is checking if a param is matching a given matcher, calls isMatching() method on NotOperator which calls isMatching() on a saved matcher and reverts the output. The same when the pretty name is needed, NotOperator calls toString() on a saved matcher and prepend the output with not().. Doing that we are separating the logic of matchers and operators, don't have to change matchers and remember to add this logic when creating a new matcher. To achieve that, still extending the Matcher class, as I need the same interface, I had to rename core match() method to isMatching() to avoid a conflict with MatchMatcher and better describe the purpose of this method. I've kept the rest of your changes intact.

[cakeinpanic]

@bendykowski okey, now I see the difference, thank you. Your solution looks better

The only one thing you forgot is toString, method for not() . matcher, I think it might be useful sometimes

[cakeinpanic]

@NagRock friendly reminder for review/merge)

[cakeinpanic]

@NagRock friendly reminder for review/merge)