Improve and Validate TestingSequence assertions

[viceroypenguin]

This PR updates the assertions made in TestingSequence:

• Per the specification, it is not a failure to call .Dispose() multiple times. The .Dispose() method is expected to be idempotent, such that it is callable multiple times without throwing an exception. Updated to allow relying on this specification via Options.AllowRepeatedDispoals.
• Per the specification, it is not a failure to call .MoveNext() after receiving a false response. The enumerator is simply expected to continue to return false for each following call. As such, we should not be afraid to take advantage of such behavior when it makes code easier (see Simplify ZipLongest implementation #905 for examples). Updated to allow relying on this specification via Options.AllowRepeatedMoveNexts.
• While most IEnumerators do not complain when calling .MoveNext() after disposal, it does indicate an error in our code to expect that .MoveNext() is a valid behavior after we have disposed the iterator. As such, we should fail directly.
• While most IEnumerators return a default or the last value when calling .Current after .MoveNext() returns false, the spec does not make any promises on the usefulness of .Current in this situation. More importantly, we should be relying on .MoveNext() return value and not attempting to reference .Current in these cases. As such, we should fail directly.
• While most IEnumerators do not complain when calling .Current after disposal, it does indicate an error in our code to expect that .Current is a valid behavior after we have disposed the iterator. As such, we should fail directly.
• Changing AllowMultipleEnumerations to an exact expectation of how many times the sequence should be enumerated.
• Disallowing multiple simultaneous enumerations.
• Adding unit tests to prove TestingSequence validates what it claims it validates; giving us better confidence that invalid MoreLinq and MoreLinq.Test code is written correctly.

[codecov]

Codecov Report

Merging #936 (1291504) into master (5b49471) will increase coverage by 0.04%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #936      +/-   ##
==========================================
+ Coverage   92.39%   92.44%   +0.04%     
==========================================
  Files         112      113       +1     
  Lines        3434     3453      +19     
  Branches     1021     1025       +4     
==========================================
+ Hits         3173     3192      +19     
+ Misses        199      196       -3     
- Partials       62       65       +3     

Impacted Files	Coverage Δ
MoreLinq/ListLike.cs	86.95% <0.00%> (-5.91%)	⬇️
MoreLinq/AssertCount.cs	96.29% <0.00%> (-0.14%)	⬇️
MoreLinq/SkipLast.cs	100.00% <0.00%> (ø)
MoreLinq/TakeLast.cs	100.00% <0.00%> (ø)
MoreLinq/CollectionLike.cs	80.00% <0.00%> (ø)
MoreLinq/PadStart.cs	97.91% <0.00%> (+0.13%)	⬆️
MoreLinq/CountDown.cs	90.32% <0.00%> (+0.32%)	⬆️
MoreLinq/FallbackIfEmpty.cs	96.22% <0.00%> (+7.54%)	⬆️

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[atifaziz]

Depends on #901.

Sorry, but I'm confused. Doesn't this superseded #901 instead of depending on it?

[viceroypenguin]

Up to you - feels like separate PRs to me, given the nature of the work, but if you feel they're the same, then this one could be pulled directly without #901. PR description here updated as if single PR.

[atifaziz]

This is really a great set of improvements and having gone the extra distance to test the TestingSequence<T> implementation is just absolute bonus! Thanks!

I have made several suggestions to simplify things. Have a look and then I think we're good to merge soon.

[viceroypenguin]

I'll address the code changes when I get a chance.

[atifaziz]

Thanks for all the improvements, follow-up and tests added with this PR! Good to merge! 🚀

[viceroypenguin]

Woot! 🎉