#52 - add support for null in should clause

[alex-bogomaz]

hi, please review this PR. it contains fix for issue #52.
Now should be null should work the same as should be Null

[sergey-tihon]

Could we save prev behavior for backward compatibility as well?

null |> should be Null
null |> should be null  

[alex-bogomaz]

first 2 test of the test-class have exact same body, but different naming:

• null should be Null
• null should fail to not be Null

So they really test same thing.
There is the same situation in be Null tests for XUnit and MsTest.
If you take a loot at NUnit tests you'll see that first 2 tests are testing what is stated in their names:

[<Test>] member test.
     ``null should be Null`` ()=
        null |> should be Null

    [<Test>] member test.
     ``null should fail to not be Null`` ()=
        shouldFail (fun () -> null |> should not' (be Null))

So I guess there should be no backward-compatibility issues and now test body corresponds to the test name, just like similar test for NUnit

[alex-bogomaz]

PR is updated. Could you please review and provide feedback.

As for back-compatibility: both null |> should be Null and null |> should be null are valid assertions. So previous behaviour is still valid. There are unit-tests for this.

PR has 3 commits - if this doesn't work for you, I'll create new PR with single commit

[sergey-tihon]

@alex-bogomaz Could we close this PR?

[alex-bogomaz]

yes