Improve BeTrue assertion

[t3mi]

1. General summary of the issue

I understand that examples specified below is how PowerShell works but it's not so obvious what assertion to use to make a proper check against the $true value even from the Should page. It would be helpful at least for BeTrue assertion to have a proper compare check.

Describe "Check if Truthy" {

    It "False as a string" {
        'False' | Should -Be $true
    }

    It "False as a string" {
        'False' | Should -BeTrue
    }

    It "Any string" {
        'Whatever' | Should -Be $true
    }

    It "Any string" {
        'Whatever' | Should -BeExactly $true
    }

    It "Any string" {
        'Whatever' | Should -BeTrue
    }
}

2. Describe Your Environment

Pester version     : 4.8.1 C:\Program Files\WindowsPowerShell\Modules\Pester\4.8.1\Pester.psd1
PowerShell version : 5.1.17763.316
OS version         : Microsoft Windows NT 10.0.17763.0

3. Expected Behavior

Tests must fail.

4.Current Behavior

Tests passed.

[renehernandez]

-BeTrue assertion will check for truthy equality, which as described here works not only for boolean values, but also for:

• Strings, as long as they aren't zero length string
• Any number different than 0
• Any non-zero length array
• Any non-null object

That left us with the 3 remaining cases that we may need to analyze:

Describe "Check if Truthy" {

    It "False as a string" {
        'False' | Should -Be $true
    }

    It "Any string" {
        'Whatever' | Should -Be $true
    }

    It "Any string" {
        'Whatever' | Should -BeExactly $true
    }
}

By looking at the code, it seems to be that this is due to the way the comparison are being performed within the -Be and -BeExactly assertions, which use -eq \ -ceq operators and it is comparing the expected value ($true) on the left-hand side against the received value (strings objects) on the right-hand side. Due to this, -eq \ -ceq will coerce the strings into a boolean value and since they aren't empty it will coerce them to the $true value, thus ending up performing a truthy comparison as well :).

[renehernandez]

Fixing this may affect people that are already relying on this unexpected behavior, so this could possible be a breaking change.

@nohwnd Any comments?

[nohwnd]

Yes, it would be a breaking change so it would need to be done in Pester v5. I want to re-visit the asseritons in v5 and I am now deciding if going in the direction of many breaking changes and better future behavior is better than keeping v5 more compatible to allow people to migrate to it.

[renehernandez]

For what it matters, I would vote to improve the future behavior and not be too concerned with the breaking changes. We are already doing breaking changes in v5, so I think we should leverage it as a relatively clean slate to revisit, fix and change as much as possible. Within reason of course :)

[nohwnd]

I think that as well. I will be talking about exactly this with the community on psconf and post summary afterwards so others can speak to it as well :)

[nohwnd]

Described here a possible way forward for should, would keep the new functionality specification for that change #1423