Add support for multiple aliases (-Alias) to Should -HaveParameter

[dtewinkel]

PR Summary

Add support for multiple mandatory aliases for -HaveParameter. -Alias parameter was changed from a String to a String[].

Allows for Should -HaveParameter Parameter -Alias Alias1, Alias2 to test if parameter Parameter has both the aliases alias1 and Alias2. If either one is missing the assert fails.

Fix #2152

[dtewinkel]

@fflaten, @nohwnd , I see the PR build breaks due to a previously introduced bug in file tst\functions\Add-ShouldOperator.ts.ps1, where it runs into the more than 32 assertion operators limitation. This builds fail since a merge to main on the 3rd of October.
I added a fix, or work-around, to the tests to prevent running into the limitation.

[fflaten]

Suggested change

	foreach ($AliasValue in $Alias) {
	$testPresenceOfAlias = $ActualValue.Parameters[$ParameterName].Aliases -contains $AliasValue
	$filters += "to$(if ($Negate) {" not"}) have an alias '$AliasValue'"
	$filters += "with$(if ($Negate) {'out'}) alias$(if ($Alias.Count -ge 2) {'es'}) $(Join-And ($Alias -replace '^|$', "'"))"
	foreach ($AliasValue in $Alias) {
	$testPresenceOfAlias = $ActualValue.Parameters[$ParameterName].Aliases -contains $AliasValue

Maybe move the filter-text outside foreach to improve readability? Thoughts?
We can utilize helpers in Format.ps1 like Join-And or Format-Collection to make it pretty.

Current PR:

Expected command Invoke-DummyFunction to have a parameter MandatoryParam, to have an alias 'First' and to have an alias 'Second', but it didn't have an alias 'Second'.

Join-And (suggestion):

Expected command Invoke-DummyFunction to have a parameter MandatoryParam, with aliases 'First' and 'Second', but it didn't have an alias 'Second'.

Format-Collection:

Expected command Invoke-DummyFunction to have a parameter MandatoryParam, with aliases @('First', 'Second'), but it didn't have an alias 'Second'.

$filters += "with$(if ($Negate) {'out'}) alias$(if ($Alias.Count -ge 2) {'es'}) $(Format-Collection $Alias)"

[dtewinkel]

Thanks for the suggestions. They help to make the output more readable. I applied them, including to the formatting of the wrong aliases.

[fflaten]

Thanks for submitting the PR!

Also thanks for pointing out the CI-issue! I've submitted a separate PR (#2248) to split the recent help-tests into its own block and to keep this PR clean. Would you mind reverting the workaround? 😇

[dtewinkel]

Thanks for submitting the PR!

Also thanks for pointing out the CI-issue! I've submitted a separate PR (#2248) to split the recent help-tests into its own block and to keep this PR clean. Would you mind reverting the workaround? 😇

Sure, no problem :-). I'll revert it as soon as it is merged to master. For now, it helps me to show that PS3 and PS4 are failing, and I need to figure out why.

[dtewinkel]

Thanks for submitting the PR!

Also thanks for pointing out the CI-issue! I've submitted a separate PR (#2248) to split the recent help-tests into its own block and to keep this PR clean. Would you mind reverting the workaround? 😇

Revert is done.

[fflaten]

Revert is done.

Thanks! Closed/reopened to trigger a working CI-run.
LGTM 👏 Waiting on @nohwnd

[nohwnd]

My only concern here is that this changes "data" of another test, would it be possible to keep those functions as they are, and introduce a new function that has two aliases.

[dtewinkel]

All tests share two functions, one with all possible options that need to be validated, and an empty one, so that was the road I continued on.
I validated that the tests around aliases were still valid for what they were testing. I believe that actually sharing the data in this case 'enhances' the tests without having to add a number of functions to validate all kinds of scenarios. It now also validates that if -Alias First is tested, that it succeeds even if there are multiple aliases.

So, my question is, is your concern a general one and was the setup of the tests perhaps not right in the first place, because they all share the same two functions here, or is there a more specific concern here that you need me to address?

[nohwnd]

It was a general concern about sharing test data than eventually become so complicated that no-one knows what is going to break. This is not severe here so let's allow it.

[nohwnd]

@dtewinkel Thank you :)