(#2398) Restore PowerShell v2 support

[corbob]

Description Of Changes

Commit c408d12 introduced a change to fix a bug with the $env:TEMP variable. As well, commit f68a242 introduced handling of msps in the install helper.

However, they both introduced PowerShell 3+ syntax. This changes those to work with PowerShell 2+.

Motivation and Context

We currently support PowerShell 2+, so it's important that the PowerShell uses v2 compatible commands.

What Have I Done To Test This

This one is specific to PowerShell 2 and the import of the helpers\chocolateyInstaller.psm1. It can be tested either on a Windows 7 with PowerShell 2 system (this is a multi-hour waiting game to build and update a system if you don't already have one), or on a newer Windows system with the PowerShell 2 compatibility feature enabled by running powershell -version 2.

From PowerShell 2 run: Import-Module src\chocolatey.resources\helpers\chocolateyInstaller.psm1 -force from the root of the repository. This should result in no errors and no output.

Change Types Made

• Bug fix (non-breaking change)
• Feature / Enhancement (non-breaking change)
• Breaking change (fix or feature that could cause existing functionality to change)

Related Issue

Fix #2398

Change Checklist

• Requires a change to the documentation
• Documentation has been updated
• Tests to cover my changes, have been added
• All new and existing tests passed.

[corbob]

This was initially in draft as I found the second commit while doing up the testing for it as I found it was still giving the error after fixing the first one. It's ready now.

[gep13]

LGTM!

[vexx32]

@corbob is there a reason you used -inotcontains instead of -notcontains for one of these? 🤔

[gep13]

@vexx32 that is a great question!

[gep13]

Based on this:

https://iannotes.wordpress.com/2011/12/25/powershell-comparison-operators/

They are equivalent, but it would be better to have them the same, right?

[vexx32]

Yep, they should be equivalent, but using different ones might confuse maintainers coming back to these code paths in the future.

[corbob]

Oh duh. For some reason I was thinking -notcontains would default to case sensitive, and for the one where we're comparing username it seemed reasonable to force it to case insensitive. Given that they're both the same, I'll change it back to -notcontains for the consistency.

[corbob]

There we go, should be updated now 😄

[gep13]

LGTM!

[gep13]

@corbob thanks for getting this fixed!