TL;DR

• Mechanical migration from Pester 4.99 → 5.7.1 across 207 test files and 2 build/CI files.
• No test logic, assertions, or expected behaviors were changed. Most edits move setup and data into BeforeDiscovery / BeforeAll blocks for Pester 5's execution model.
• Test coverage is effectively retained (14,696 → 15,264 discovered tests). P5 actually discovers 568 more tests than P4 could. 3 of 4 P4 hangs resolved.

How to review: Most diffs are mechanical — look for the pattern of data moving into BeforeDiscovery {} and setup moving into BeforeAll {}. The two non-test files (build.psm1, tools/ci.psm1) contain the Pester version bump and invocation changes.

Why

Pester 4 is no longer maintained. This updates PowerShell's test infrastructure to the supported framework version and aligns the suite with Pester 5's discovery/execution model.

Before → After

Both baselines captured on Windows, same machine, same environment, 600s per-file timeout.

P5 runs 568 more tests and resolves 3 of 4 P4 hangs. The higher failed/skipped counts come from tests that P4 couldn't enumerate now being properly discovered — many are environment-dependent (remoting, services, admin rights) and correctly skip or fail. 342 files have identical test counts between P4 and P5.

What changed

Build / CI (2 files)

Test files (207 files)

The dominant migration pattern adapts to Pester 5's discovery/run phase separation:

# BEFORE (Pester 4) — data at script scope, available everywhere
$testCases = @(
    @{ Name = 'foo'; Expected = 1 }
)
Describe "Tests" {
    It "works with <Name>" -TestCases $testCases { ... }
}

# AFTER (Pester 5) — data in BeforeDiscovery, setup in BeforeAll
BeforeDiscovery {
    $testCases = @(
        @{ Name = 'foo'; Expected = 1 }
    )
}
Describe "Tests" {
    BeforeAll {
        # runtime setup that was previously at script scope
    }
    It "works with <Name>" -TestCases $testCases { ... }
}

Applied across 207 files:

What did NOT change

• ✅ All test assertions (Should -Be, Should -Throw, etc.) — unchanged
• ✅ Test logic and expected behaviors — unchanged
• ✅ No tests were deleted or force-skipped to make the migration pass
• ✅ File structure and naming — unchanged
• ✅ 342 files have identical test counts between P4 and P5

Scope by area

Possible improvements (not in this PR)

• Wait-Process.Tests: genuine hang in both versions. Adding -Timeout 30 to Wait-Process calls would fix it.
• CompatiblePSEditions.Module.Tests: 107 failures from test environment lacking Windows PowerShell 5.1 WinCompat modules. Tests discover and run correctly.