Fix flaky Debug-Runspace attach event test - #27883

PR Summary

Fixes the flaky Debug-Runspace test Should write attach event and mark runspace as having a remote debugger attached, which intermittently fails in CI at:

$waitTask.AsyncWaitHandle.WaitOne(5000) | Should -BeTrue
Expected $true, but got $false.

Root cause (test-side race, not a product bug)

The test did:

$debugTarget = [PowerShell]::Create()          # runspace created but NOT opened
$null = $debugTarget.AddCommand('Wait-Event')...
$waitTask = $debugTarget.BeginInvoke()         # opens runspace + starts pipeline asynchronously
$debugger.AddCommand('Debug-Runspace').AddParameter('Id', $debugTarget.Runspace.Id)
$debugTask = $debugger.BeginInvoke()           # starts immediately, no synchronization

BeginInvoke only queues work. Nothing guaranteed the target runspace had finished opening before Debug-Runspace executed. When Debug-Runspace won the race, either:

1. DebugRunspaceCommand.EndProcessing operated on a runspace that was not Opened, producing a terminating InvalidOperation error before ever reaching the event generation, or
2. _runspace.Events?.GenerateEvent(...) (DebugRunspaceCommand.cs:269) was skipped by the null-conditional because the runspace's execution context / event manager was not yet available.

Either way OnDebugAttach was never delivered and the 5-second wait timed out.

Note the fix must gate on runspace readiness, not merely on subscriber registration: an event generated before Wait-Event reaches ProcessRecord is not lost, because WaitEventCommand.ScanEventQueue drains Events.ReceivedEvents on entry.

Fix

Test-only changes — no production code modified:

• Explicitly create and Open() the target runspace and assign it to the PowerShell instance, eliminating the "runspace not yet open" window and guaranteeing a live event manager.
• Before invoking Debug-Runspace, gate on real readiness with the existing Wait-UntilTrue helper: the pipeline's InvocationStateInfo.State is Running and the runspace's RunspaceAvailability is Busy. This is a bounded readiness gate, not a fixed sleep or an inflated timeout.
• Wait deterministically for IsRemoteDebuggerAttached to return to $false, since the cmdlet resets it while unwinding asynchronously with respect to Stop().
• Wrap the body in try/finally so the debugger PowerShell, target PowerShell, and the runspace are stopped and disposed even when an assertion fails.

All four original assertions are preserved: OnDebugAttach received, IsRemoteDebuggerAttached becomes $true, stopping throws PipelineStoppedException, and IsRemoteDebuggerAttached returns to $false.

PR Context

Removes a source of intermittent Windows CI failures.

PR Checklist

• PR has a meaningful title
• Summarized changes
• Make sure all .h, .cpp, .cs, .ps1 and .psm1 files have the correct copyright header
• This PR is ready to merge and is not work in progress
  • User-facing changes: N/A (test-only)
  • Testing - New and existing tests pass

Test Results

Built locally (Start-PSBuild, Windows, Debug) and run against the fresh pwsh:

Describing Debug-Runspace
  [+] Debugging a runspace should fail if the name is ambiguous
  [+] Debugging a runspace should fail if the name is not found
  [+] Debugging a runspace should fail if the runspace is not open
  [+] Debugging a runspace should fail if the runspace has no debugger
  [+] Should write attach event and mark runspace as having a remote debugger attached
Tests Passed: 5, Failed: 0

Repeated 10 consecutive standalone runs: P=5 F=0 every time.

Via the repo harness, including the closely related debugger suite:

Start-PSPester -Path test/.../Debug-Runspace.Tests.ps1, test/.../RunspaceBreakpointManagement.Tests.ps1
Tests Passed: 25, Failed: 0, Skipped: 0, Pending: 0, Inconclusive: 0