Remove vestigial block setup/teardown

[nohwnd]

Draft, stacked on #2907 (which is stacked on #2906). Follow up to Frode's comment on #2907.

The block level setup/teardown is dead. New-EachBlockSetup, New-EachBlockTeardown and New-OneTimeBlockTeardown (and the never set New-OneTimeBlockSetup that #2907 already removed) are only ever called from the runtime tests. Real BeforeAll / BeforeEach / AfterAll / AfterEach go through the *Test* variants (New-OneTimeTestSetup, New-EachTestSetup, ...), so in a real run the Block.EachBlockSetup / OneTimeBlockSetup / EachBlockTeardown / OneTimeBlockTeardown properties are always null and the two reads in Invoke-Test are no-ops.

Removes:

• the three functions
• the four ScriptBlock properties on Pester.Block
• the two always null reads in Invoke-Test
• the runtime tests that exercised them. The two data availability tests keep their teardown check by switching to New-OneTimeTestTeardown.

The plugin block lifecycle (OneTimeBlockSetupStart, EachBlockTeardownEnd, ...) is a separate, live mechanism and is untouched.

These are genuinely dead so I deleted them rather than reaching for [ExcludeFromCodeCoverage()].

Minor breaking change, fine for 6.1.0.

Verification

Full test.ps1 run passes.

[fflaten]

About OneTimeBlock*, see #1219. Is BeforeAll -Context still planned? If so, we can't remove them after all.

[nohwnd]

Good point, let's keep them. The EachBlockSetup/EachBlockTeardown side is actually already executed in Invoke-Test (the outer setup/teardown around each child block), it just has no public way to set it, so it is scaffolding for #1219 rather than dead. OneTimeBlock* is not wired at runtime yet. Closing this, I'll look at wiring up #1219 (BeforeEach/AfterEach -Context) instead.