Optimize formatting for template strings and numbers

[fflaten]

PR Summary

Restricts template strings to one level of object recursion in formatting to avoid hangs on large complex objects.

To address the issue above, this PR also:

• Refactors Format2 to use a decreasing and internally configurable -MaxDepth instead of increasing -Depth with hard limit. Easier to restrict something a max 1 level than start at depth 9.
• Uses Format-Numbers* for integers and not just decimals. Integers was formatted as ValueType, requring -MaxDepth 2 for formatting simple arrays.
  • Same applies to [char] vs [string] etc, but not fixed for now.

Fix #2865

PR Checklist

• PR has meaningful title
• Summary describes changes
• PR is ready to be merged
  • If not, use the arrow next to Create Pull Request to mark it as a draft. PR can be marked Ready for review when it's ready.
• Tests are added/update (if required)
• Documentation is updated/added (if required)

[fflaten]

This will still print a massive multi-line name for something like (Get-Command Invoke-Pester), but at least it doesn't hang. Questions:

• Should we special case CommandInfo to just return the name like v5? Or let it be so the author will identify a bad practice (not referencing the name property) and fix it.
• Should tests names expand arrays/dictionaries etc in general?
• Should tests names allow multi-line at all? Messy console output. Trim or run through [Pester.Formatter]::EscapeControlChars()?

[fflaten]

Just cosmetic to get consistent naming in code and tests <Type.Name> for both PS 5.1 and 7+

[nohwnd]

Hey @fflaten 👋 I pushed a commit onto this branch to iterate on it — hope that's ok, happy to revert if you'd rather take it a different way.

Your -MaxDepth refactor + integer fix are exactly the right foundation, so I built on top rather than starting over. The one thing left over was that a complex object referenced whole in a name (e.g. <cmd> instead of <cmd.Name>, where cmd is a CommandInfo) still expanded one full level — for a CommandInfo that's the entire Definition and ends up ~80k chars, so the name is huge even though it no longer hangs.

The commit finishes the type-map idea that was already sketched in the Get-DisplayProperty2 comment: match by base type (one CommandInfo entry covers FunctionInfo/CmdletInfo/AliasInfo/ExternalScriptInfo/…) and register CommandInfo -> Name. Format-Nicely2 consults the registry before expanding an object, so a registered type renders a compact summary in both assertion messages and test-name templates:

<cmd>  ->  Management.Automation.FunctionInfo{Name='Invoke-Pester'}

Anything not in the registry keeps today's behaviour (full expansion in assertion messages), and in a name it just collapses to [SomeType] so an unknown object can never blow up a test name. Taming a new noisy type is then a one-line registry entry — that's the "least maintenance" part.

Added tests for the base-type match, the compact summary, and an RSpec acceptance test for the #2865 repro. Clean build is green; only the unrelated pre-existing SkipRemainingOnFailureCount failures remain. Let me know what you think!