Reject non-function bindings for single-case and partial active pattern names

[Copilot]

The compiler previously rejected non-function multi-case active patterns ((|A|B|)) but still allowed equivalent invalid forms for single-case ((|A|)) and partial ((|A|_|)) names. This left declarations that tooling recognized as active patterns but could not be used in pattern positions.

• Checker behavior aligned across active pattern forms

  • Updated post-inference active-pattern validation to run the function-typed check for all active pattern bindings, not only multi-case ones.
  • This makes FS1209 (Active pattern '...' is not a function) consistently apply to single-case and partial declarations as well.

• Conformance coverage expanded

  • Extended E_ActivePatternNotAFunction.fs with explicit single-case and partial non-function declarations.
  • Updated expected diagnostics in Named.fs to assert all three FS1209 errors.

• Example of newly-rejected declarations

  let (|C|) = 3
  let (|D|_|) = None
  // now reported as FS1209: Active pattern '...' is not a function

[github-actions]

❗ Release notes required

---

✅ Found changes and release notes in following paths:

Change path	Release notes path	Description
src/Compiler	docs/release-notes/.FSharp.Compiler.Service/11.0.100.md

[T-Gro]

The fix is correct and well-scoped. Removing the when _apinfo.ActiveTags.Length > 1 guard causes doesActivePatternHaveFreeTypars to be called for all active pattern bindings, which triggers the existing FS1209 side-effect check (activePatternIdentIsNotFunctionTyped) for single-case and partial patterns too, while FS1210 (free typars) remains restricted to multi-case via the if condition.

Correctness — verified that stripFunTy on a non-function type returns ([], ty) with no spurious free-typar error, so the error cascade is clean.

Tests — the RelaxWhitespace2.fs fixups make sense: those active patterns were never valid for pattern matching, and making them function-typed aligns with the test's real purpose (whitespace parsing, not active-pattern semantics).

Minor note (pre-existing, not blocking): doesActivePatternHaveFreeTypars reports FS1209 as a side effect despite its name suggesting a pure predicate. The same function is called from PatternMatchCompilation.fs at use-sites (lines 651, 720, 1415, 1623), meaning a non-function active pattern can get a duplicate FS1209 at each use. This was already the case for multi-case patterns before this PR, so it's not a regression — just something to be aware of for a future cleanup pass that separates the validation concern from the predicate.