chore(python): cover overloaded-ctor casing on lowercase types

[dbrattli]

Follow-up to #4807.

1. Regression test for the overload-suffix casing path

#4807 fixed Python class/member declarations and references disagreeing on casing for lowercase-named types, but its test (type shape with a single constructor) never exercised the overload-suffix path called out in the PR description (shape__ctor_Z524259A4 declared but shape__ctor_z524259a4 imported).

This gives shape a second, overloaded constructor and references both across files. Transpiled output confirms declaration and import agree on the lowercased suffix:

# Misc/util2.py
def shape__ctor_z524259a4(size: int32) -> Shape: ...
def shape__ctor_z721c83c5(label: str) -> Shape: ...
# test_misc.py
from Misc.util2 import (..., shape__ctor_z524259a4, Shape as Shape_1, shape__ctor_z721c83c5, ...)

Before the #4807 fix, the declaration kept ...Z524259A4 while the import used ...z524259a4 → ImportError.

2. Accurate comments on the composed-name toPythonNaming

While reviewing the double toPythonNaming application in getMemberDeclarationName, I traced (via quicktest) exactly what the unconditional outer call does:

• It mirrors the re-casing applied at import/reference sites, which is what keeps declarations and references consistent (the whole point of fix(python): make class declarations and references agree on type names #4807).
• As a side effect, being unconditional it also re-cases [<CompiledName>] members. This is a no-op for class/nested-module members (uppercase entity/module prefix), so their compiled name survives — but a root-level [<CompiledName>] function starting lowercase gets snake_cased ([<CompiledName("rootLevelCompiled")>] → root_level_compiled).

This is a pre-existing tension, not something this PR changes: cross-file references to such a function were already broken before #4807 (ImportError), so #4807 traded "compiled name honored but cross-file broken" for "always consistent but not honored verbatim". A fully correct fix would need the import printer to be CompiledName-aware (it only sees a bare string today) — out of scope here. This PR just documents the real behavior instead of the earlier misleading "idempotent, harmless" note.

No compiler behavior change — comments only.

Testing

./build.sh test python — all 2405 tests pass (run against the initial version of this PR; item 2 is comments-only).

🤖 Generated with Claude Code

[github-actions]

Python Type Checking Results (Pyright)

Metric	Value
Total errors	34
Files with errors	4
Excluded files	4
New errors	✅ No

Excluded files with errors (4 files)

These files have known type errors and are excluded from CI. Remove from pyrightconfig.ci.json as errors are fixed.

File	Errors	Status
temp/tests/Python/test_hash_set.py	18	Excluded
temp/tests/Python/test_applicative.py	12	Excluded
temp/tests/Python/test_nested_and_recursive_pattern.py	2	Excluded
temp/tests/Python/fable_modules/thoth_json_python/encode.py	2	Excluded