Potential duplicate test or missing typing.Optional coverage in test_type_conversion.py

[phalberg]

First Check

• I added a very descriptive title here.
• I used the GitHub search to find a similar question and didn't find it.
• I searched the documentation, with the integrated search.
• I already searched in Google "How to X" and didn't find any information.
• I already read and followed all the tutorials in the docs and didn't find an answer.

Commit to Help

• I commit to help with one of those options 👆

Example Code

def test_optional():
    app = typer.Typer()

    @app.command()
    def opt(user: str | None = None):
        if user:
            print(f"User: {user}")
        else:
            print("No user")

    result = runner.invoke(app)
    assert result.exit_code == 0
    assert "No user" in result.output

    result = runner.invoke(app, ["--user", "Camila"])
    assert result.exit_code == 0
    assert "User: Camila" in result.output


def test_union_type_optional():
    app = typer.Typer()

    @app.command()
    def opt(user: str | None = None):
        if user:
            print(f"User: {user}")
        else:
            print("No user")

    result = runner.invoke(app)
    assert result.exit_code == 0
    assert "No user" in result.output

    result = runner.invoke(app, ["--user", "Camila"])
    assert result.exit_code == 0
    assert "User: Camila" in result.output

Description

Hello,
I read trough some of the code for typer, namely more concretely in tests/test_type_conversion.py where it seems to start with this [see above for the code] which are currently identical test with different names. I did some digging and when support was dropped for python 3.9, the conversion: Optional[str] might have resulted in str | None, see #1546, by the help of some tool like ruff or something.
Before opening a PR I wanted to make sure I am aligned, as I see it there are two options:

1. Remove test_optional() as it is no longer needed or,
2. Adjust test_optional() to use def opt(user: Optional[str] = None): (and import it)

I would think reverting back to Optional is perhaps not the preferred option, but it is still widely used, so just making sure!

Operating System

Windows

Operating System Details

No response

Project Version

v.0.26.7

Python Version

3.10.11

Additional Context

No response

[laypatel13]

Hey @phalberg,

I looked into this and traced it through PR #1546. Originally, test_optional() tested the legacy Optional[str] syntax, while test_union_type_optional() was gated behind @needs_py310 and tested Optional[tuple[int, int]] — a genuinely different case. That PR correctly modernized both to | syntax as part of dropping Python 3.9 support.

At some point after #1546, test_union_type_optional()'s content changed to test str | None as well, which is where the duplication actually comes from — not #1546 itself.

Your Option 2 seems like the right direction: adjusting test_optional() to explicitly use Optional[str] = None would restore its original purpose (verifying the legacy typing.Optional syntax still works), while test_union_type_optional() continues to cover the modern | syntax. That keeps both tests distinct rather than duplicated.

I'd lean against just deleting test_optional() (Option 1) — even though both tests currently check the same runtime behavior, they're meant to guard two separate code paths. Removing it would leave no coverage for the old-style Optional[str] syntax, so if that path ever broke, tests would still pass while real users on Optional[str] wouldn't.

If it's helpful, I'd be glad to propose a PR for this fix.

[laypatel13]

Quick update - while checking into this issue as mentioned, and testing it on my local machine, I came across @svlandeg's comment in #1855 about an active refactor of type resolution happening in Typer right now.

Since this touches type annotations in the same test file, just wanted to mention it in case it's relevant timing-wise before this goes any further.