[rahul05ranjan]

Fixes #215.

Problem

The timeout marker registration documents a disable_debugger_detection keyword (it also appears in pytest --markers output), but _parse_marker() has no branch for it, so:

@pytest.mark.timeout(60, disable_debugger_detection=True)
def test_foo(): ...

kills the whole session with INTERNALERROR> TypeError: Invalid keyword argument for timeout marker: disable_debugger_detection.

The CLI flag and ini setting of the same name work; only the documented marker keyword is broken.

Fix

Parse disable_debugger_detection in _parse_marker() as a keyword-only argument and include it in the returned Settings, so the existing _validate_disable_debugger_detection() validation in _get_item_settings() applies to it like the other keyword arguments. Unknown keywords are still rejected.

Verification

• Before the fix, the reproduction above fails the session with the INTERNALERROR above; after the fix the test passes.
• New regression test test_timeout_marker_disable_debugger_detection_keyword covers the keyword both alone and combined with method/func_only via pytester.runpytest_subprocess().
• Direct unit check: _parse_marker(pytest.mark.timeout(60, disable_debugger_detection=True)) now returns Settings(timeout=60, method=None, func_only=None, disable_debugger_detection=True), while pytest.mark.timeout(10, bogus=True) is still rejected with TypeError.
• Note: HEAD's parser.addini(..., type="float") requires pytest >= 8.4, and my system pytest is 8.3.5, so I could not run the full suite locally as-is; I verified the end-to-end behavior with a local one-line addini shim (coercing the new "float" ini type, unrelated to this change) on top of this patch. The new pytester test runs in the project's normal CI matrix.

[flub]

I'd appreciate if you do not let LLMs write the text of issues. It's really hard to read. Please put in the effort to communicate in your own words.

[flub]

Please read the commit message, this needed only two lines of explanation to communicate clearly.