Assertion error when running `basilisp test` wihtout speicfying a path

[ikappaki]

Hi

there appears to be an general issue with Basilisp 0.3.6 when running basilisp test without explicitly specifying a path:

AssertionError: Must have at least one module name

To reproduce,

1. create a new Poetry Basilisp project, and create a tests/issue_test.lpy file with a minimal test case (or clone https://github.com/ikappaki/issue-bas-test-path-dot)

(ns tests.issue-test
  (:require [basilisp.test :refer [deftest is]]))

(deftest abc
  (is (= 5 0)))

2. Run basilisp test the above error is returned:

$ poetry run basilisp test
=================================================== test session starts ===================================================
platform win32 -- Python 3.11.4, pytest-8.3.4, pluggy-1.5.0
rootdir: d:\bas\issue-bas-test-path-dot
configfile: pyproject.toml
plugins: basilisp-0.3.6
collected 0 items / 1 error

__________________________________________ ERROR collecting tests/issue_test.lpy __________________________________________
.venv\Lib\site-packages\basilisp\contrib\pytest\testrunner.py:270: in collect
    module = self._import_module()
.venv\Lib\site-packages\basilisp\contrib\pytest\testrunner.py:247: in _import_module
    assert modnames, "Must have at least one module name"
E   AssertionError: Must have at least one module name
E   assert []
...

A workaround is to include the FQD name of the project directory when running basilisp test:

$ poetry run basilisp test -p <project dir>
platform win32 -- Python 3.11.4, pytest-8.3.4, pluggy-1.5.0
rootdir: d:\bas\issue-bas-test-path-dot
configfile: pyproject.toml
plugins: basilisp-0.3.6
collected 1 item

tests\issue_test.lpy F                                                                                               [100%]

___________________________________________________________ abc ___________________________________________________________
FAIL in (abc) (issue_test.lpy:5)
    Test failure: (= 5 0)

    expected: 5
      actual: 0
...

This issue seems to be related to a regression in (#1176), which I thought I previously tested with success, but it was in a slightly more complex setup that did not cover this basic case.

PR to follow.

Thanks