Description

The pyutils-unusedcode tool crashes with a subprocess.CalledProcessError when it encounters pytest fixtures that are used as parameters rather than called directly.

Environment

• Tool: pyutils-unusedcode from python-utility-scripts
• Python: 3.12
• OS: Linux

Steps to Reproduce

1. Create a pytest fixture in a test file:

   @pytest.fixture
   def sample_clusters() -> list[ClusterInfo]:
       return [...]

2. Use the fixture as a parameter in test functions:

   def test_something(sample_clusters: list[ClusterInfo]) -> None:
       # Use the fixture
       assert len(sample_clusters) == 2

3. Run pyutils-unusedcode on the codebase

Expected Behavior

The tool should recognize that pytest fixtures are used as parameters and either:

• Skip them automatically
• Handle the git grep exit code gracefully
• Continue processing other functions

Actual Behavior

The tool crashes with:

subprocess.CalledProcessError: Command '['git', 'grep', '-wE', 'sample_clusters(.*)']' returned non-zero exit status 1.

Full Stack Trace:

Traceback (most recent call last):
  File "/path/to/pyutils-unusedcode", line 10, in <module>
    sys.exit(get_unused_functions())
  File "/path/to/click/core.py", line 1442, in __call__
    return self.main(*args, **kwargs)
  [... stack trace continues ...]
  File "/path/to/unused_code.py", line 78, in process_file
    _func_grep_found = subprocess.check_output(["git", "grep", "-wE", f"{func.name}(.*)"], shell=False)
  File "/path/to/subprocess.py", line 466, in check_output
    return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
  File "/path/to/subprocess.py", line 571, in run
    raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command '['git', 'grep', '-wE', 'sample_clusters(.*)']' returned non-zero exit status 1.

Root Cause

1. The tool uses git grep -wE 'function_name(.*)' to find function usage
2. Pytest fixtures are used as parameters (e.g., def test_foo(sample_clusters):), not called with parentheses
3. When git grep finds no matches, it returns exit code 1
4. The tool uses subprocess.check_output() which raises an exception on non-zero exit codes

Suggested Fix

In unused_code.py line 78, handle the case where git grep returns no matches:

try:
    _func_grep_found = subprocess.check_output(
        ["git", "grep", "-wE", f"{func.name}(.*)"], 
        shell=False,
        stderr=subprocess.DEVNULL
    )
except subprocess.CalledProcessError as e:
    if e.returncode == 1:  # No matches found
        _func_grep_found = b""  # Treat as empty result
    else:
        raise  # Re-raise for other errors

Workaround

Add pytest fixture names to the exclusion list:

pyutils-unusedcode --exclude-function-prefixes "fixture_name1,fixture_name2"

Additional Context

This issue commonly occurs with pytest fixtures, which are a standard testing pattern. The tool should handle this case gracefully rather than requiring manual exclusions for every fixture.

The tool should ideally:

1. Detect pytest fixtures automatically (functions decorated with @pytest.fixture)
2. Use different search patterns for fixtures (look for parameter usage instead of function calls)
3. Handle git grep exit codes gracefully for any function that might not have matches