subprocess.CalledProcessError: Command '['dot', '-Tsvg']' died with <Signals.SIGSEGV: 11>

[pawamoy]

Describe the bug

Running dot in a subprocess from Python in GHA results in this error:

    Traceback (most recent call last):
      File "<code block: session comment_blocks; n1>", line 25, in _render_call_graph
      File "/opt/hostedtoolcache/Python/3.11.9/x64/lib/python3.11/subprocess.py", line 466, in check_output
        return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
      File "/opt/hostedtoolcache/Python/3.11.9/x64/lib/python3.11/subprocess.py", line 571, in run
        raise CalledProcessError(retcode, process.args,
    subprocess.CalledProcessError: Command '['dot', '-Tsvg']' died with <Signals.SIGSEGV: 11>.

To Reproduce

I'm not able to reproduce it with a simple example. I tried this:

    - name: Reproduce issue
      run: python -c 'import subprocess; print(subprocess.check_output(["dot", "-Tsvg"], input="digraph D {\n  A -> B\n}", text=True))'

...but it works fine. Maybe the actual contents are large enough to make dot request too much memory or something? Completely out of my depth here.

The error can be seen here: https://github.com/mkdocstrings/griffe/actions/runs/9922970574/job/27412614124. Scroll down a bit to find a WARNING log with the above traceback.

Expected behavior

No segfault.

Additional context

Just opening for visibility in case someone else experiences the same issue. I searched on the web, and other users have this issue, but generally with programs like PyTorch and CUDA, so GPU stuff. Not sure why our lovely dot here suffers from the same fate. Feel free to close since I'm not able to build an MRE.