BUG: A trivial C++ module crashes if you import it after pandas

[kapacuk]

Pandas version checks

• [X] I have checked that this issue has not already been reported.

• [X] I have confirmed this bug exists on the latest version of pandas.

• [x] I have confirmed this bug exists on the main branch of pandas.

Reproducible Example

cat >spam.cpp <<END

#include <Python.h>
#include <filesystem>
PyMODINIT_FUNC PyInit_spam(void)
{
    std::filesystem::path p1("/usr/share/icons");
    std::filesystem::path p2("hicolor");
    auto x = p1 / p2;
    return NULL;
}

END
g++-12 -I/usr/include/python3.9 -g -fPIC -std=c++20 -fsanitize=address -shared -o spam.cpython-39-x86_64-linux-gnu.so spam.cpp
python -c "import pandas; import spam"

Issue Description

The above example is a trivial python module with no classes or methods. It does not even try to create a PyModule object, so if you import it on its own it throws this exception: SystemError: initialization of spam failed without raising an exception which is expected.

However, if you import this module after pandas it crashes printing a single error message: munmap_chunk(): invalid pointer

I tried it with python3.9, 3.10, and 3.11, and with gcc-11 and gcc-12. It crashes consistently, although sometimes I get a different the error message.

Expected Behavior

That example should throw the SystemError exception regardless of whether you import it before or after pandas.

Installed Versions

INSTALLED VERSIONS ------------------ commit : 2e218d10984e9919f0296931d92ea851c6a6faf5 python : 3.9.2.final.0 python-bits : 64 OS : Linux OS-release : 5.10.0-13-amd64 Version : #1 SMP Debian 5.10.106-1 (2022-03-17) machine : x86_64 processor : byteorder : little LC_ALL : None LANG : None LOCALE : en_US.UTF-8 pandas : 1.5.3 numpy : 1.24.1 pytz : 2022.7.1 dateutil : 2.8.2 setuptools : 66.1.1 pip : 22.3.1 Cython : None pytest : None hypothesis : None sphinx : None blosc : None feather : None xlsxwriter : None lxml.etree : None html5lib : None pymysql : None psycopg2 : None jinja2 : None IPython : None pandas_datareader: None bs4 : None bottleneck : None brotli : None fastparquet : None fsspec : None gcsfs : None matplotlib : None numba : None numexpr : None odfpy : None openpyxl : None pandas_gbq : None pyarrow : None pyreadstat : None pyxlsb : None s3fs : None scipy : None snappy : None sqlalchemy : None tables : None tabulate : None xarray : None xlrd : None xlwt : None zstandard : None tzdata : None None

[Jython1415]

I'm struggling to reproduce this, but I suspect it's due to problems on my end...

Would you mind elaborating on the other error you get sometimes, and under what conditions you get the other error message?

[kapacuk]

Would you mind elaborating on the other error you get sometimes, and under what conditions you get the other error message?

I've just checked, I still get the munmap_chunk(): invalid pointer error with gcc-12 and gcc-13, but with gcc-10 the error is different:

$ python -c "import pandas; import spam"
double free or corruption (out)
Aborted
$

Please let me know if you still have problems reproducing it, I can build a Docker container for you.