Log file remains locked after spdlog::shutdown() in Jupyter notebook with SWIG-wrapped C++ library

[ArnaultT]

Hi,

First of all, thank you so much for your work! The open-source community is truly amazing.
I’m using spdlog in a C++ library that is wrapped in Python using SWIG. As a scientist, I work with a lot with notebooks and Jupyter Lab.

My program creates an instance of a "Simulation" class, which contains a shared_ptr to a custom logger class that just encapsulate spdlog stuff.

At the start of the simulation, I initialize the logger, and once the simulation is complete, I call spdlog::shutdown(). All of this is hidden from the Python user.

The issue is that the log file remains locked even after calling spdlog::shutdown(). I can’t delete or move it. The only way to unlock the file is to kill the notebook kernel. I’ve also tried deleting the Python variable holding the Simulation object, but that doesn’t help.
I suspect this might be a Jupyter-specific issue, but do you have any ideas on how to resolve it? I wonder if it’s related to memory management, but I’m not sure.

Thanks a lot for your help!

[tt4g]

spdlog does not acquire locks for files.
Isn't the wrapper library acquiring locks for the log files?

[ArnaultT]

Yes! I found the problem—thank you for your help in pointing me in the right direction!
The issue was that my Simulation instance had been copied into another class. As a result, I had two instances of the simulation in Python, not just one. After deleting both instances, the log file was correctly released and could be deleted.