Potential codegen bug with MinGW-w64 in spread_sorter tests

[Morwenn]

Since very recently the test spread_sorter_defaults.cpp segfaults on my computer (currently running MinGW-w64 GCC 10, only in Release mode), and considering how printing additional information in this test makes the bug disappear, I fear that it might be a codegen bug or something along these lines that I can't do much about. I wouldn't be the first one on this platform.

This doesn't affect CI at all, so unless it becomes a bigger issue it will probably remain unfixed for a while.