In principle, handling of memory failures is tested in the libzz. Though, maybe it does make sense to keep some basic tests here as well, to check that it really works for a Python extension.
But we need to find a better way for this. E.g. current testing affected by the presence of the hypothesis plugin:
$ pytest -q tests/test_memory.py
... [100%]
3 passed in 16.80s
$ pytest -p no:hypothesispytest -q tests/test_memory.py
... [100%]
3 passed in 30.10s
We could run tests in subprocesses instead and also use ulimit (available in MinGW) to set memory/stack limits.
See also #73.
In principle, handling of memory failures is tested in the libzz. Though, maybe it does make sense to keep some basic tests here as well, to check that it really works for a Python extension.
But we need to find a better way for this. E.g. current testing affected by the presence of the hypothesis plugin:
We could run tests in subprocesses instead and also use ulimit (available in MinGW) to set memory/stack limits.
See also #73.