The problem is that our py-kms tries to use an invalid shell (which is provided by docker or even systemd sometimes) and therefore loops indefinitely. This patch forces docker to provide a valid shell (see here), which should be a acceptable fix for the problem.
The best solution would be to modify py-kms directly by checking at startup for a valid shell (this would then fix this bug for any problematic environment), but this can be really hard (because sometimes we would have a valid shell but e.g. the /dev/null as input).
This addresses issue #38 - but for docker only.
The problem is that our py-kms tries to use an invalid shell (which is provided by docker or even systemd sometimes) and therefore loops indefinitely. This patch forces docker to provide a valid shell (see here), which should be a acceptable fix for the problem.
The best solution would be to modify py-kms directly by checking at startup for a valid shell (this would then fix this bug for any problematic environment), but this can be really hard (because sometimes we would have a valid shell but e.g. the /dev/null as input).