Fixing nested timeout contexts using both thread and signal based Timeout contexts

[mmatera]

This PR would provide a more complete fix for #17, handling timeouts with both kinds of Timeout context managers.

[kod-kristoff]

@mmatera I started fixing issues with this repo here: https://github.com/spraakbanken/stopit2
My plan is to release this as stopit2 to PyPi.

I am interested of including this PR (I recreated it here: spraakbanken#5 including adapting code to my changes of the project)
Does that sound ok? Have you used you fork, and if so does it work as intended?

[mmatera]

@mmatera I started fixing issues with this repo here: https://github.com/spraakbanken/stopit2 My plan is to release this as stopit2 to PyPi.

I am interested of including this PR (I recreated it here: spraakbanken#5 including adapting code to my changes of the project) Does that sound ok? Have you used you fork, and if so does it work as intended?

Hi, @kod-kristoff. Sorry for the delay in the answer. I am glad to know that you are working on update this project.
Regarding Mathics3, what we are using now is the version in this fork: https://github.com/Mathics3/stopit, but it would be nice to have a compatible version in a Pypi repo.

This is how we are using it:
https://github.com/Mathics3/mathics-core/blob/master/mathics/eval/datetime.py

and works in ubuntu, windows and osx. What is not working is in "emscripten", a platform for which it would be nice to port it too.