Halfspace Strain solution contains Nans

[hvasbath]

The same as with the displacement solution. Fix was the same setting cosB to EPS in case of zero. However, I couldnt find a triangle configuration that raises in a test. Found the Nans in a more complex setup. So sending here only a PR no updated matlab test. My complex setup works after that.

[tbenthompson]

Thanks!

[tbenthompson]

@hvasbath I'll push a new release of cutde this weekend now that these bugs are all fixed up. Hope it's working well for your problems now!

[hvasbath]

Sounds good! For now I haven't stumbled upon another issue. Fingers crossed it stays like this 😉. The forward model is working, now lets see how the inversion goes. Without cutde this would be impossible to do in reasonable runtime. Again thanks a lot for putting it together!

[tbenthompson]

And thank you for finding the bugs!

[tbenthompson]

Just a heads up that I pushed a new version to PyPI. The pip install should now use the new version. In a few hours, the conda package will be updated too - that happens almost automatically but has a bit of time delay.