Short edges

[dgud]

The problem is that to small lengths will eventually cause a crash somewhere since the error will propagate and become larger for each computation.
That is the fact of using float(or in our case doubles) when doing computations.

The simple way to avoid this is to avoid to small numbers which is why it is done the way it is.
The inconsistency is not nice but comes from different developers having fixed it in different ways,
when the bug reports have come. And we have missed it for some functionality, you are welcomed
to send a patch which fixes this consistently all over the code if you want.