Add (invalid) test cases of proportions == 0.0

[molpopgen]

See #166.
See also popsim-consortium/demes-python#576

[grahamgower]

Thanks for getting this rolling... I have a couple of additional test cases (for the defaults section), and the corresponding fixes for the reference implementation. I just ran out of time to commit and pull request yesterday. I can just add them on top of your PR, if you don't mind?

[molpopgen]

@grahamgower -- go for it!

[molpopgen]

@grahamgower -- I just rebased this onto main so that CI works.

[grahamgower]

@molpopgen, could you please run your eye over the changes I added here? If you're ok with it, I think it can be merged.

[molpopgen]

LGTM but not sure why some changes concern selfing/cloning rates?

[grahamgower]

In the schema definitions, we have (a) proportion for ancestry proportions and pulse proportions, where 0 < proportion <= 1, and (b) we have rate for migration rates, selfing/cloning, where 0 <= rate <= 1. The code in the reference implementation instead used the term 'fraction' and this was used for both proportions and rates, so I removed 'fraction' in favour of using the same terms and definitions as the schema.

[molpopgen]

Seems like this one is ready to go!