Fwdpy11

Closes a set of nagging issues: 246, 280, 339, 365, 386, 397

This is a major feature release. The changes include all those listed for the various
release candidates (0.6.0rc0, 0.6.0rc1, and 0.6.0rc2) plus the following:

* Several back-end issues are fixed:
[Issue 388](https://github.com/molpopgen/fwdpy11/issues/388)
[Issue 389](https://github.com/molpopgen/fwdpy11/issues/389)
[Issue 390](https://github.com/molpopgen/fwdpy11/issues/390)
[Issue 392](https://github.com/molpopgen/fwdpy11/issues/392)
* fwdpy11.TableCollection.fs added.
[PR387](https://github.com/molpopgen/fwdpy11/pull/387)
[PR399](https://github.com/molpopgen/fwdpy11/pull/399)
* Creating populations from msprime input improved.
[PR395](https://github.com/molpopgen/fwdpy11/pull/395)
* Added PendingDeprecationWarning to fwdpy11.evolve_genomes.
[PR396](https://github.com/molpopgen/fwdpy11/pull/396)

This is a major new feature release. Finally, we have support for demographic events involving multiple demes, meaning that merges, splits, size changes, migration, etc. are all possible now.

This release incorporates changes from several PRs. See the change log in the [manual](https://fwdpy11.readthedocs.io) for details.

Future release candidates will focus on improving the docs and fixing any bugs that get found along the way.

Fixed 362 and 363

Note: these changes could result in different results being obtained when mutating tree sequences **after** restoring a population from a file. Previous results were correct in distribution, however. The bug fixes mean that an internal lookup table is now properly restored after converting a population back from binary. This restoration means that the output from `fwdpy11.infinite_sites` could now differ from earlier versions.