We are historically lazy in doing the effort of preparing and realising a package.
We should at least try to have an automatic versioning based on some github id/constant/counter related to the latest commit or on the date, in such a way that we avoid having tens of different versions all marked 0.2.0.
Maybe something like 0.2.0_200426_155300 is ok (but I'm pretty sure there is an official pythonic way of versioning pre-releases; just grab in the net)
This is also a prerequisite to properly release a new pip package: once we'll be happy we can tag the release and it should automatically be build and released with the proper version.
Five minutes later we'll find a bug and we should automatically increase the version when releasing the fix
We are historically lazy in doing the effort of preparing and realising a package.
We should at least try to have an automatic versioning based on some github id/constant/counter related to the latest commit or on the date, in such a way that we avoid having tens of different versions all marked 0.2.0.
Maybe something like 0.2.0_200426_155300 is ok (but I'm pretty sure there is an official pythonic way of versioning pre-releases; just grab in the net)
This is also a prerequisite to properly release a new pip package: once we'll be happy we can tag the release and it should automatically be build and released with the proper version.
Five minutes later we'll find a bug and we should automatically increase the version when releasing the fix