Bump to next non-breaking version

[nickrobinson251]

Please may we release a version with the new eweights functionality added in #401

v0.30.0...master

We're pre-1.0, and this is backwards compatible with v0.30, so just incremented to 0.30.1
:)

[rofinn]

I'd still vote to make it a minor release. I like to always reserve patch releases for bug fixes. Appveyor failure is a testing issue on fweights (sqrt on negative number) that should be fixed at some point.

[nickrobinson251]

I'd still vote to make it a minor release. I like to always reserve patch releases for bug fixes.

I am strongly against that in the pre-1.0 world, because having only two digits to play with, we should only distinguish breaking v non-breaking changes (and this is non-breaking).

Because 0.2 is considered incompatible with 0.3 etc., it's annoying to label backward compatible versions as incompatible. You end up with either:

• PkgA listing compat for a bunch of 0.Y versions for StatsBase in the Project.toml: this is painful as a maintainer, it does not distinguish actually breaking from non-breaking versions, and it's annoying to check if/when something should be dropped from a long compat list in the future
• PkgA listing only one StatsBase 0.Y version, e.g. 0.1, and PkgB listing only e.g. 0.2, but if 0.1 and 0.2 are compatible then we're going to get unnecessary "unsatsfiable requirement" errors when using PkgA and PkgB

[ararslan]

This should still be a minor release. If someone wants to start depending on eweights in their code, they'll need to adjust their version of StatsBase anyway.

[ararslan]

Closing in favor of #502.