six8
goals for 1.0?
It seems like maybe the primary thing missing is windows support.
I don't personally care about that, but I believe it should be not-too-hard to convert the bits of clom that use sh(1) to use the python stdlib, and AppVeyor is a free-for-opensource Windows CI service that would prevent it from regressing.
Stabilization of the API. Are all the common use cases handled? Example, we just added support for |. What about (cmd1 & cmd2) for AND(cmd1, cmd2) etc. I think the API as is works as advertised and that might be good enough for 1.0 as long as there's no immediate plans to introduce breaking changes into the API. I would expect that the API introduced in 1.0 works until 2.0.
I don't know (or care about really) windows support. Clom is based on posix standards for command line usage. Windows commands have different ways of passing in options, no concept of pipes, etc. I don't know if it's worth it.
Are you using clom in your projects these days? Are there pain points in the API that need addressed?
It seems relatively solid now. You might call it 1.0 beta.
I'm using it quite a bit for a lot of fabric automation for deployment/builds/docker. The API has been complete for my use cases. I'll push out 0.8 and validate it a bit and call it 1.0.
I unfortunately can't use python 3.x much so I'd like validation that it works well there.
Certainly all your tests pass under py3: https://travis-ci.org/six8/python-clom/builds/48931990
If there are issues under python3, it should be quite easy to regression test them, and any resulting version bump would be a bugfix bump I think. I think the tests already show that the API is sound under python3.