this one is a bit more controversial.
in hwaf I do store a fair bit of non-os.environ stuff in the ctx.env (dicts of projects trees, stack of operations modules applied on the environment, etc...)
passing the full environment as in that callstack:
I think this should be okay. It looks like this change will still use the extra kwds for resolving but not otherwise update the deconf dictionaries with them.
this one is a bit more controversial. in hwaf I do store a fair bit of non-os.environ stuff in the ctx.env (dicts of projects trees, stack of operations modules applied on the environment, etc...)
passing the full environment as in that callstack:
https://github.com/hwaf/worch/blob/master/orch/waffuncs.py#L57 https://github.com/hwaf/worch/blob/master/orch/pkgconf.py#L99
breaks hwaf.