Namespace conflicts between RIO and RIO.Map about rio

[snoyberg]

The only reason it didn't recommend a qualified name is that I having written that documentation yet, but it's definitely intended to be qualified. PR welcome if you'd like to add the missing comment.

…

On Sat, Jan 27, 2018, 12:53 AM tejon ***@***.***> wrote: RIO exports Data.Foldable.null. RIO.Map (implicitly) exports Data.Map.Strict.null. Being not technically the same function (despite the former being defined as the latter), they conflict. There are quite a few other examples of this (e.g. most of Foldable) -- null is just the first one I ran into. This comes down to base Data.Map being intended for qualified import, so they didn't use unique names; whereas Rio doesn't suggest a default qualifier, so I assume it's intended for use without one. On the upside, the fix should be simply hiding the conflicts in RIO.Map's import. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#35>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AADBB81Jt9ML5jMYi0GMvrBsiPjAEMIpks5tOldvgaJpZM4Ru_K_> .

[akhra]

D'oh! What I get for assuming. I can take care of that. Does this apply to everything that would be qualified under standard Prelude? At the moment it looks like only RIO.ByteString has a canonical qualifier.

[snoyberg]

PRs are definitely welcome. However, for the moment, there's a bit of a debate about the correct naming, see #12. I'd say that, for now, any kind of recommended naming is fine, and when #12 is resolved we can update the recommendations.

[snoyberg]

Addressed by #36, closing.