Hi, very well written proposal, the motivation section is exemplary. In particular excited about the ability to locally open a module (something that I began to like when writing Ocaml).
Would it be equivalent to say that all traditional import declarations create local modules?
When reading the proposal, that was somehow automatically my mental model. So I hope it is true, and that we can explain things this way. This would actually rule out the alternative
Drop point (8) allowing extension of existing local modules. because `import qualified Foo as A; import qualified Bar as A` only works if local modules can be extended.
Should -XLocalModules be required to import a local module?
My gut feeling is yes. The knowledge that `Foo.Bar.baz` comes from an imported `Foo.Bar` should still be applicable, unless a language extension clearly tells you that something is different. So I am supportive. I could understand if someone argued “too big of a change for too little impact; makes Haskell more complicated” … but I won’t be that someone. Cheers, Joachim -- Joachim Breitner mail@joachim-breitner.de http://www.joachim-breitner.de/