I have a question regarding the visibility of some more specific impl blocks (just impl, not impl Trait).
Suppose library A defines type Foo<T>. Then there's library B that defines type Bar and also an impl Foo<Bar> block that contains some Bar-specific member functions.
What are the visibility rules of these member functions? Does user code need to use the Bar type in order to have them in scope?
Good question. Any methods inside impl Foo<Bar> are inherent methods, so you don't need anything at all in scope to use them. You can have access to a Foo<Bar> value without having any of those in scope: Let's pretend that self.xyz.foo is a Foo<Bar> value. Then you can call the inherent methods like self.xyz.foo.methodname().
Really? Wow, cool. Thanks for answer 
I suppose given how UFCS works, this makes sense...
Ok, to address the different crates part of the question, the B crate part is not allowed. It can't add inherent methods to Foo. So for it to happen, all the inherent methods would have to be in the first crate.
Ah, ok, that's a bit of a bummer. Glad it's cleared up for me though...