Ariel Ben-Yehuda

Inside each function: ``` fn f(...) { ... } ``` just replace the `T: Foo` bound by a `WF(T: Foo)` bound. So the caller (actually, the expander of the type-scheme)...

> If `|- A` and `Implied(A) |- T: Trait`, then `|- T: Trait`. This feels like it ignores the parametric context. While this should be enough for translation, rustc also...

@scalexm That's why I added the `WF(Γ)` as a parameter to the parametric definition. We have `WF(Γ) |- T: Copy`, so order is restored. Also, because impl-trees must be finite...

Some comments from discord that someone might want to document permanently: Expand > arielbyToday at 10:23 PM In the Chalk WellFormed/FromEnv proposal how do you get a FromEnv(ProjectionTy)? do you...

Occurred for `ceramic` again in http://cargobomb-reports.s3.amazonaws.com/pr-43880/b96d5eea76e758bdd82fd4169eb4d5f7e3ce4fe2/reg/ceramic-0.3.0/log.txt

How do we translate something like ` 'c ends` is legal, the ordering `'c ends -> 'b ends` is legal, but their combination `'a ends -> 'c ends -> 'b...

I think the best way to solve this is to A) treat every free lifetime to the combination of all CFG nodes in the current function + an "skolemized continuation...

I think "post-end skolemized variables" is a better name than "end points" - they aren't just points (if you "inlined" a function, they would turn into the subgraph of the...

I'm using "skolemized" for what rustc calls "skolemized" ~~actually closer to "Herbrandized"~~actually that depends on your POV - to mean free lifetimes that are supposed to be "closed over" by...

@nrc I'm quite sure you can do dot completion using the type-context you get after compilation - it already has the type information for all nodes.