This reverts commit 755bdaa.

This reverts commit 17df455.

This reverts commit a985634.

This reverts commit 35e78b5.

This reverts commit d4df52c.

This reverts commit 13efaf0.

Always using root environment for now.

This gives each type inference variable a notion of universe but doesn't
do anything with it. We can always get the "current universe" from
infer_ctxt. This relies on the property of type variables that they can
never interact with siblings.

This is sort of confusing "side step". All it does is to change the
representation of a skolemized region. but the source of that universe
index is not the inference context, which is what we eventually want,
but rather an internal counter in the region inference context.

We'll patch that up later. But doing this now ought to help with
confusing diffs later.

This simple change says that if you have a constraint like

``` exists<'a> { for<'b> { 'a: 'b } } ```

you can solve it by making 'a static. This is "good enough" for now,
though not super smart, since for example

``` exists<'a> { for<'b> { if ('a: 'b) { 'a: 'b } } } ```

will still force 'a to be 'static, though it should not. (We don't
intend to fix that in this PR series; current rustc kind of sidesteps
the need to consider such concerns at the moment.)

These are only used to provide a better hint for code that was phased
out long ago. We can just remove it. It's been long enough.

We'll need this in order to start tracking skolemizatoins here, and it's
easier to update all the field accesses now rather than later.

remove all failing tests so we can do a crater run