These instances are the same as those for e.g. the later modality.

This fixes issue #187.

It now turns the goal into `P` and `<pers> Q`, which is dual to
`iDestruct`, which turns `P ∧ <pers> Q` into `P` and `□ Q`.

Extend ElimModal with Boolean flags to specify whether it operates on the persistent/spatial context.

Fixes #177

Split [FromAssumption] into three, depending on wheter the parameters are evars. This is to avoid loops in TC search.

Based on an earlier MR by @jung.

This change is slightly more invasive than expected: in monPred we were
using the embedding before the BI was defined. With the new setup, this
is no longer possible, because in order to make an instance of the
embedding, we need to know that `monPred` is a BI. As such, we define
`emp`, `⌜ _ ⌝` and friends directly in the model of `monPred` and later
prove that they are equal to a version in terms of the embedding.

This requires changing the Hint Mode of the [Frame] type class because it should not fail if its parameter is an evar, but instantiate it instead. In order to prevent all the other instances of [Frame] to intantiate this evar themselves, we create a new type class [KnwonFrame], which corresponds to the old behavior.