Also, rewrite `iIntoEmpValid`. Now, instead of using Ltac to traverse
the type of the term and generate goals for the premises, we repeatedly
apply a series of lemmas. This has the advantage that it works up to
convertability, and we no longer need the `eval ...` hacks.

This avoids weird Ltac behaviors like those in !272.

Also, change `before_tc` keyword into `as` to be consistent with other tactics.

Joseph Tassarotti authored May 24, 2019 and Robbert Krebbers committed May 24, 2019

Also fixes pre-existing bug in iCombine error messages.

It now supports the specialization pattern `(H spat1 .. spatn)`, which first
recursively specializes the hypothesis `H` using the specialization patterns
`spat1 .. spatn`.

Split it up into more logical parts.

Thanks to @Blaisorblade for reporting.

pm_reduce just reduces away proof mode terms using cbv; pm_prettify just prettifies user-visible connectors using cbn.
Most uses of pm_default are converted to default to keep the desired reduction behavior.

List literals reduce with a spurious `emp` at the end, which is not pretty.

With a pretty proof by Robbert

make pm_maybe_wand a BI connective; reduce BI connectives and option combinators in the proofmode with cbn

We already do the same for the goal, this avoids some scope
delimiters being displayed.