Proof-outcome-case
Case macro for the different kinds of proof-outcome structures.
This is an ACL2::fty sum-type case macro,
typically introduced by fty::defflexsum or fty::deftagsum. It
allows you to safely check the type of a proof-outcome structure, or to split
into cases based on its type.
Short Form
In its short form, proof-outcome-case allows you to safely check the type of
a proof-outcome structure. For example:
(proof-outcome-case x :assertion)
is essentially just a safer alternative to writing:
(equal (proof-outcome-kind x) :assertion)
Why is using proof-outcome-case safer? When we directly inspect the
kind with equal, there is no static checking being done to
ensure that, e.g., :assertion is a valid kind of proof-outcome structure. That means there is nothing to save you
if, later, you change the kind keyword for this type from :assertion to something else. It also means you get no help
if you just make a typo when writing the :assertion
symbol. Over the course of developing VL, we found that such
issues were very frequent sources of errors!
Long Form
In its longer form, proof-outcome-case allows you to split into cases based
on the kind of structure you are looking at. A typical example would be:
(proof-outcome-case x
:assertion ...
:fail ...
:error ...)
It is also possible to consolidate ``uninteresting'' cases using
:otherwise.
For convenience, the case macro automatically binds the fields of x for
you, as appropriate for each case. That is, in the :assertion case,
you can use fty::defprod-style foo.bar style accessors for x
without having to explicitly add a assertion b*
binder.
