defexpect with a single expression should behave like deftest

[seancorfield]

Currently, a defexpect whose body is one or two expressions with no embedded expect is treated as shorthand for an expect with those expressions. That's fine for two expressions:

(defexpect my-name (expected-expr) (actual-expr))
;; becomes
(defexpect my-name (expect (expected-expr) (actual-expr)))

But for a single expression, this means a difference of behavior compared to deftest:

(defexpect my-name (actual-expr))
;; becomes
(defexpect my-name (expect (actual-expr)))
;; which is treated as
(deftest my-name (is (actual-expr)))

But a single expression in the body of a deftest is just "odd" and behaves like a regular function:

(deftest my-name (actual-expr))

That just runs the code as-is, no assertions. I think the single-bodied defexpect above should behave the same way (perhaps with a warning on macroexpansion).

[seancorfield]

I made 0- and 1-form bodies throw exceptions as a test case for our codebase at work and ran into no issues.

(deftest foo) is legal and creates a test with an empty body -- (defexpect foo) is illegal (wrong number of arguments to expect) -- that's a bug.

I intend to make the following changes:

• Fix the bug above and make 0- and 1-form defexpect (without expect) behave exactly like deftest

Since this is potentially breaking -- (defexpect foo (produces-falsey)) would be a failing test right now -- I'll bump to 2.x after this change but I really don't expect anyone to be relying on that (and it's never been documented).