Add examples from OOPSLA-2022 artifact as tests

[dvdvgt]

Description

Closes #63 by porting the examples from the artifact and adding them as tests. I also had a look at the case studies listed there, but it seems there are already tests that closely resemble them.

Since some examples were using the timestamp function and this is of course inherently impossible to test, I instead used a dummy counter instead.

[dvdvgt]

Are these tests failing because of the usage of unsupported features of the respective backend, or rather are actually bugs?

[b-studios]

Thanks for testing -- these are all expected:

Rank-2 not supported in ML

• println[A]() in this test leads to this error

This is why this error shows up:

-[error] examples/pos/global_capabilities.effekt:74:35: Cannot typecheck call.
-There are multiple overloads, which all fail to check:
-
-Possible overload: effekt.println$4 of type String => Unit
-  Wrong number of type arguments 1
-
-Possible overload: effekt.println of type Int => Unit
-  Wrong number of type arguments 1
-
-Possible overload: effekt.println$2 of type Double => Unit
-  Wrong number of type arguments 1
-
-Possible overload: effekt.println$1 of type Unit => Unit
-  Wrong number of type arguments 1
-
-Possible overload: effekt.println$3 of type Boolean => Unit
-  Wrong number of type arguments 1
-
-        def myprintln[A](msg) = { println[A](msg); resume(()) }
-                                  ^^^^^^^^^^^^^^^
+Hello!
+Hello!
...

The ML backend needs to have monomorphic definitions of println. You could define it at String and then use the specific show function (on int for example).

Boxing (especially non-empty captures) not supported in lift-based implementations (Chez lift, ML, and LLVM)

• boxing in this test is not expected to work.