Include some integration testing against cl-quil

[notmgsk]

cl-quil-tests run pretty long. I wonder if executing a subset of the tests would suffice.

Another option (probably healthier) is to create an issue to make sure that magicl test coverage is at least as comprehensive as the code paths exercised by quilc. I noticed with the LAPACK-CSD function that it was not well tested in magicl.

Originally posted by @kilimanjaro in #121 (comment)