Change Log

NOTE: Since Clojars introduced a Verified Group Names policy, no new libraries could be pushed to the expectations group, and doo filters out JAR artifacts that begin clojure- for self-hosted ClojureScript environments (i.e., planck), so continuing to use clojure-test for the artifact name is not viable. Accordingly, the 2.x versions of this library are published as com.github.seancorfield/expectations.

Stable Releases

Only accretive/fixative changes will be made from now on.

• 2.0.160 -- 2022-03-26

  • Fix #28 by recognizing qualified calls to expect (to suppress legacy behavior in more cases).
  • Update build-clj to v0.8.0.

• 2.0.157 -- 2022-01-25

  • Fix a small regression in how classes are treated when used as predicates in expect.

• 2.0.156 -- 2022-01-19

  • Address #26 by adding an example combining more-> and more-of around destructuring ex-info data.
  • Fix #24 by using a local (gensym) for the actual value in more and more-> so it is only evaluated once.
  • Update build-clj to v0.6.7 and automate snapshot/release builds.

• 2.0.143 -- 2021-12-01

  • Fix #23 by adding support for set-in-set expectations.
  • Documentation updates.
  • Build deps updates.

• 2.0.137 -- 2021-11-07

  • Address #22 by adding clj-kondo.exports (this is just a first pass; the :lint-as mappings will probably be replaced by hooks in the future).
  • Fix #19 by supporting regex/patterns dynamically inside =? (as well as the compile-time support already in expect).
  • Update build-clj to v0.5.0.
  • Switch to build.clj/tools.build for CI/deploy.

2.0.x Prereleases

• 2.0.0-alpha2 -- 2021-06-09

  • Mostly a documentation refresh, compared to Alpha 1.

• 2.0.0-alpha1 -- 2021-06-05

  • Make (defexpect foo) and (defexpect foo (bar)) match the behavior of deftest, without wrapping the body in (expect ,,,). This is potentially breaking insofar as (defexpect foo (produces-falsey)) would have been a failing test in 1.x but now silently just runs (produces-falsey) in the same way that (deftest foo (produces-falsey)) does.
  • Bring in several test-running functions from clojure.test, for convenience in dev/test so users don't need to require clojure.test as well.
  • Implement cljs.test's version of use-fixtures: accepts functions or hash maps (containing :before and/or :after keys with 0-arity functions).
  • Add various macro-like constructs back into the source code to improve the documentation (in, from-each, more-of, more->, more are really only syntactic constructs inside expect).
  • Support (self-hosted) ClojureScript via planck -- see https://github.com/clojure-expectations/clojure-test/pull/16 for details (@kkinear).

Previous Releases

These versions required users to also require clojure.test and were not as directly comparable to clojure.test behaviors.

• 1.2.1 -- 2019-12-09

  • Fix cljdoc.org index (Collections was missing).

• 1.2.0 -- 2019-12-09

  • Improve failure reporting for in; allow it to be combined with more etc. #11
  • Add support for mocking return values in side-effects.
  • Add support for optional message argument in expect. #9
  • Added article-style documentation for cljdoc.org. #6, #7, #8, #10
  • Add example of more-> equivalent to thrown-with-msg?. #5

• 1.1.2 -- 2019-12-07

  • Adds between and between' for inclusive and exclusive range checking.
  • Fix in with a hash map to correctly detect failing cases.
  • Add a first round of tests (finally!). Verified support for Clojure 1.8 (without Spec expectations). Verified full support for Clojure 1.9 and 1.10.1.
  • Clean up :require .. :refer in README to list all public symbols. #4
  • Fixes links in README. PR #3 (@marekjeszka)
  • Add/improve docstrings. Add ^:no-doc metadata for cljdoc.org.

• 1.1.1 -- 2019-01-14

  • An expectation can now use a qualified keyword spec to test conformance of the actual value. Failures are reported with the spec explanation. #2
  • If Paul Stadig's Humane Test Output is available (on the classpath), failure reporting is automatically made compatible with it. Expectations that use data structure "equality" (the =? extension to is) will produce "humane" output for failures, showing differences. #1

• 1.1.0 -- 2019-01-08

  • (broken version)

• 1.0.1 -- 2019-01-02

  • Initial version