Help wanted

I welcome your help.

Midje itself is, in core places, fairly tricksy code because it works via extensively rewriting the facts you give it. Adding new features may be a big leap. Some of the minor features in the Issues list might be a good place to start. That's especially true if they depend on Clojure's use of the JVM (protocols, for example), since I'm JVM-ignorant.

One of my long-term goals for Midje is that most syntax errors by users will be met by something more useful than "Argh! Fail! Here's 3000 lines of stack trace." Error checks could be added piecemeal, and be a way to learn the code.

My days as an Emacs elisp hacker are far behind me. I'm not familiar with current conventions and helper libraries. I bet Midje-mode could be greatly improved, both in how it does what it does now, and by adding new features. My style is to work on tests in the source file, not in a separate test file, so it could probably be greatly improved for those with a different style. It does coexist with clojure-test mode, but awkwardly. (I used the same key bindings for similar functions, but clojure-test mode sometimes keeps hold of them when I'd expect it not to.) There are clojure-test mode features that could be ported over.

Both lein midje and cake midje are fairly limited. There are probably abilities that lein test and cake test have that could be ported over.

There's a double-loading problem in cake - see the Issues section of the Cake-midje page.