Fully evaluate lazy seqs when macroexpanding

[yuhan0]

Fixes #886

Note: I also added a doall in what seemed like the right spot for the cljs expander, but wasn't able to easily test whether it had any effect. There's no existing test coverage for that code path so I'm fine with removing it in case it breaks anything.

---

Before submitting a PR make sure the following things have been done:

• The commits are consistent with our contribution guidelines
• You've added tests to cover your change(s)
• All tests are passing
• You've updated the README
• Middleware documentation is up to date
  • Please check out and modify the cider.nrepl ns which has all middleware documentation.
  • Run lein docs afterwards, and commit the results.

Thanks!

[vemv]

Thanks!

LGTM

[yuhan0]

Oh.. I only ran lein test, good thing the CI caught this bug. Somehow I couldn't replicate the test failure locally using make tools-deps-test but hopefully the following commit should fix it.

[alexander-yakushev]

Unfortunately, it is not enough to call doall to force all internal parts of the constructed sexp. See https://clojure-goes-fast.com/blog/clojures-deadly-sin/#no-way-to-force-everything.

Try making a test case where the lazy part of the macro is somewhere deeper in the top-level expression. I think you'll have to print-str or something like that.

[alexander-yakushev]

This PR is superseded by #894 which should handle the lazy expansion problem fully besides other things.

[alexander-yakushev]

@yuhan0 Thanks for reporting the issue and proposing a fix! This led to uncovering a deeper problem with macroexpansion middleware.