[djblue]

I'm looking for feedback one how we want to test split. Also, I have a feeling that edge cases will be dominated by implementation defined behavior, so do we want to capture that as well?

[Copilot]

🟡 Changes recommended

Two ClojureScript expectations incorrectly include leading empty strings and would fail.

1 open issue

• Correct ClojureScript split expectations for empty regex · New

What changed in this PR

Adds clojure.string/split coverage to the cross-dialect test suite.

Changes:

• Tests basic splitting, limit handling, and whitespace.
• Records dialect-specific empty-pattern behavior.
• Adds examples from ClojureDocs.

File	Description
test/​clojure/​string_test/​split.cljc	Adds guarded split tests with shared and dialect-specific expectations.

_{Review effort: Balanced}

[jeaye]

This is already great coverage. The most important thing is to test what's covered by the doc string.

Splits string on a regular expression. Optional argument limit is
the maximum number of parts. Not lazy. Returns vector of the parts.
Trailing empty strings are not returned - pass limit of -1 to return all.

I don't think we should worry too much about all of the implementation details. We just want to provide a good coverage of the essential success cases and essential failure cases.

[djblue]

I think we have enough coverage to satisfy the docstring.

[jeaye]

Great work!