Speculoos library changelog

Breaking changes

• Now require version 2 of `fn-in` lib so that can handle subvectors. See https://github.com/blosavio/fn-in/issues/2.
• Now require version Clojure 1.12.0.

Non-breaking changes

• All functions in Speculoos can now handle subvector types.

• map-2 → map*
• map-indexed-2 → map-indexed*
• reduce-2 → reduce*
• reduce-indexed-2 → reduce-indexed*
• validate-fn-meta-spec → validate-fn

Breaking changes

• Adjusted `scalars-without-predicates` to return a set of all-paths elements, instead of a sequence, to be consistent with `collections-without-predicates`.
• Re-named functions in collection-functions namespace from `blah-2` to `blah` to emphasize they're not merely a new version, but that they operate on *any* Clojure collection type, and to be consistent with `fn-in` namespace members.
• Removed Speculoos project's internal 'fn-in' namespace to an external lib. 'fn-in' is a new dependency.
• Moved `recover-literal-path` and helper functions from utility namespace to core namespace; moved attendant tests from utility-tests to core-tests. In preparation for adding feature to `validate-collections`.
• GitHub Issue #1: Changed validation report key returned `validate-set-elements` from `:datum` to `:datums-set` to emphasize that *every* datum in the set is validated.
• Removed un-serious utility functions of dubious usefulness.
• Removed function arg-vs-ret scalar/collection validation feature from `validate-fn-with`. It was too complex, and difficult to use.
• Re-named `validate-fn-meta-spec` to `validate-fn`.

Non-breaking changes

• Added `thoroughly-valid-scalars?` and `thoroughly-valid-collections?` utilities.
• Adjusted `thoroughly-valid?` to internally use new functions.
• Removed hard-coded collection hierarchy derivation and replaced with a `load` from a single file.
• Expanded docstrings and examples to provide clarity, completeness, and better convey how functions behave and what they're intended for, especially for a user who looks at the API doc before html documentation.
• Internalized `reduce-indexed` into 'core' namespace, plus unit tests into core-tests. Removed 'collection-functions' namespace because Speculoos does not use `map*`, `map-indexed*`, nor `reduce*`.
• Removed Speculoos project's own version of the derived collection hierarchy; now refers to the one provided by the 'fn-in' dep.
• Upgraded `validate-collections` report so that entries now indicate the literal path to the collection datum, at keyword `:path-datum`.
• GitHub Issue #2: Upgraded `validate-scalars` to accept 'bare' scalar and 'bare' predicate. This is intentionally undocumented behavior, but provided for convenience and least surprise.
• GitHub Issue #1: Fixed incorrect `:path` reporting in `validate-set-elements`.
• GitHub Issue #3: Fixed bug in `validate-collections` that validated un-paired collection predicates against `nil`.
• Added `predicates-without-collections` to complete feature matrix.
• GitHub Issue #4: Fixed ignored return collection validations in `valdiate-fn-with`.
• Adjusted `validate-fn-with` to take advantage of `validate-scalars` upgraded ability to validate bare scalars with bare predicates.
• Added a completely new argument/return relationship validation to `validate-fn-with` that is much cleaner. Old, complex way remains.
• Added `:canonical` option to `exercise` and `exercise-fn` to use canonical values instead of randomly-generated values.
• GitHub Issue #9: Fixed `valid-collections?` to properly handle collection predicates that yield truthy/falsey values.

• collections-without-specs → collections-without-predicates
• data-with-specs → scalars-with-predicates
• data-without-specs → scalars-without-predicates
• valid-collection-spec? → valid-collections?
• valid-macro-spec? → valid-macro?
• valid-scalar-spec? → valid-scalars?
• validate-collection-spec → validate-collections
• validate-scalar-spec → validate-scalars

Breaking changes

• Re-named functions for clarity, correctness, and conciseness. Also updated documentation to correspond to new function names.
• Re-naming functions to more accurately convey purpose and action. `data-without-specs` to `scalars-without-predicates`, `data-with-specs` to `scalars-with-predicates`, and `collections-without-specs` to `collections-without-predicates`.

Non-breaking changes

• Upgraded `defpred` predicate definition macro that can inspect a function S-expression and create a random sample generator. Also added documentation and illustrations.
• Added `thoroughly-valid?` utility function.
• Improved `data-without-specs`, `specs-without-data`, `data-with-specs` to properly handle scalars within a set.
• Removed un-needed/redundant api.clj and api.html pages because they're automatically generated by Codox.
• Added a `*max-tries*` dynamic var that governs random sample generators.
• Added tests to verify handling of sequences emitted by `interpose`, `interleave`, `lazy-cat`, `mapcat`, and `zipmap`.
• Added `lazy-seq?` utility predicate.
• Various documentation edits spanning 2024July10 through 2024July26.

Breaking changes

Non-breaking changes