Changelog for mockcat-1.0.0.0

Changelog for `mockcat`

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to the Haskell Package Versioning Policy.

DSL Reboot: Replaced |> with ~> as the primary parameter chain operator (representing the "mock arrow").
Terminology Shift: Standardized terminology to "called" instead of "applied" throughout the library and error messages.
Simplified creating/stubbing API: f <- mock $ ... is now the canonical way.
Expanded structural diffing support for nested records and lists.
Unified verification API: All verification is now handled via shouldBeCalled.
Strict by Default: makeMock and makePartialMock now default to strict return values (implicit monadic return is disabled). makeAutoLiftMock was introduced for the previous behavior.

Deep Structural Diff: Enhanced error messages with precise caret pointers for complex nested data structures.
STM-based concurrency for mock registration and call recording.
Infinite arity support for mock/stub building.

Backward compatibility with 0.x.x APIs (stubFn, createMock, applied, etc.).
makeMockWithOptions, makePartialMockWithOptions, and MockOptions (internalized to simplify API).

This release is a complete reboot. Previous code will break.

Operator Change: Replace |> with ~>.

-- Old
createStubFn $ "arg" |> "result"

-- New
stub $ "arg" ~> "result"

Mock Creation: Use mock / stub instead of createMock / createStubFn.

-- Old
f <- createMock $ "arg" |> "result"

-- New
f <- mock $ "arg" ~> "result"

Verification: Use shouldBeCalled (unified API).

-- Old
f `shouldApplyTo` "arg"

-- New
f `shouldBeCalled` "arg"

Template Haskell Generics: makeMock is now strict by default (requires explicit pure for IO actions).
- Use makeAutoLiftMock for old implicit behavior.
- Or stick to makeMock and add pure to your return values.

Removed the upper limit on variable arguments when creating stub functions. Previously, there was a restriction on the maximum number of arguments, but this limitation has been removed, allowing stub functions to accept an arbitrary number of arguments.

Aliases expectApplyTimes and expectNever (preferred names) for pre-run expectation declarations.

README (EN/JA) now recommends expectApplyTimes / expectNever over legacy applyTimesIs / neverApply.
Clarified that expectApplyTimes n is the canonical form; expectNever is sugar for expectApplyTimes 0.

Legacy names remain exported for backward compatibility (no deprecation pragma yet). They may receive a soft deprecation notice in a future minor release after community feedback.

Race causing lost/double count in concurrent stub applications (strict modifyTVar').

MonadUnliftIO instance for MockT (initial groundwork for later parallel support).