Load the in-repo fixtures once and feed the decoded cases straight into
the parameterized test, instead of collecting IDs and re-reading every
file to find the matching case per ID. PredicateConformanceFixtureCase
gains Identifiable + Sendable (for the arguments) and a
CustomTestStringConvertible extension so each case still displays by id.

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

Asserting an exact warning count coupled the fixtures to engine
internals. Match warnings by substring only (count-agnostic); an empty
`contains` now asserts that no warning is emitted, preserving the
"does not warn" cases. The three count-only fixtures gain the
"missing variable" substring they actually emit.

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

Drop the khepri single-file plumbing (defaultFixtureURL, fixtureURL,
the no-arg loadCases(), the env var, and the unused fixtureNotFound
error). This PR only runs the in-repo PredicateFixtures/ directory; the
khepri conformance loader lands with its test in the conformance PR.

Co-authored-by: Cursor <cursoragent@cursor.com>

Implements the json-logic-js some/all iteration predicates. Operator coverage
is expressed as JSON predicate fixtures (some.json / all.json) run by the
shared fixture runner introduced downstack, replacing hand-written unit tests.

Co-authored-by: Cursor <cursoragent@cursor.com>
(cherry picked from commit d11529c)

Implements the json-logic-js min/max operators (variadic flat list, ±∞ for
empty input, NaN propagation for non-numeric operands). Coverage is expressed
as JSON predicate fixtures (min.json / max.json) run by the shared fixture
runner, replacing hand-written unit tests.

Co-authored-by: Cursor <cursoragent@cursor.com>
(cherry picked from commit e73eabe)

Completes the json-logic-js iteration family (none/map/filter/reduce) on top
of some/all. Coverage is expressed as JSON predicate fixtures
(none.json / map.json / filter.json / reduce.json) run by the shared fixture
runner, replacing hand-written unit tests.

Co-authored-by: Cursor <cursoragent@cursor.com>
(cherry picked from commit 46d6925)

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

# Conflicts:
#	RevenueCat.xcodeproj/project.pbxproj

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

# Conflicts:
#	RevenueCat.xcodeproj/project.pbxproj

Co-authored-by: Cursor <cursoragent@cursor.com>

# Conflicts:
#	RevenueCat.xcodeproj/project.pbxproj

Co-authored-by: Cursor <cursoragent@cursor.com>

# Conflicts:
#	RevenueCat.xcodeproj/project.pbxproj

Expresses the "empty-string key resolves to the whole scope and is not
missing" case as a `{"!!":{"missing":[""]}}` -> false fixture; the string
coercion the other missing fixtures use can't distinguish [] from [""].
Bumps the pinned fixture count to 237. Keeps the corpus in sync with the
purchases-android counterpart.

Co-authored-by: Cursor <cursoragent@cursor.com>

Adds fixtures for the empty-segment dot-path splits, var default-vs-null-leaf
behavior, and the non-numeric missing_some threshold (7 new cases), bumping
the pinned count to 244. Also adds the var-null-path Swift test so the
non-fixturizable cases match the purchases-android counterpart. Keeps the
fixture corpus byte-identical across both repos.

Co-authored-by: Cursor <cursoragent@cursor.com>

…-logic-iteration-operators

Co-authored-by: Cursor <cursoragent@cursor.com>

# Conflicts:
#	Tests/RulesEngineInternalTests/PredicateFixtureTests.swift

…json-logic-min-max-operators

Co-authored-by: Cursor <cursoragent@cursor.com>

# Conflicts:
#	Tests/RulesEngineInternalTests/PredicateFixtureTests.swift

…on-logic-iteration-mapping-operators

Co-authored-by: Cursor <cursoragent@cursor.com>

# Conflicts:
#	Tests/RulesEngineInternalTests/PredicateFixtureTests.swift

Exposes the JSON Logic engine's evaluate entry point (and Value / RuleError) at the RulesEngineInternal module level for SDK call sites. No SDK wiring yet. SDK-4308.

Co-authored-by: Cursor <cursoragent@cursor.com>

Nest the error type under the RulesEngine namespace and update all module and test references.

Co-authored-by: Cursor <cursoragent@cursor.com>

Stops misreporting unexpected errors as parse failures, and trims evaluate's doc comment.

Co-authored-by: Cursor <cursoragent@cursor.com>

Match the file name to the RulesEngine namespace it declares.

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

fromJSONObject already throws EvaluationError, so the rethrow catch is unnecessary.

Co-authored-by: Cursor <cursoragent@cursor.com>

Co-authored-by: Cursor <cursoragent@cursor.com>

Brings back the lock-guarded `LoggerStorage` so the default logger can be
replaced (host-SDK wiring will expose this in a follow-up PR), while keeping
everything internal to the module.

Co-authored-by: Cursor <cursoragent@cursor.com>

…8-ios-rules-engine-api-surface

Lets SDK call sites install their own logger via the module-internal public surface.

Co-authored-by: Cursor <cursoragent@cursor.com>

Addresses review feedback on #6966 suggesting Swift Testing for the new
test file, matching the style of PredicateFixtureTests.

Co-authored-by: Cursor <cursoragent@cursor.com>

…-rules-engine-api-surface

Co-authored-by: Cursor <cursoragent@cursor.com>

# Conflicts:
#	RulesEngineInternal/Logger.swift
#	RulesEngineInternal/RulesEngine.swift