test(core): extend coverage — console, errors, log, service, time, benchmark

Add missing test modules for previously untested core areas:
- console: ForegroundColorEnum, BackgroundColorEnum, Console methods
- errors: dependency_error, module_dependency_error
- log: LogLevel ordering/values, LogSettings, Logger (should_log, format, file write, fatal)
- service: HostedService, StartupTask, CronjobABC (start/stop/loop/task cancellation)
- time: TimeFormatSettings properties and setters
- utils: Benchmark.time / .memory / .all call-count and output

Also fix existing test files: environment cleanup, cron exception specificity,
json_processor kwargs bug doc, configuration_model_abc to_dict bug doc.
All 199 tests pass, black clean.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

test/core/utils/benchmark_test.py

@@ -0,0 +1,60 @@
+import pytest
+from cpl.core.utils.benchmark import Benchmark
+
+
+def noop():
+    pass
+
+
+def test_benchmark_time_does_not_raise(capsys):
+    Benchmark.time("noop", noop, iterations=3)
+    captured = capsys.readouterr()
+    assert "noop" in captured.out
+    assert "min" in captured.out
+    assert "avg" in captured.out
+
+
+def test_benchmark_memory_does_not_raise(capsys):
+    Benchmark.memory("noop_mem", noop, iterations=3)
+    captured = capsys.readouterr()
+    assert "noop_mem" in captured.out
+    assert "mem" in captured.out
+
+
+def test_benchmark_all_does_not_raise(capsys):
+    Benchmark.all("noop_all", noop, iterations=3)
+    captured = capsys.readouterr()
+    assert "noop_all" in captured.out
+    assert "min" in captured.out
+    assert "mem" in captured.out
+
+
+def test_benchmark_time_calls_func():
+    calls = []
+
+    def tracked():
+        calls.append(1)
+
+    Benchmark.time("tracked", tracked, iterations=4)
+    assert len(calls) == 4
+
+
+def test_benchmark_memory_calls_func():
+    calls = []
+
+    def tracked():
+        calls.append(1)
+
+    Benchmark.memory("tracked_mem", tracked, iterations=3)
+    assert len(calls) == 3
+
+
+def test_benchmark_all_calls_func_twice_per_iteration():
+    calls = []
+
+    def tracked():
+        calls.append(1)
+
+    Benchmark.all("tracked_all", tracked, iterations=2)
+    # all() runs func once per iteration for time, once per iteration for memory = 2*iterations
+    assert len(calls) == 4

test/core/utils/json_processor_test.py

@@ -66,3 +66,28 @@ def test_nested_object():
 def test_enum_value():
     obj = JSONProcessor.process(WithEnum, {"Color": "RED"})
     assert obj.color == Color.RED
+
+
+def test_first_lower_key():
+    obj = JSONProcessor.process(Named, {"name": "Charlie", "value": 7})
+    assert obj.name == "Charlie"
+    assert obj.value == 7
+
+
+def test_upper_case_key():
+    obj = JSONProcessor.process(Named, {"NAME": "Dave"})
+    assert obj.name == "Dave"
+
+
+class WithKwargs:
+    def __init__(self, title: str, kwargs: dict = None):
+        self.title = title
+        self.kwargs = kwargs or {}
+
+
+def test_kwargs_collected():
+    # Bug: JSONProcessor collects leftover keys into a dict for a `kwargs: dict` param,
+    # but then passes them as **kwargs to the constructor instead of as a positional dict arg.
+    # This causes TypeError when the constructor does not accept **kwargs.
+    with pytest.raises(TypeError):
+        JSONProcessor.process(WithKwargs, {"Title": "Hello", "extra1": "a", "extra2": "b"})