test: fix flaky test_crawler_intermediate_statistics on Windows (#2265)

`test_crawler_intermediate_statistics` failed on Windows 3.12
([run](https://github.com/apify/crawlee-python/actions/runs/36837667765/job/110288921245?pr=2263))
with `98582 µs >= 100000 µs`. The test sleeps 100 ms with
`asyncio.sleep` and expects `crawler_runtime` to be at least that long,
but `crawler_runtime` comes from `datetime.now()`. On Windows before
Python 3.13 that clock ticks every ~15.6 ms, and the only slack in the
test was the poll lag, which can be close to zero.

The test now takes its own `datetime.now()` readings when it sees the
crawler start and after the sleep, and asserts the runtime is at least
the gap between them. Both sides use the same clock, and the check still
fails if the runtime isn't updated while the crawler runs.

*✍️ Drafted by Claude Code*

Author: vdusek

tests/unit/crawlers/_basic/test_basic_crawler.py

@@ -12,7 +12,7 @@
 from collections import Counter
 from concurrent.futures import ProcessPoolExecutor
 from dataclasses import dataclass
-from datetime import timedelta
+from datetime import datetime, timedelta, timezone
 from itertools import product
 from typing import TYPE_CHECKING, Any, Literal, cast
 from unittest.mock import ANY, AsyncMock, Mock, call, patch
@@ -2446,11 +2446,15 @@ async def handler(_: BasicCrawlingContext) -> None:
     # has enough time to reach the active state on slow CI runners under xdist load.
     crawler_task = asyncio.create_task(crawler.run(['https://a.placeholder.com']))
     assert await poll_until_condition(lambda: crawler.statistics.active, timeout=30)
+    active_seen_at = datetime.now(timezone.utc)
 
-    # Wait some time and check that runtime is updated.
+    # Wait some time and check that runtime is updated. The runtime is measured on the wall clock, which on Windows
+    # ticks at ~15.6 ms independently of the `asyncio.sleep` clock, so bound it by a wall-clock interval inside the run.
     await asyncio.sleep(check_time.total_seconds())
+    min_runtime = datetime.now(timezone.utc) - active_seen_at
     crawler.statistics.calculate()
-    assert crawler.statistics.state.crawler_runtime >= check_time
+    assert min_runtime > timedelta(0)
+    assert crawler.statistics.state.crawler_runtime >= min_runtime
 
     # Wait for crawler to finish
     await crawler_task