Skip to content

Use Claude Code, Codex, VSCode, Pi, and OpenCode (and 6 other harnesses) for free (1.3B+ free tokens) from your terminal, app, IDE, or phone, and now from the browser with native browser sessions (multi-harness + multi-model) like OpenClaw (voice supported + ToS friendly)

Python

Make provider retry tests deterministic (#1070)

Ali Khokharf2f5b08714c4
diff --git a/tests/providers/support.py b/tests/providers/support.py
index 5cdaf4c4..1f3266cc 100644
--- a/tests/providers/support.py
+++ b/tests/providers/support.py
@@ -31,6 +31,23 @@ class PassthroughProviderRateLimiter(ProviderRateLimiter):
return await fn(*args, **kwargs)
+class ImmediateRetryProviderRateLimiter(ProviderRateLimiter):
+ """Run the real retry policy without exercising wall-clock backoff."""
+
+ async def execute_with_retry(
+ self,
+ fn: Callable[..., Any],
+ *args: Any,
+ **kwargs: Any,
+ ) -> Any:
+ kwargs.update(base_delay=0.0, max_delay=0.0, jitter=0.0)
+ return await super().execute_with_retry(fn, *args, **kwargs)
+
+ def extend_reactive_block(self, seconds: float) -> None:
+ """Leave reactive timing to the limiter's dedicated unit tests."""
+ del seconds
+
+
def passthrough_rate_limiter() -> ProviderRateLimiter:
"""Return a fresh limiter test double for one provider instance."""
return PassthroughProviderRateLimiter()
@@ -51,8 +68,8 @@ def profiled_provider(
def retrying_rate_limiter() -> ProviderRateLimiter:
- """Return a fresh real limiter for provider retry-policy tests."""
- return ProviderRateLimiter(
+ """Return a limiter that exercises retry policy without elapsed time."""
+ return ImmediateRetryProviderRateLimiter(
rate_limit=1_000_000,
rate_window=1.0,
max_concurrency=1_000,
diff --git a/tests/providers/test_openai_compat_5xx_retry.py b/tests/providers/test_openai_compat_5xx_retry.py
index dfb62d57..af8fff00 100644
--- a/tests/providers/test_openai_compat_5xx_retry.py
+++ b/tests/providers/test_openai_compat_5xx_retry.py
@@ -68,7 +68,6 @@ async def test_nim_stream_retries_on_openai_5xx_then_streams(status_code):
"create",
new_callable=AsyncMock,
) as mock_create,
- patch("asyncio.sleep", new_callable=AsyncMock),
):
mock_create.side_effect = [_internal_5xx(status_code), mock_stream()]
events = [e async for e in provider.stream_response(req)]
@@ -113,7 +112,6 @@ async def test_nim_stream_retries_on_pre_stream_connection_error_then_streams():
"create",
new_callable=AsyncMock,
) as mock_create,
- patch("asyncio.sleep", new_callable=AsyncMock),
):
mock_create.side_effect = [_connection_error(), mock_stream()]
events = [e async for e in provider.stream_response(req)]
@@ -148,7 +146,6 @@ async def test_nim_stream_connection_error_exhausted_emits_cause_chain():
new_callable=AsyncMock,
side_effect=error,
) as mock_create,
- patch("asyncio.sleep", new_callable=AsyncMock),
patch("free_claude_code.providers.openai_chat.provider.trace_event") as trace,
pytest.raises(ExecutionFailure) as exc_info,
):
@@ -202,7 +199,6 @@ async def test_nim_stream_openai_5xx_exhausted_emits_user_message(
"create",
new_callable=AsyncMock,
) as mock_create,
- patch("asyncio.sleep", new_callable=AsyncMock),
):
mock_create.side_effect = _internal_5xx(status_code)
with pytest.raises(ExecutionFailure) as exc_info: