Process completed with exit code 1.

In the span streaming path (lines 191-208, 234-235), a StreamedSpan is created via `sentry_sdk.traces.start_span()` but never started (via `.start()` or `with` context manager) before calling `.end()`. The `end()` method calls `__exit__()` which expects `_context_manager_state` to be set by `__enter__()`. Since the span was never entered, `__exit__` fails silently (caught by `capture_internal_exceptions()`), and `_end()` is never called, causing the span to be lost. The same issue affects `on_validate` (lines 242-245, 259-260) and `on_parse` (lines 267-270, 282-283).

In the span streaming path (lines 191-208, 234-235), a StreamedSpan is created via `sentry_sdk.traces.start_span()` but never started (via `.start()` or `with` context manager) before calling `.end()`. The `end()` method calls `__exit__()` which expects `_context_manager_state` to be set by `__enter__()`. Since the span was never entered, `__exit__` fails silently (caught by `capture_internal_exceptions()`), and `_end()` is never called, causing the span to be lost. The same issue affects `on_validate` (lines 242-245, 259-260) and `on_parse` (lines 267-270, 282-283).

In the span streaming path (lines 191-208, 234-235), a StreamedSpan is created via `sentry_sdk.traces.start_span()` but never started (via `.start()` or `with` context manager) before calling `.end()`. The `end()` method calls `__exit__()` which expects `_context_manager_state` to be set by `__enter__()`. Since the span was never entered, `__exit__` fails silently (caught by `capture_internal_exceptions()`), and `_end()` is never called, causing the span to be lost. The same issue affects `on_validate` (lines 242-245, 259-260) and `on_parse` (lines 267-270, 282-283).

In `_sentry_execute_command`, if `await old_execute_command()` raises an exception, `db_span.__exit__()` and `cache_span.__exit__()` are never called because there's no try/finally block. This differs from the sync version in `_sync_common.py` which properly wraps the command execution in try/finally (lines 141-150). When Redis operations fail (connection errors, timeouts, etc.), spans will remain unclosed, potentially causing resource leaks or incomplete trace data.

In `_sentry_execute_command`, if `await old_execute_command()` raises an exception, `db_span.__exit__()` and `cache_span.__exit__()` are never called because there's no try/finally block. This differs from the sync version in `_sync_common.py` which properly wraps the command execution in try/finally (lines 141-150). When Redis operations fail (connection errors, timeouts, etc.), spans will remain unclosed, potentially causing resource leaks or incomplete trace data.