Relax lint failures and fix async tests#7
Merged
santoshkumarradha merged 3 commits intomainfrom Nov 11, 2025
Merged
Conversation
AbirAbbas
added a commit
that referenced
this pull request
Apr 10, 2026
Adds 9 additional regression tests covering scenarios closer to the
production failure shape and edge cases that could surface more bugs.
Concurrency / realistic workloads (test_agent_ai_deadlock_recovery.py):
- test_parallel_hangs_some_succeed_some_recover_via_pool_reset:
The exact production scenario from extract_all_entities. Spawns 10
concurrent ai() calls via asyncio.gather; calls #3 and #7 hang on a
"stale socket". Asserts the 8 healthy ones return successfully (a
hung call's safety-net firing must not corrupt unrelated in-flight
requests), the 2 hung ones surface as TimeoutError, the pool is
reset, AND a follow-up batch of 5 calls all succeed (recovery is
durable, not one-shot).
- test_cascading_sequential_hangs_each_recover_independently:
Three calls in a row each hang then time out, then a fourth call
succeeds. Catches a future regression where reset accidentally
caches state (e.g. someone adds a `_already_reset` flag). Each
timeout must trigger a fresh reset.
- test_fallback_model_used_after_primary_hangs:
Primary model hangs, fallback model is configured, fallback must be
invoked successfully and the pool reset must run between them.
Exercises the interplay between `_make_litellm_call` and
`_execute_with_fallbacks`.
- test_tool_calling_loop_recovers_from_hang:
The tool-calling loop has its own copy of the timeout + reset logic
in `_tool_loop_completion._make_call`. This test routes through
`execute_tool_call_loop` (mocked to just call make_completion once)
and verifies the tool-loop path also recovers from hangs. Catches
the regression where someone fixes one path and forgets the other.
Reset robustness:
- test_concurrent_resets_are_safe:
20 concurrent resets via asyncio.gather. None should raise; all
should observe the final cleared state. Pins down the property so
a future "optimization" cannot break it.
- test_reset_swallows_exceptions_from_broken_cache:
If `in_memory_llm_clients_cache.clear()` raises (e.g. third-party
plugin replaced the cache with a broken object), the reset must
NOT propagate — otherwise it would mask the original TimeoutError
the caller is trying to surface.
- test_reset_does_not_clobber_unrelated_module_attrs:
Catches a regression where someone changes the implementation to
iterate over `dir(litellm_module)` and accidentally wipes config
flags, callbacks, or api_key. Asserts suppress_debug_info,
set_verbose, api_key, and success_callback all survive.
/debug/tasks robustness (test_agent_server.py):
- test_debug_tasks_endpoint_survives_cancelled_and_done_tasks:
The endpoint must remain responsive even when the live task set
contains tasks in pathological states (cancelled, done with
exception). JSON must still be well-formed.
- test_debug_tasks_endpoint_reports_done_and_cancelled_state:
Pins down the schema: each task entry must include `done=` and
`cancelled=` markers so operators can quickly distinguish "stuck on
await" from "finished but not yet collected" when diagnosing a hang.
Total: 63 passing tests across test_agent_ai.py + test_agent_server.py +
test_agent_ai_deadlock_recovery.py.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
5 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Testing