hermes-agent/tests/agent/test_custom_pool_mismatch_guard.py
Teknium 6b81590c55
test: prune low-value tests suite-wide (wave 1) — 46,820 → 28,106 test functions
Systematic prune per AGENTS.md test policy, one pass over every major
test tree (gateway, hermes_cli, tools, agent, run_agent, plugins, cli,
cron, tui_gateway, honcho/openviking, root-level):

- DELETE: source-reading tests (read_text/getsource on prod files),
  change-detector tests (exact catalog counts, model-name snapshots,
  config version literals), mock-echo tests (assert a mock returns what
  it was told), assertion-free/trivial tests, near-duplicate
  parametrizations (boundaries + one representative kept), async/sync
  twin duplicates, cosmetic within-file variations.
- KEEP (mandatory): security/redaction/approval guards, message-role
  alternation invariants, prompt-caching/deterministic-call-id
  invariants, issue-number regression tests (deduped), E2E tests.
- 6 test files deleted outright (script-style/no-assert or fully
  redundant); conftest.py, fakes/, fixtures/ untouched.
- tests/acp/conftest.py added: autouse fixture stubs the live
  models.dev/GitHub/Copilot/Anthropic inventory fetches that ACP server
  tests performed on every session create — test_server.py 147s → 3.4s,
  and the tests are now genuinely hermetic.
- Sleep-based slowness shrunk where safe (codex_ttfb_watchdog,
  compression_concurrent_fork, etc.); no wall-clock assertion tightened.

Verification: full hermetic suite via scripts/run_tests.sh —
2439 files, 31,130 tests passed, 0 failed, 0 flaky retries, 315s wall
(baseline: 583s wall, 13,564s subprocess CPU).
2026-07-29 13:10:23 -07:00

71 lines
2.6 KiB
Python

"""Regression tests for the credential-pool provider-mismatch guard with
custom providers (Bernard's Fireworks report, June 2026).
Custom endpoints carry two naming conventions for the same provider: the
agent's ``provider`` attribute is the generic ``"custom"`` label while the
pool is keyed ``custom:<normalized-name>`` (``CUSTOM_POOL_PREFIX``). The
defensive guard in ``recover_with_credential_pool`` compared the two
literally, logged "Credential pool provider mismatch: pool=custom:<name>,
agent=custom", and skipped recovery — so 401/429 recovery (refresh,
rotation) never ran for ANY custom-provider user.
The fix accepts the pair only when the agent's current base_url resolves to
the same pool key, preserving the guard's original purpose (#33088/#33163:
never mutate the primary's pool while a fallback provider is active).
"""
from unittest.mock import MagicMock, patch
import pytest
from agent.agent_runtime_helpers import recover_with_credential_pool
from agent.error_classifier import FailoverReason
FIREWORKS_URL = "https://api.fireworks.ai/inference/v1"
def _agent(provider, base_url, pool_provider):
agent = MagicMock()
agent.provider = provider
agent.base_url = base_url
pool = MagicMock()
pool.provider = pool_provider
agent._credential_pool = pool
return agent, pool
class TestCustomPoolMismatchGuard:
def test_unrelated_custom_pool_still_guarded(self):
"""agent=custom pointed at a DIFFERENT endpoint than the pool's
custom provider must still skip pool mutation."""
agent, pool = _agent(
"custom", "https://other-endpoint.example/v1", "custom:fireworks"
)
with patch(
"agent.credential_pool.get_custom_provider_pool_key",
return_value="custom:other",
):
recovered, _ = recover_with_credential_pool(
agent,
status_code=401,
has_retried_429=False,
classified_reason=FailoverReason.auth,
)
assert recovered is False
assert not pool.method_calls
def test_fallback_provider_still_guarded(self):
"""Original #33088/#33163 contract: when a fallback provider is
active (agent.provider != pool.provider, non-custom), the pool is
never mutated."""
agent, pool = _agent("openai-codex", "https://chatgpt.com/backend-api", "custom:fireworks")
recovered, _ = recover_with_credential_pool(
agent,
status_code=401,
has_retried_429=False,
classified_reason=FailoverReason.auth,
)
assert recovered is False
assert not pool.method_calls