routstr-core

mirror of https://github.com/Routstr/routstr-core.git synced 2026-07-30 15:26:14 +00:00

Author	SHA1	Message	Date
9qeklajc	a383db5ee8	test: drop superseded billing regression tests	2026-07-24 21:08:06 +02:00
9qeklajc	6fcad8bcb9	Merge main into fix/zero-cost-fallback-billing	2026-07-24 21:02:06 +02:00
thefux	701c351ce7	fix: eliminate zero-cost billing fallback that gave free inference When adjust_payment_for_tokens raised during a streaming finalize, the code hardcoded total_msats=0 — giving users free inference and leaking the reserved balance (never released). This was the most critical billing bug still live on main. Changes: - Add _safe_finalize_billing helper that logs at CRITICAL, releases the reserved_balance (atomic clamp-to-zero pattern matching auth.py), and returns a non-zero cost dict using the reserved max_cost as best-effort - Replace all 6 zero-cost fallback sites in base.py: - handle_streaming_chat_completion main finalize - handle_streaming_chat_completion finalize_db_only - handle_streaming_chat_completion finalize_without_usage - handle_streaming_messages_completion main finalize - handle_streaming_messages_completion finalize_db_only - handle_streaming_messages_completion finalize_without_usage - Narrow except clauses from catch-all Exception to specific types (SQLAlchemyError, UpstreamError, OSError) so transient DB hiccups no longer silently result in free service - Log at CRITICAL (was ERROR) so operators are alerted to money leaks - Add 9 functional tests in test_zero_cost_fallback_fix.py covering: non-zero cost return, CRITICAL logging, reserved balance release, child key release, release failure handling, narrow exception types All 5 existing RED tests in test_zero_cost_fallback.py now pass. Full suite: 975 passed, 13 skipped, 0 failures.	2026-07-21 05:58:33 +00:00
thefux	7481ebd8cd	Merge PR #622 : fix: retry critical Cashu storage writes	2026-07-20 18:10:24 +00:00
thefux	7be1db8fd9	Merge PR #621 : fix: propagate Cashu storage errors	2026-07-20 18:10:15 +00:00
thefux	4a4443f98f	Merge PR #620 : fix: guard fee payouts against crash double-payments	2026-07-20 18:10:06 +00:00
thefux	fecf5b27b8	Merge PR #619 : test: vulnerability-reproducing RED tests + coverage # Conflicts: # .gitignore # tests/unit/test_coverage_admin.py # tests/unit/test_coverage_base2.py # tests/unit/test_coverage_payment_helpers.py # tests/unit/test_wallet_money_paths.py	2026-07-20 18:09:52 +00:00
thefux	59bcc3cbbf	test: add 17 base.py coverage tests (41%→42%) Tests for previously untested methods: - _extract_upstream_error_message (5 tests): JSON error, text error, empty body - on_upstream_error_redirect (2 tests): 402, 429 status codes - _fold_cache_into_input_tokens (2 tests): no cache, preserves total - get_cached_models / get_cached_model_by_id (2 tests) - get_x_cashu_cost (2 tests): with/without usage data - get_balance / create_account / refresh_models_cache / fetch_models (4 tests) ruff: clean, mypy: clean, 884 pass, 10 RED, 14 skip	2026-07-17 16:27:07 +00:00
thefux	25f75643c2	test: add wallet money-path tests, strengthen billing RED tests - New test_wallet_money_paths.py (10 tests): is_mint_connection_error, classify_redemption_error, store_cashu_transaction success path, get_balance, periodic task structure verification - Fixed test_messages_streaming_no_silent_billing_failure: was (always pass), now properly asserts the silent pass pattern must NOT exist - ruff: all clean, mypy: all clean 10 RED failures (correct), 867 pass, 14 skip	2026-07-17 16:19:33 +00:00
thefux	c68c1936d3	chore: fix ruff lint issues (29 fixes, 0 remaining)	2026-07-17 16:10:54 +00:00
thefux	49d285571e	test: rewrite vulnerability tests as RED (assert correct behavior) Rewrite vulnerability-documenting tests to assert CORRECT behavior so they FAIL against current buggy main. These are TRUE TDD RED tests. RED tests (10 failures — correct, these document live bugs): - test_store_cashu_raises_on_db_failure (DB errors must propagate) - test_retry_wrapper_exists (store_cashu_transaction_with_retry must exist) - test_emergency_refund_no_try_except_pass (must not silently lose tokens) - test_fee_payout_has_crash_guard (must have lock before pay) - test_billing_error_must_not_hardcode_zero_cost (must not give free service) - test_billing_error_must_release_reserved_balance (stuck funds) - test_billing_error_catch_is_too_broad (narrow exception type) - test_billing_error_must_log_critical (CRITICAL not ERROR) New coverage tests (45 pass, zero regressions): - test_coverage_base.py (17 tests): preparers, builders, injectors - test_coverage_admin.py (11 tests): withdraw validation, slugs, auth - test_coverage_proxy.py (13 tests): JSON parsing, model extraction Coverage gains: - middleware.py: 38% → 90% - helpers.py: 52% → 60% - proxy.py: 47% → 51% - admin.py: 35% → 36% Test suite: 857 pass, 10 RED failures, 14 skipped (zero regressions)	2026-07-17 16:05:56 +00:00
Paperclip Deployment Engineer	eb108a4a5a	test: add vulnerability-reproducing and coverage-filling tests Adds 40 new tests across 5 test files that document critical bugs and fill coverage gaps in the routstr-core codebase: - test_emergency_refund_integrity.py (5 tests): Documents the try/except/pass vulnerability in emergency refund paths (base.py:3643-3653 and base.py:4607-4617) where DB store failures silently lose minted tokens. Verifies store_cashu_transaction catches all exceptions and send_token mints before DB persistence. - test_zero_cost_fallback.py (6 tests): Documents the hardcoded zero-cost fallback (base.py:1012-1030) where exceptions from adjust_payment_for_tokens() result in total_msats=0, giving users free service with permanently reserved balances. - test_db_and_payout_resilience.py (8 tests): Confirms store_cashu_transaction_with_retry was reverted (#600→#604). Documents the fee payout pay-then-reset crash window and wallet caching mechanism. - test_coverage_middleware.py (11 tests): Fills middleware.py coverage gap (was 38%) — tests LoggingMiddleware, _should_log filters, request_id_context, and middleware exports. - test_coverage_payment_helpers.py (10 tests): Fills payment/helpers.py coverage gap (was 52%) — tests check_token_balance, estimate_tokens, create_error_response, and image token calculation helpers. All tests pass against current main (830 passed, 13 skipped).	2026-07-17 15:12:48 +00:00

Compare commits

Diff Content Not Available