Repository navigation
Conversation
|
Thanks for opening your first pull request in the Codna CLI! Please make sure the checks are green and that the change follows CONTRIBUTING.md. A maintainer will review it shortly. |
![codna-ai[bot]](https://avatars.githubusercontent.com/in/4061960?s=60&v=4){kind=small}
| if not target: | ||
| raise ConfigError( | ||
| f"model.key is not applicable to the keyless provider {provider!r} " | ||
| "(ollama/lmstudio/openai-compatible/litellm accept requests without a credential)" | ||
| ) |
There was a problem hiding this comment.
🟡 MEDIUM · correctness — Key rejected for openai-compatible/litellm even when provider requires a credent
The new block at lines 157-161 raises ConfigError when model.key is set for any KEYLESS_PROVIDERS member (ollama, lmstudio, openai-compatible, litellm). However, openai-compatible and litellm deployments commonly require an API key — a self-hosted LiteLLM proxy behind authentication, or an openai-compatible endpoint with token auth, is a standard setup. The blanket rejection prevents those valid configurations: users will get a hard error ("model.key is not applicable to the keyless provider 'litellm'") even when their deployment requires a credential. The design intent ("keyless by default") should use a warning or be limited to ollama/lmstudio which never use keys.
Reply @codna fix and Codna opens a verified, test-green, risk-gated fix PR.
| DURABLE_RANGES: Final[dict[str, str]] = { | ||
| "temporalio": ">=1.34,<2", | ||
| "langgraph": ">=1,<2", | ||
| "langchain-core": ">=1,<2", | ||
| } |
There was a problem hiding this comment.
⚪ LOW · correctness — DURABLE_RANGES missing langgraph-checkpoint entry
DURABLE_PINS declares four requirements including langgraph-checkpoint (minimum=(4,), maximum=(5,)), but DURABLE_RANGES only contains three entries — langgraph-checkpoint is absent. When require_durable() raises DurableUnavailableError for an out-of-range langgraph-checkpoint version it calls DURABLE_RANGES.get(requirement.distribution, ""), which returns "" for langgraph-checkpoint. The error details.range field is therefore empty, making the error message less actionable: the user is told the version is unsupported but not what range is accepted.
| DURABLE_RANGES: Final[dict[str, str]] = { | |
| "temporalio": ">=1.34,<2", | |
| "langgraph": ">=1,<2", | |
| "langchain-core": ">=1,<2", | |
| } | |
| DURABLE_RANGES: Final[dict[str, str]] = { | |
| "temporalio": ">=1.34,<2", | |
| "langgraph": ">=1,<2", | |
| "langchain-core": ">=1,<2", | |
| "langgraph-checkpoint": ">=4,<5", | |
| } |
Reply @codna fix and Codna opens a verified, test-green, risk-gated fix PR.
Automated mirror of cli/ (flattened), the root allowlist (server.json, glama.json, LICENSE, .gitleaks.toml) and the public-root/ community tree from thyn-ai/codna at 912b399d9a843964fd7681662ae563dff5e2769d. Merges itself via auto-merge once the mirror's required checks (codna review, Analyze (python)) pass and the codna-ai review approves; if a later sync lands first, this PR is closed as superseded -- every sync PR is a full snapshot of the exported surface.