mirror of
https://github.com/QwenLM/qwen-code.git
synced 2026-07-10 01:29:17 +00:00
a76183f9c6
Some checks failed
Qwen Code CI / Classify PR (push) Waiting to run
Qwen Code CI / Lint (push) Blocked by required conditions
Qwen Code CI / Test (macos-latest, Node 22.x) (push) Blocked by required conditions
Qwen Code CI / Test (ubuntu-latest, Node 22.x) (push) Blocked by required conditions
Qwen Code CI / Test (windows-latest, Node 22.x) (push) Blocked by required conditions
Qwen Code CI / Post Coverage Comment (push) Blocked by required conditions
Qwen Code CI / CodeQL (push) Blocked by required conditions
E2E Tests / E2E Test (Linux) - sandbox:docker (push) Waiting to run
E2E Tests / E2E Test (Linux) - sandbox:none (push) Waiting to run
E2E Tests / E2E Test - macOS (push) Waiting to run
Sync cua-driver to Aliyun OSS / Mirror cua-driver binaries to Aliyun OSS (push) Has been cancelled
* fix(core): coerce numeric string params in SchemaValidator for MCP tools
LLMs frequently emit numeric parameters as strings (e.g. `{"depth": "3"}`
instead of `{"depth": 3}`), which strict MCP servers like Playwright reject
with schema validation errors ("params/depth must be number"). The
SchemaValidator already coerces boolean strings and stringified JSON but
was missing numeric coercion.
Add `fixNumericValues()` that converts string values to integers/numbers
when the schema expects integer/number but does not accept string,
consistent with the existing `fixBooleanValues` pattern.
Co-Authored-By: Qwen-Coder (Qwen Code) <noreply@qwen.ai>
* fix(core): prevent silent truncation in numeric string coercion
- Skip coercion for decimal strings when schema only accepts integer
- Prefer parseFloat when number type is accepted (fixes anyOf[integer, number])
- Add tests for integer-only decimal rejection, anyOf coercion, and array elements
Co-Authored-By: Qwen-Coder <noreply@qwen.ai>
* fix(core): coerce whole-number decimal strings for integer schemas
Replace the syntactic `hasDecimal` guard with a semantic `num % 1 !== 0`
check so whole-number decimal strings ("3.0", "-10.0") coerce to integers
while genuine non-integers ("5.5") are still left for the LLM to
self-correct. Python-based models commonly serialize float(3) as "3.0".
Co-Authored-By: Qwen-Coder <noreply@qwenlm.ai>
---------
Co-authored-by: Qwen-Coder (Qwen Code) <noreply@qwen.ai>
Co-authored-by: Qwen-Coder <noreply@qwenlm.ai>
|
||
|---|---|---|
| .. | ||
| acp-bridge | ||
| channels | ||
| cli | ||
| core | ||
| desktop | ||
| sdk-java | ||
| sdk-python | ||
| sdk-typescript | ||
| vscode-ide-companion | ||
| web-shell | ||
| web-templates | ||
| webui | ||
| zed-extension | ||