Skip to content

Codineer v0.6.7

Choose a tag to compare

@andeya andeya released this 06 Apr 09:30
· 253 commits to main since this release

✨ New Features

  • Multimodal Input (Image Attachments):

  • Attach images via @image.png, paste from clipboard (Ctrl+V / /image), or drag-and-drop into the terminal.

  • Supports Anthropic, OpenAI, and other multimodal-capable providers.

  • Added support for Google Gemini with a free API key via the OpenAI-compatible endpoint.

  • Customizable Model Aliases:

  • Define your own short model names in settings.json using the modelAliases field.

  • Removed built-in model aliases, giving users full control over model naming.

  • codineer models now displays your configured aliases.

🛠️ Improvements

  • Enhanced Error Reporting: API error messages now include the request URL for easier debugging.
  • Refined @-mention Processing:
  • Support for specifying line numbers or ranges in file attachments (e.g., @src/main.rs:10-50).
  • Binary files are now attached with metadata (size, type) instead of attempting to read their content.
  • General Code Refinements: Various internal code improvements and dependency updates for stability and future enhancements.

🐞 Bug Fixes

  • Addressed an issue where some model aliases were not correctly resolved internally.

✨ 新功能

  • 多模态输入(图片附件):

  • 通过 @image.png 附加图片、从剪贴板粘贴(Ctrl+V / /image)或将文件拖拽到终端。

  • 支持 Anthropic、OpenAI 及其他兼容多模态的 Provider。

  • 新增支持 Google Gemini,可通过 OpenAI 兼容端点使用免费 API Key。

  • 可自定义模型别名:

  • 现在可以在 settings.json 中使用 modelAliases 字段定义自己的模型短名。

  • 移除了内置模型别名,赋予用户完全的模型命名控制权。

  • codineer models 命令现在会显示你配置的别名。

🛠️ 改进

  • 增强错误报告: API 错误消息现在包含请求 URL,以便更轻松地进行调试。
  • 优化 @ 提及处理:
  • 文件附件现在支持指定行号或行范围(例如 @src/main.rs:10-50)。
  • 二进制文件现在会附加元数据(大小、类型),而不再尝试读取其内容。
  • 通用代码优化: 各种内部代码改进和依赖更新,以提高稳定性和为未来的增强做准备。

🐞 Bug 修复

  • 修复了某些模型别名在内部未能正确解析的问题。

Full Changelog: v0.6.6...v0.6.7

Full Changelog: v0.6.6...v0.6.7