Hi, thanks for maintaining this package!
Currently, ollama is listed as a hard dependency. However, many users (including myself) already run a standalone Ollama installation or use a remote Ollama server. Having it as a required dependency forces a redundant installation and can cause port conflicts on 11434. Would you consider moving ollama from depends to optdepends? This would allow the package to remain lightweight for those who already have their backend configured while still guiding new users on what they need for a local setup.
Something like: optdepends=('ollama: for local LLM inference support')
Thanks for your work on this!
Pinned Comments
zxp19821005 commented on 2025-01-15 04:33 (UTC)
The upstream has utilized the third download links. Despite the file being updated, the filename remains unchanged. Therefore, if there are errors during package installation, please make the package outdated.