Features
- Add support for OpenCode - howto here: #253 (comment)
- Bump internal
llama.cppto b10068
Internals
- Thread
CancellationTokenthrough all the Paddler APIs, so everything can be canceled cleanly inpaddler_client
llama.cpp to b10068CancellationToken through all the Paddler APIs, so everything can be canceled cleanly in paddler_client