Name and Version
$ ./build/bin/llama-server --version
ggml_cuda_init: GGML_CUDA_FORCE_MMQ: no
ggml_cuda_init: GGML_CUDA_FORCE_CUBLAS: no
ggml_cuda_init: found 1 CUDA devices:
Device 0: NVIDIA GeForce RTX 5070 Ti, compute capability 12.0, VMM: yes
version: 6533 (51f5a45)
built with cc (GCC) 15.2.1 20250808 (Red Hat 15.2.1-1) for x86_64-redhat-linux
Operating systems
Mac
Which llama.cpp modules do you know to be affected?
llama-server
Command line
./build/bin/llama-server -hf ggml-org/gpt-oss-120b-GGUF -c 0 --jinja -ncmoe 31 --host 0.0.0.0 --port 59000
Problem description & steps to reproduce
The server OS doesn't matter, this is purely a WebUI issue.
Access the WebUI from macOS Safari, type in the "Ask anything..." box in Japanese. It will submit halfway after pressing the "word confirm" Enter key despite I'm still not finished writing yet.
(You may need to use the keyCode property which becomes 229 while on IME, as regular isComposing doesn't work correctly on Safari. Or just have an option to disable the "Enter to Submit" shortcut.)
First Bad Commit
b6500
Relevant log output