Git commit

c262bed

Operating systems

Windows

GGML backends

CPU

Problem description & steps to reproduce

• build with msvc22 and std cpp < 202000L
• see compilation failure:

llama-chat.cpp: 
error C2664: 'bool llm_chat_detect_template::<lambda_1>::operator ()(const char *) const': 
cannot convert argument 1 from 'const char8_t [9]' to 'const char *'

because of this line

First Bad Commit

No response

Compile command

cl.exe llama-chat.cpp /std-c++17 ...

Relevant log output

llama-chat.cpp(150): 
error C2664: 'bool llm_chat_detect_template::<lambda_1>::operator ()(const char *) const': 
cannot convert argument 1 from 'const char8_t [9]' to 'const char *'

Possible solution

#define LU8(x) reinterpret_cast<const char*>(u8##x)

seems to please the compiler