Misc. bug: call to llama_model_meta_val_str for array values metadata returns nothing

[solomatov]

Name and Version

❯ llama-cli --version
ggml_metal_device_init: tensor API disabled for pre-M5 and pre-A19 devices
ggml_metal_library_init: using embedded metal library
ggml_metal_library_init: loaded in 0.014 sec
ggml_metal_device_init: GPU name: Apple M4 Max
ggml_metal_device_init: GPU family: MTLGPUFamilyApple9 (1009)
ggml_metal_device_init: GPU family: MTLGPUFamilyCommon3 (3003)
ggml_metal_device_init: GPU family: MTLGPUFamilyMetal4 (5002)
ggml_metal_device_init: simdgroup reduction = true
ggml_metal_device_init: simdgroup matrix mul. = true
ggml_metal_device_init: has unified memory = true
ggml_metal_device_init: has bfloat = true
ggml_metal_device_init: has tensor = false
ggml_metal_device_init: use residency sets = true
ggml_metal_device_init: use shared buffers = true
ggml_metal_device_init: recommendedMaxWorkingSetSize = 115448.73 MB
version: 7150 (3d07caa)
built with Apple clang version 17.0.0 (clang-1700.4.4.1) for arm64-apple-darwin25.1.0

Operating systems

Mac

Which llama.cpp modules do you know to be affected?

No response

Command line

Problem description & steps to reproduce

I invoke llama_model_meta_val_str with a string value of 'general.tags' on loaded gpt-oss model. I get a null value as a result

First Bad Commit

No response

Relevant log output

[pwilkin]

Due to the following code:

    // get metadata as string
    for (int i = 0; i < gguf_get_n_kv(ctx); i++) {
        gguf_type type = gguf_get_kv_type(ctx, i);
        if (type == GGUF_TYPE_ARRAY) {
            continue;
        }
        const char * name = gguf_get_key(ctx, i);
        const std::string value = gguf_kv_to_str(ctx, i);
        gguf_kv.emplace(name, value);
    }

you are not going to get general.tags because the type of that metadata entry is ARRAY.

This would require a modification of the GGUF metadata handling code to load arrays and provide a separate function llama_model_meta_val_array.