[ggml, llama-cpp] Update to 0.24.0 and 0.4.1

[miyan (miyanyan)]

Updates ggml from 0.11.1#1 to 0.24.0 and llama-cpp from build 9030 to 0.4.1. These updates are paired because llama.cpp 0.4.1 uses ggml 0.24.0. The llama-cpp port switches to upstream's semantic versioning scheme and v release tags, see ggml-org/ggml#1579

For llama-cpp, adapt the system dependency targets, replace the obsolete curl download dependency with cpp-httplib's OpenSSL feature, and make server depend on tools. Install the newly separated Python conversion package alongside the conversion script and update the installed tool list. Skip Metal tuning with system ggml because it uses private ggml source headers.

The unified app is disabled. Embedded Web UI building and automatic prebuilt UI downloads are also disabled, so the packaged server currently has no embedded Web UI. This behavior change needs review before this draft is marked ready.

• Changes comply with the maintainer guide.
• SHA512s are updated for each updated download.
• The "supports" clause reflects platforms that may be fixed by this new version, or no changes were necessary.
• Any fixed CI baseline and CI feature baseline entries are removed from that file, or no entries needed to be changed.
• All patch files in the port are applied and succeed.
• The version database is fixed by rerunning ./vcpkg x-add-version --all and committing the result.
• Exactly one version is added in each modified versions file.

[miyan (miyanyan)]

try remove fix-dequant_funcs.diff, cause glslang version is already >= 15.3.0, ggml-org/llama.cpp#13419 (comment)

[Billy O'Neal (BillyONeal)]

This looks good to me. One licensing nitpick from GPT 5.6 Sol:

• The 0.4.1 update newly builds vendor-hash and links it into the installed mtmd tool. That retained implementation includes BSD-2-Clause xxHash plus public-domain and MIT-licensed hash code, while the manifest, installed copyright, and SPDX metadata remain MIT-only.

Would you consider miyanyan#9 ?

[Billy O'Neal (BillyONeal)]

Thanks!