llama-cpp: bump version, enable RPC-server

dev-cpp/llama-cpp/llama_cpp-b4663.recipe

@@ -13,7 +13,7 @@ COPYRIGHT="Georgi Gerganov"
 LICENSE="MIT"
 REVISION="1"
 SOURCE_URI="https://github.com/ggerganov/llama.cpp/archive/refs/tags/$portVersion.tar.gz"
-CHECKSUM_SHA256="a1b82ab73aa7aa287089ea1752066e60f98adc32859cde40beab190ba0bd97de"
+CHECKSUM_SHA256="c630e767ce858e0ea16980bc38a1fa008c25a320610d07aa136b147446a1b3ef"
 SOURCE_DIR="llama.cpp-$portVersion"
 SOURCE_URI_2="https://github.com/microsoft/mimalloc/archive/refs/tags/v2.1.7.tar.gz"
 CHECKSUM_SHA256_2="0eed39319f139afde8515010ff59baf24de9e47ea316a315398e8027d198202d"
@@ -28,6 +28,7 @@ PROVIDES="
 	lib:libggml = $portVersion
 	lib:libggml_base = $portVersion
 	lib:libggml_cpu = $portVersion
+	lib:libggml_rpc = $portVersion
 	lib:libllama = $portVersion
 	lib:libllava_shared = $portVersion
 	cmd:convert_hf_to_gguf.py
@@ -73,6 +74,7 @@ PROVIDES="
 	cmd:llama_speculative_simple
 	cmd:llama_tokenize
 	cmd:llama_tts
+	cmd:rpc_server
 	"
 REQUIRES="
 	haiku
@@ -84,6 +86,7 @@ PROVIDES_devel="
 	devel:libggml = $portVersion
 	devel:libggml_base = $portVersion
 	devel:libggml_cpu = $portVersion
+	devel:libggml_rpc = $portVersion
 	devel:libllama = $portVersion
 	devel:libllava_shared = $portVersion
 	"
@@ -118,8 +121,8 @@ BUILD()
 		-DLLAMA_BUILD_SERVER=ON \
 		-DLLAMA_BUILD_TESTS=OFF \
 		-DLLAMA_BUILD_EXAMPLES=ON \
-		-DLLAMA_NATIVE=OFF \
 		-DGGML_NATIVE=OFF \
+		-DGGML_RPC=ON \
 		-DCMAKE_EXE_LINKER_FLAGS="$sourceDir/build_mimalloc/mimalloc.o"
 
 	make -C build $jobArgs
@@ -128,11 +131,13 @@ BUILD()
 INSTALL()
 {
 	make -C build install
+	cp build/bin/rpc-server $binDir
 
 	prepareInstalledDevelLibs \
 		libggml \
 		libggml-base \
 		libggml-cpu \
+		libggml-rpc \
 		libllama \
 		libllava_shared