Merge commit 'c342c3b93de358a4571941b41c35dc5ba2081145' into concedo_experimental

# Conflicts:
#	.github/workflows/build.yml
#	CODEOWNERS
#	scripts/sync_vendor.py

Author: LostRuins

common/debug.cpp

@@ -45,6 +45,8 @@ static float common_ggml_get_float_value(const uint8_t * data,
     return v;
 }
 
+#define INDENT "    "
+
 template <bool abort>
 void common_debug_print_tensor(uint8_t * data, ggml_type type, const int64_t * ne, const size_t * nb, int64_t n) {
     GGML_ASSERT(n > 0);
@@ -60,41 +62,41 @@ void common_debug_print_tensor(uint8_t * data, ggml_type type, const int64_t * n
         }
     }
     for (int64_t i3 = 0; i3 < ne[3]; i3++) {
-        LOG_ERR("                                     [\n");
+        LOG(INDENT "[\n");
         for (int64_t i2 = 0; i2 < ne[2]; i2++) {
             if (i2 == n && ne[2] > 2 * n) {
-                LOG_ERR("                                      ..., \n");
+                LOG(INDENT INDENT "..., \n");
                 i2 = ne[2] - n;
             }
-            LOG_ERR("                                      [\n");
+            LOG(INDENT INDENT "[\n");
             for (int64_t i1 = 0; i1 < ne[1]; i1++) {
                 if (i1 == n && ne[1] > 2 * n) {
-                    LOG_ERR("                                       ..., \n");
+                    LOG(INDENT INDENT INDENT "..., \n");
                     i1 = ne[1] - n;
                 }
-                LOG_ERR("                                       [");
+                LOG(INDENT INDENT INDENT "[");
                 for (int64_t i0 = 0; i0 < ne[0]; i0++) {
                     if (i0 == n && ne[0] > 2 * n) {
-                        LOG_ERR("..., ");
+                        LOG("   ..., ");
                         i0 = ne[0] - n;
                     }
                     const float v = common_ggml_get_float_value(data, type, nb, i0, i1, i2, i3);
-                    LOG_ERR("%12.4f", v);
+                    LOG("%12.4f", v);
                     if (i0 < ne[0] - 1) {
-                        LOG_ERR(", ");
+                        LOG(", ");
                     }
                 }
-                LOG_ERR("],\n");
+                LOG("  ],\n");
             }
-            LOG_ERR("                                      ],\n");
+            LOG(INDENT INDENT "],\n");
         }
-        LOG_ERR("                                     ]\n");
-        LOG_ERR("                                     sum = %f\n", sum);
+        LOG(INDENT "]\n");
+        LOG(INDENT "sum = %f\n", sum);
     }
 
     if constexpr (abort) {
         if (std::isnan(sum)) {
-            LOG_ERR("encountered NaN - aborting\n");
+            LOG("encountered NaN - aborting\n");
             exit(0);
         }
     }
@@ -137,9 +139,9 @@ template <bool abort_on_nan> bool common_debug_cb_eval(struct ggml_tensor * t, b
     }
 
     if (matches_filter) {
-        LOG_ERR("%s: %24s = (%s) %10s(%s{%s}, %s}) = {%s}\n", __func__, t->name, ggml_type_name(t->type),
-                ggml_op_desc(t), src0->name, common_ggml_ne_string(src0).c_str(), src1 ? src1_str : "",
-                common_ggml_ne_string(t).c_str());
+        LOG("%s: %24s = (%s) %10s(%s{%s}, %s}) = {%s}\n", __func__, t->name, ggml_type_name(t->type),
+            ggml_op_desc(t), src0->name, common_ggml_ne_string(src0).c_str(), src1 ? src1_str : "",
+            common_ggml_ne_string(t).c_str());
     }
 
     const bool is_host = ggml_backend_buffer_is_host(t->buffer);

ggml/src/ggml-metal/ggml-metal.cpp

@@ -7,6 +7,9 @@
 #include "ggml-metal-context.h"
 #include "ggml-metal-ops.h"
 
+#include <mutex>
+#include <string>
+
 #define GGML_METAL_NAME "MTL"
 #define GGML_METAL_MAX_DEVICES 16
 

src/models/qwen3next.cpp

@@ -265,13 +265,13 @@ std::pair<ggml_tensor *, ggml_tensor *> llm_build_qwen3next::build_delta_net_chu
     cb(g_diff, "g_diff", il); // shape: (chunk_size, 1, n_chunks, H_v * n_seqs)
 
     ggml_tensor * g_diff_exp = ggml_exp(ctx0, g_diff);
-    ggml_tensor * g_diff_exp_t = ggml_reshape_4d(ctx0, g_diff_exp, 
+    ggml_tensor * g_diff_exp_t = ggml_reshape_4d(ctx0, g_diff_exp,
                                                  1, chunk_size, n_chunks, g_diff_exp->ne[3]);
 
     ggml_tensor * key_gdiff = ggml_mul(ctx0, k, g_diff_exp_t);
     cb(key_gdiff, "key_gdiff", il); // shape: (S_k, chunk_size, n_chunks, H_v * n_seqs)
 
-    ggml_tensor * key_gdiff_t = ggml_cont(ctx0, ggml_transpose(ctx0, key_gdiff)); 
+    ggml_tensor * key_gdiff_t = ggml_cont(ctx0, ggml_transpose(ctx0, key_gdiff));
     cb(key_gdiff_t, "key_gdiff_t", il); // shape: (chunk_size, S_k, n_chunks, H_v * n_seqs)
 
 

vendor/cpp-httplib/CMakeLists.txt

@@ -3,9 +3,14 @@ license_add_file("cpp-httplib" "LICENSE")
 
 find_package(Threads REQUIRED)
 
+llama_add_compile_flags()
+
 add_library(${TARGET} STATIC httplib.cpp httplib.h)
-if (NOT MSVC)
-    # disable warnings in 3rd party code
+
+# disable warnings in 3rd party code
+if (CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
+    target_compile_options(${TARGET} PRIVATE /w)
+else()
     target_compile_options(${TARGET} PRIVATE -w)
 endif()
 
@@ -146,6 +151,23 @@ elseif (LLAMA_OPENSSL)
     endif()
 endif()
 
+# disable warnings in 3rd party code
+if(LLAMA_BUILD_BORINGSSL OR LLAMA_BUILD_LIBRESSL)
+    if (CMAKE_CXX_COMPILER_ID STREQUAL "MSVC")
+        target_compile_options(ssl PRIVATE /w)
+        target_compile_options(crypto PRIVATE /w)
+        if(LLAMA_BUILD_BORINGSSL)
+            target_compile_options(fipsmodule PRIVATE /w)
+        endif()
+    else()
+        target_compile_options(ssl PRIVATE -w)
+        target_compile_options(crypto PRIVATE -w)
+        if(LLAMA_BUILD_BORINGSSL)
+            target_compile_options(fipsmodule PRIVATE -w)
+        endif()
+    endif()
+endif()
+
 if (CPPHTTPLIB_OPENSSL_SUPPORT)
     target_compile_definitions(${TARGET} PUBLIC CPPHTTPLIB_OPENSSL_SUPPORT) # used in server.cpp
     if (APPLE AND CMAKE_SYSTEM_NAME STREQUAL "Darwin")