ci : add an option to fail on compile warning (llama/3952)

* feat(ci): add an option to fail on compile warning

* Update CMakeLists.txt

* minor : fix compile warnings

ggml-ci

* ggml : fix unreachable code warnings

ggml-ci

* ci : disable fatal warnings for windows, ios and tvos

* ggml : fix strncpy warning

* ci : disable fatal warnings for MPI build

* ci : add fatal warnings to ggml-ci

ggml-ci

---------

Co-authored-by: Georgi Gerganov <[email protected]>

ggml-backend.c

@@ -1006,6 +1006,7 @@ static int ggml_backend_sched_backend_from_buffer(ggml_backend_sched_t sched, gg
         }
     }
     GGML_ASSERT(false && "tensor buffer type not supported by any backend");
+    return -1; // silence warning
 }
 
 #if 0

ggml-metal.m

@@ -176,7 +176,7 @@ struct ggml_metal_context {
 // MSL code
 // TODO: move the contents here when ready
 //       for now it is easier to work in a separate file
-//static NSString * const msl_library_source = @"see metal.metal";
+// static NSString * const msl_library_source = @"see metal.metal";
 
 // Here to assist with NSBundle Path Hack
 @interface GGMLMetalClass : NSObject

ggml.c

@@ -868,7 +868,7 @@ do {                                                              \
     const __m128 t0 = _mm_add_ps(_mm256_castps256_ps128(x[0]),    \
                                  _mm256_extractf128_ps(x[0], 1)); \
     const __m128 t1 = _mm_hadd_ps(t0, t0);                        \
-    res = _mm_cvtss_f32(_mm_hadd_ps(t1, t1));                     \
+    res = (ggml_float) _mm_cvtss_f32(_mm_hadd_ps(t1, t1));        \
 } while (0)
 // TODO: is this optimal ?
 
@@ -1149,7 +1149,7 @@ inline static void __wasm_f16x4_store(ggml_fp16_t * p, v128_t x) {
         x[i] = _mm_add_ps(x[i], x[offset+i]);                     \
     }                                                             \
     const __m128 t0 = _mm_hadd_ps(x[0], x[0]);                    \
-    res = _mm_cvtss_f32(_mm_hadd_ps(t0, t0));                     \
+    res = (ggml_float) _mm_cvtss_f32(_mm_hadd_ps(t0, t0));        \
 }
 // TODO: is this optimal ?
 
@@ -2086,6 +2086,7 @@ void ggml_numa_init(enum ggml_numa_strategy numa_flag) {
         }
     }
 #else
+    GGML_UNUSED(numa_flag);
     // TODO
 #endif
 }
@@ -3219,7 +3220,7 @@ const char * ggml_get_name(const struct ggml_tensor * tensor) {
 }
 
 struct ggml_tensor * ggml_set_name(struct ggml_tensor * tensor, const char * name) {
-    strncpy(tensor->name, name, sizeof(tensor->name));
+    strncpy(tensor->name, name, sizeof(tensor->name) - 1);
     tensor->name[sizeof(tensor->name) - 1] = '\0';
     return tensor;
 }
@@ -18575,7 +18576,9 @@ static enum ggml_opt_result linesearch_backtracking(
         (*step) *= width;
     }
 
-    GGML_UNREACHABLE();
+    GGML_ASSERT(false && "line search failed");
+
+    return GGML_LINESEARCH_FAIL;
 }
 
 static enum ggml_opt_result ggml_opt_lbfgs(
@@ -18843,7 +18846,9 @@ static enum ggml_opt_result ggml_opt_lbfgs(
         step[0] = 1.0;
     }
 
-    GGML_UNREACHABLE();
+    GGML_ASSERT(false && "lbfgs failed");
+
+    return GGML_OPT_DID_NOT_CONVERGE;
 }
 
 struct ggml_opt_params ggml_opt_default_params(enum ggml_opt_type type) {