Fix an overflow bug with f32 to s32 conversion.

resources/format_conversions.txt

@@ -15,6 +15,7 @@
 # clip [output] [a]     -> output = clamp(a, -1, 1)
 #
 # int [name] -> Declare an uninitialized 32-bit integer
+# lng [name] -> Declare an uninitialized 64-bit integer
 # flt [name] -> Declare an uninitialized 32-bit float
 
 # r = (x - 128) << 8
@@ -159,7 +160,7 @@ f32->s24 {
 # r = clip(x) * (0x7FFFFFFF + sign(x))
 f32->s32 {
     flt c;
-    int s;
+    lng s;
     clip c x;
     sig s x;
     add s s 2147483647;

tools/malgen/source/malgen.cpp

@@ -455,6 +455,9 @@ std::string malgen_generate_code__conversion_func_inst(malgen_context* pContext,
     if (strcmp(pInst->name, "int") == 0) {
         code += "int "; code += pInst->params[0];
     }
+    if (strcmp(pInst->name, "lng") == 0) {
+        code += "long long "; code += pInst->params[0];
+    }
     if (strcmp(pInst->name, "flt") == 0) {
         code += "float "; code += pInst->params[0];
     }