[GH-667] core: Avoid value truncation on 32bit architectures

Replace `labs()` with `llabs()` to avoid truncation on 32bit
architectures. Also do not tolerate value truncation in the CI build.

Author: mlasch

core/utils.c

@@ -218,7 +218,7 @@ size_t utils_intToText(int64_t data,
             string[int64_min_strlen] = '\0';
             return int64_min_strlen;
         }
-        result = utils_uintToText((uint64_t)labs(data), string + 1, length - 1);
+        result = utils_uintToText(llabs(data), string + 1, length - 1);
         if(result != 0)
         {
             result += 1;

wakaama.cmake

@@ -124,6 +124,7 @@ add_compile_options(
     -Wshadow
     -Wswitch-default
     -Wwrite-strings
+    -Werror=absolute-value
     -pedantic
 )