%f is double, so an explicit conversion is useful

as a hint to the compiler that we know what we're doing

diff --git a/src/lib/util/value.c b/src/lib/util/value.c
index 42fec15dc6dc3886daaf829dd7d531410f6d5ebf..67b2ab72c2b300e4bdeeca0eff9729bcac673d0c 100644 (file)
--- a/src/lib/util/value.c
+++ b/src/lib/util/value.c
@@ -3421,7 +3421,7 @@ char *fr_value_box_asprint(TALLOC_CTX *ctx, fr_value_box_t const *data, char quo
                break;
 
        case FR_TYPE_FLOAT32:
-               p = talloc_typed_asprintf(ctx, "%f", data->vb_float32);
+               p = talloc_typed_asprintf(ctx, "%f", (double) data->vb_float32);
                break;
 
        case FR_TYPE_FLOAT64:
@@ -3607,7 +3607,7 @@ size_t fr_value_box_snprint(char *out, size_t outlen, fr_value_box_t const *data
                return snprintf(out, outlen, "%" PRId64, data->vb_int64);
 
        case FR_TYPE_FLOAT32:
-               return snprintf(out, outlen, "%f", data->vb_float32);
+               return snprintf(out, outlen, "%f", (double) data->vb_float32);
 
        case FR_TYPE_FLOAT64:
                return snprintf(out, outlen, "%g", data->vb_float64);