Allow overflow/underflow when parsing floating constants

1 files changed, 1 insertions, 1 deletions

diff --git a/expr.c b/expr.c
index 92cd4d9..9e92fd3 100644
--- a/expr.c
+++ b/expr.c
@@ -208,7 +208,7 @@ primaryexpr(struct scope *s)
 			/* floating constant */
 			errno = 0;
 			e->constant.f = strtod(tok.lit, &end);
-			if (errno)
+			if (errno && errno != ERANGE)
 				error(&tok.loc, "invalid floating constant '%s': %s", tok.lit, strerror(errno));
 			if (!end[0])
 				e->type = &typedouble;