Detect overflow on large precision

test/format-test.cc

@@ -934,6 +934,9 @@ TEST(format_test, precision) {
   EXPECT_THROW_MSG((void)fmt::format(runtime("{:.{}e}"), 42.0,
                                      fmt::detail::max_value<int>()),
                    format_error, "number is too big");
+  EXPECT_THROW_MSG(
+      (void)fmt::format("{:.2147483646f}", -2.2121295195081227E+304),
+      format_error, "number is too big");
 
   EXPECT_EQ("st", fmt::format("{0:.2}", "str"));
 }

test/fuzzing/float.cc

@@ -12,7 +12,7 @@
 
 void check_round_trip(fmt::string_view format_str, double value) {
   auto buffer = fmt::memory_buffer();
-  fmt::format_to(buffer, format_str, value);
+  fmt::format_to(std::back_inserter(buffer), format_str, value);
 
   if (std::isnan(value)) {
     auto nan = std::signbit(value) ? "-nan" : "nan";