[flang][runtime] Underflow real input to -0. when negative (#82443)

When the result of real input editing underflows to zero, return -0.
when the input field had a minus sign.

flang/lib/Decimal/decimal-to-binary.cpp

@@ -14,6 +14,7 @@
 #include <cinttypes>
 #include <cstring>
 #include <ctype.h>
+#include <utility>
 
 namespace Fortran::decimal {
 
@@ -275,7 +276,12 @@ ConversionToBinaryResult<PREC> IntermediateFloat<PREC>::ToBinary(
         if (guard != 0) {
           flags |= Underflow;
         }
-        return {Binary{}, static_cast<enum ConversionResultFlags>(flags)};
+        Binary zero;
+        if (isNegative) {
+          zero.Negate();
+        }
+        return {
+            std::move(zero), static_cast<enum ConversionResultFlags>(flags)};
       }
     }
   } else {

flang/unittests/Runtime/NumericalFormatTest.cpp

@@ -916,7 +916,7 @@ TEST(IOApiTests, EditDoubleInputValues) {
       {"(RU,F7.0)", "-1.e999", 0xffefffffffffffff, 0}, // -HUGE()
       {"(E9.1)", " 1.0E-325", 0x0, 0},
       {"(RU,E9.1)", " 1.0E-325", 0x1, 0},
-      {"(E9.1)", "-1.0E-325", 0x0, 0},
+      {"(E9.1)", "-1.0E-325", 0x8000000000000000, 0},
       {"(RD,E9.1)", "-1.0E-325", 0x8000000000000001, 0},
   };
   for (auto const &[format, data, want, iostat] : testCases) {