diff options
Diffstat (limited to 'src/corelib/serialization/qjsonvalue.cpp')
| -rw-r--r-- | src/corelib/serialization/qjsonvalue.cpp | 7 |
1 files changed, 2 insertions, 5 deletions
diff --git a/src/corelib/serialization/qjsonvalue.cpp b/src/corelib/serialization/qjsonvalue.cpp index 27a2f0e227..29c29184c1 100644 --- a/src/corelib/serialization/qjsonvalue.cpp +++ b/src/corelib/serialization/qjsonvalue.cpp @@ -449,11 +449,8 @@ void QJsonValue::swap(QJsonValue &other) noexcept also affects this function. A floating point value that is either an infinity or NaN will be converted - to a null JSON value. Since Qt 6.0, QJsonValue can store the full precision - of any 64-bit signed integer without loss, but in previous versions values - outside the range of ±2^53 may lose precision. Unsigned 64-bit values - greater than or equal to 2^63 will either lose precision or alias to - negative values, so QMetaType::ULongLong should be avoided. + to a null JSON value. The values outside the range of ±2^53 may lose precision, + because they are converted to a double QJsonValue. For other types not listed above, a conversion to string will be attempted, usually but not always by calling QVariant::toString(). If the conversion |