A bug in the current _M_insert_float/_M_group_float

Paolo Carlini pcarlini@unitus.it
Fri Dec 19 17:35:00 GMT 2003


Martin Sebor wrote:

> ...                                                            (even
> though the intent should be obvious, i.e., thousands_sep should only
> be allowed in the integral part of the number).

Thanks. Indeed, the parsing code implicitly makes this assumption.

Paolo.



More information about the Libstdc++ mailing list