Enrico Forestieri wrote: > I did that. It turned out that the compiler num_get implementation > finishes like this: > > if (__beg == __end) > __err |= ios_base::eofbit; > return __beg; > > So, if there's nothing after the last character part of the number being > read, the eofbit is set, invalidating the read itself! > > I don't know if this is a bug (it seems so to me), but anyway...
I don't think that it is a bug: If there is nothing after the last character you can't know whether there was an error or whether the stream was really finished. > Note that it would suffice appending any character not being part of > the current numpunct facet, so the '\0' is pretty ok. If it works... (but I would have used a space, '\0' can be surprising). Georg
