Re: [Quantlib-dev] default 0*Days value for Period in QLXL

[Luigi B. <lui...@gm...>]

On Mon, 2009-10-12 at 16:42 +0200, Ferdinando Ametrano wrote:
> I could provide a non-explicit string constructor for
> QuantLib::Period(std::string s) as solution, but something tells me
> Luigi might frown at this (or God forbid, even balk at this ;-)

At the very least, I'd raise an eyebrow :)

Would it be feasible to use a little adapter class in QLA? Something
like:

class QlaPeriod {
    QuantLib::Period p_;
  public:
    // input in as many ways as you like...
    QlaPeriod(QuantLib::Period p) : p_(p) {}
    QlaPeriod(std::string s) : p_(PeriodParser::parse(s)) {}

    // ...then pass it wherever a period is needed:
    operator QuantLib::Period() const { return p_; }
};

Your various ObjectHandler::convert2 functions would return QlaPeriods
and the underlying functions would accept them after the automatic
conversion.

Luigi

P.S. On a related note, how about the hideous 0*Years vs 0*Days thing?
I'm still shaking from that one :)


-- 

I'd never join any club that would have the likes of me as a member. 
-- Groucho Marx