[MPFRCPP] Re: MPFRCPP

[Alexey Beshenov]

1. MPFRCPP does not wrap features which are not present in MPFR.

2. Also, MFPRCPP does not support operations between fixed precision types and 
mpfr::Real due to potential loss of precision (so don't expect '+' or '*' to 
be overloaded for mpfr::Real---double).

On Tuesday 08 April 2008 22:29, jtl wrote:
> I advanced in using your library.
> There is a little feature lacking : float construction assignement and
> conversion. Was it voluntary
> For me its necessary to have the feature. I do not know if  mpfr
> supports floats but the lines below make the
> join for me.
>
>
>            Real& operator= (const float& y) throw() {
>                 mpfr_set_d (getMpfrT(), y,
> getParameters().getDefaultRoundMode().getMpfrRndT());
>                 return *this;
>             }
>
>            operator float () const throw (ConversionDoesNotFitsError)  {
>                 if (isFits<float>())
>                     return mpfr_get_d (getMpfrT(),
> getParameters().getDefaultRoundMode().getMpfrRndT());
>                 else
>                     throw ConversionDoesNotFitsError ();
>             }
>
>            Real (const float& x, const RoundMode& rm) throw() {
>                 mpfr_init2(getMpfrT(),
> getParameters().getDefaultPrecision().getMpfrPrecT());
>                 mpfr_set_d (getMpfrT(), x, rm.getMpfrRndT ());
>             }
>
>             Real (const float& x, const Precision& pr =
> getParameters().getDefaultPrecision(), const RoundMode& rm =
> getParameters().getDefaultRoundMode()) throw() {
>                 mpfr_init2(getMpfrT(), pr.getMpfrPrecT());
>                 mpfr_set_d (getMpfrT(), x, rm.getMpfrRndT ());
>             }
>
>
> Do you think it could be added ?
>
>                   thanks
>
>                                   JT     Lapresté

-- 
Alexey Beshenov <address@hidden>
http://beshenov.ru/

pgpHtc9Wr9tQ_.pgp
Description: PGP signature