Boost logo

Boost Users :

Subject: Re: [Boost-users] Boost_1_54_0: boost::math::sph_bessel(n, x) crash at x=0 n>0
From: JE Campagne (campagne_at_[hidden])
Date: 2013-10-18 06:22:10

Your fix pass all my personnal tests. Thanks.
Sorry for the trouble, according to the limit close to zero.

Bye the way do you know when 1_55 version will be publicly available?


Le 18/10/2013 11:01, John Maddock a écrit :
>> 1) in your patch for T operator()() part, what's the "n" variable,
>> looking at 1_54_0 version there is no mention to this variable.
> There are no changes to operator()(), the n variable appears in the
> next part of the diff in
> inline T sph_bessel_j_imp(unsigned n, T x, const Policy& pol)
>> 2) is there a way to get your current version of the bessel.hpp file ?
> Attached.
> Note I had to change the previous fix somewhat as it failed for types
> with extended exponent ranges (I realised that just as I was dozing
> off to sleep last night!)
>> That's true that for x close to zero but positive there are also
>> problems to the use of power series close to 0 is certainly the best
>> thing to do. I am astonished that this not yet used.
> Sorry I don't understand: the problem was occurring inside the power
> series for small x.
> John.
> _______________________________________________
> Boost-users mailing list
> Boost-users_at_[hidden]

Boost-users list run by williamkempf at, kalb at, bjorn.karlsson at, gregod at, wekempf at