help-gsl
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Help-gsl] cspline: gsl_spline_eval


From: Alex
Subject: Re: [Help-gsl] cspline: gsl_spline_eval
Date: Mon, 12 Dec 2011 03:27:44 +0100

Hi,

thank you for  the explanation.

If i got it right, i can use the following line to stay with the old 
functionality?

spline->interp->type->eval (spline->interp->state, spline->x, spline->y, 
spline->interp->size, MyXValue, acc, &MyYValue);

I would like to check the return value. GSL_SUCCESS is 0?

Sincerely,
Alex

-----Urspr√ľngliche Nachricht-----
Von: Rhys Ulerich [mailto:address@hidden 
Gesendet: Montag, 12. Dezember 2011 01:50
An: Alex
Cc: address@hidden
Betreff: Re: [Help-gsl] cspline: gsl_spline_eval

> Hi, i upgraded from openSuse 11.3 to 12.1 and got a new gsl version....
>
> But now i'm getting always error messages "interp.c:150: ERROR: 
> interpolation error". I think it's because i want y values for x 
> values lower/bigger than the x values in my interpolation points. How can i 
> get these values? I don't want to add additional
interpolation points because that will change my spline.

Changed in http://bzr.savannah.gnu.org/lh/gsl/trunk/revision/4704.

Judging from the line
   GSL_ERROR_VAL("interpolation error", GSL_EDOM, GSL_NAN); within interp.c, I 
don't think you can get the older, extrapolating
behavior.

The related mailing list thread discussing the change was here:
http://lists.gnu.org/archive/html/bug-gsl/2010-11/msg00010.html

Your options seem to be:
1) Don't extrapolate.
2) Hack your local version of the code to remove the extrapolation check.
3) Use some basic facts about splines to write an extrapolation-friendly 
version which defers to the library version.

Sorry,
Rhys




reply via email to

[Prev in Thread] Current Thread [Next in Thread]