This is the mail archive of the
gsl-discuss@sourceware.cygnus.com
mailing list for the GSL project.
bugs(s) in specfunc
- To: gsl-discuss at sourceware dot cygnus dot com
- Subject: bugs(s) in specfunc
- From: Thomas Walter <walter at pctc dot chemie dot uni-erlangen dot de>
- Date: Mon, 19 Jun 2000 09:52:58 +0200
Hello,
while writing some interface functions between 'octave' and 'gsl' I
think I found a few bugs.
The files 'gsl_sf_coulomb.h' and 'coulomb.c' seem out of sync.
The header file declares
int gsl_sf_coulomb_wave_FGp_impl()
and the source file defines
int
gsl_sf_coulomb_wave_FGp_array_impl()
The file 'log.c' defines the functions
int
gsl_sf_log_1plusx_impl()
and
int
gsl_sf_log_1plusx_mx_impl()
I think the used series expansion for small x
if(fabs(x) < GSL_ROOT5_DBL_EPSILON)
has the wrong sign for the constant 'c6'.
Also the series seems too short to fill a 'double' with valid digits.
You can verify this especially for 'gsl_sf_log_1plusx_mx_impl()' if
you call this function with values near the border
'GSL_ROOT5_DBL_EPSILON'. I suggest to extend the series at least with
the constants 'c7 = -1/8' and 'c8 = 1/9'.
Bye
Thomas
--
Warum hat eine Leiterbahn keine 90 Grad Winkel sondern 2 mal 45 Grad?
Weil die Elektronen sonst immer gegen die Wand laufen 8-)))
----------------------------------------------
Dipl. Phys. Thomas Walter
Inst. f. Physiklische Chemie II
Egerlandstr. 3 Tel.: ++9131-85 27326 / 27330
91058 Erlangen, Germany email: walter@pctc.chemie.uni-erlangen.de