Recent changes to bugs

bug in Hist2D_get

Anonymous — Mon, 18 Jun 2007 22:31:25 -0000

A small bug in the 2D histogram get method. The method checks the number of bins in the x direction instead of the y direction for the index j.

The line 226 in Hist"D.c reads

if (nj < 0 || nj > ptr->nx)
rb_raise(rb_eRangeError, "Index j is outside of allowed range.");

but it should read
if (nj < 0 || nj > ptr->ny)
...

samples/matrix_complex.rb segfaults

Tobias Hunger — Sun, 07 Mar 2004 10:24:49 -0000

Maybe due to the #ifdef I chaged to get around bug no.
911312?

Vector_complex fails to build

Tobias Hunger — Sun, 07 Mar 2004 10:24:02 -0000

Vector_complex_allocate is not defined. I just changed
#ifdef HAVE_DEFINE_ALLOC_FUNCTION to #ifdef 0 to get
around this problem.

GSL renamed constants

Anonymous — Sun, 07 Mar 2004 10:18:51 -0000

Hi!

GSL seems to have renamed its constants:
GSL_CONST_CGS_* turned into GSL_CONST_CGSM_* and
GSL_CONST_MKS_ turned into GSL_CONST_MKSA_*.

Make failes in Consts.c due to this rename. It is
trivial to fix with a regexp replacing CGS with CGSM
and MKS with MKSA.

Afterwards you need to do the same change in ruby-gsl.h
as well.

cygwin problems

Arno Erpenbeck — Sat, 03 May 2003 16:20:57 -0000

Looks like the prebuilt GSL lib on cygwin is missing two
functions.
The extconf.rb should check the platform and provide some
fix (comment out) to the relevant files.

crash in MonteCarlo integration

Arno Erpenbeck — Sat, 19 Apr 2003 16:53:00 -0000

When using many iterations in MonteCarlo integration,
the program crashes from time to time:

---
p2 = Proc.new {|k, p|
p / (1.0 - cos(k[0]) * cos(k[1]) * cos(k[2]))
}

a = 1.0 / (M_PI * M_PI * M_PI)
dim = 3

f = Function.new dim, a, p2
plain = Plain.new dim

xl = [0.0, 0.0, 0.0]
xu = [M_PI, M_PI, M_PI]
calls = 10000
rng = Random::RNG.new

r = plain.integrate f, xl, xu, calls, rng
result = r[0]
abserr = r[1]
---

Crash occurs in Monte.c (MFunction_function) when the
call to the Proc object is made (rb_funcall).

I suspect it is something with the memory management.
Probably the COPYCARRAY/COPYRUBYARRAY macros
do some harm if called too often.

BTW: Could this be a Ruby problem (see attached file)?

bug in Vector.c

Pernici Mario — Tue, 15 Apr 2003 08:56:14 -0000

Hi Arno,
there is a bug in Vector.c:
static VALUE Vector_subvector(VALUE obj, VALUE o, VALUE
n) {
gsl_vector *vector, *v2;
gsl_vector_view vtmp;
Data_Get_Struct(obj, gsl_vector, vector);
v2 = gsl_vector_alloc(NUM2INT(n));

// vtmp = gsl_vector_subvector(vector, NUM2INT(o), n);
// becomes
vtmp = gsl_vector_subvector(vector, NUM2INT(o),
NUM2INT(n));

*v2 = vtmp.vector;
return Data_Wrap_Struct(rbgsl_cVector, 0,
gsl_vector_free, v2);
}
================================
require "GSL"
include GSL

v1 = Vector.new((0..9).to_a)
v2 = v1.subvector(4,3)
v2.to_a #-> [4.0, 5.0, 6.0]
=================================
So long
Mario

misprint in Vector.c

Pernici Mario — Mon, 14 Apr 2003 11:20:11 -0000

Dear Arno,
In Vector.c , there is a misprint:

//rb_define_alias(rbgsl_cVector, "*", "div");
rb_define_alias(rbgsl_cVector, "/", "div");

Best regards
Mario