On 08/30/2016 01:00 AM, Patrick Alken wrote:
On 08/28/2016 08:37 PM, Peter Johansson wrote:
I also ran our test suite for a package that links in GSL and with
2.1.91 I get some errors. Looks like functionality related to either
svd or the matrix struct has changed. I'll try to boil down the
problem and come back, but if you have any idea of things that have
changed, that would be very useful. Thanks.
Cheers,
Peter
Neither the gsl_matrix struct nor the gsl_linalg_svd routines should
have changed as far as I know. If you post the errors you get I might
be able to help further. Does your code work with GSL 2.1?
Thanks, but since the error is all wrapped up in our classes and test
code, I don't think it would make much sense to you. The other function
I saw that is called in the class is 'gsl_linalg_cholesky_decomp'. I'll
try to peel down to a minimal test case, but busy atm so will not have
time the next couple of days, and since 2.2 already is out I reckon
there is no rush to identify the potential problem.
Cheers,
Peter
