This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
RE: possible bug in linalg/balance.c
Aaron Schweiger writes:
> Note, my suggested fix to linalg/balance.c is to check whether s is
> infinite, and if so, invoke GSL_ERROR. However - I don't know if this is
> the 'correct' behavior of the balance algorithm - as I am not familiar with
> it.
Thanks for the test program. The problem is now fixed in CVS. When
the input matrix contains inf/nan, the resulting SVD will contain
NaNs rather than going into an infinite loop.
--
Brian Gough