[Gsll-devel] Wrong dimension is assumed when an optional parameter is not specified in sv-solve

[Liam Healy]

Pushed; thanks for the patch.

On Wed, Jan 6, 2010 at 2:55 PM, Evgeniy Zhemchugov <jini.zh at gmail.com> wrote:
> SV-decomposition of matrix A (m x n) produces matrices U (m x n) and V
> (n x n) and a vector s representing the diagonal matrix S (n x n) so
> that A = U S V'. The solution of the equation Ax = b, x, is of
> dimension n that is equal to (dim0 V) rather than (dim0 U) as is
> written in SV-solve for its optional parameter. Hence the patch.
> Obviously, bug appears only for non-rectanular A, so it passed the
> tests.