From: Kresimir Fresl (fresl_at_[hidden])
Date: 2007-03-18 11:42:29
Karl Meerbergen wrote:
> What is your point?
You said: "The implementation of leading_dimension() should not rely on
size_m, but rather on size1() and size2() depending on the orientation
of the matrix."
Original version (that is, all versions prior to revision 1.16) did not
rely on size_m, but on size2(). So why did you change size2() into
size_M(), as quoted CVS log shows, if you think that it "is a problem"?