blas error and crash when doing ldivide/rdivide of a scalar by a vector

[Jason Riedy]

And Ben Abbott writes:
> Based upon the posts below, it may be that the lapack-3.1.0 sources
> are fixed

Nope.  I'm using it straight from svn (which still isn't
publically available).  Neither the 3.1 branch nor the trunk are
fixed.

> 	https://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=80&start=0&postdays=0&postorder=asc&highlight=dgelsd
> 	https://icl.cs.utk.edu/lapack-forum/viewtopic.php?t=278

Well, now I know that I'm not to blame.  ;)  The person who fixed
the workspace formula forgot to make the tests consistent.  Lots
of people are out on winter break and holidays, so I don't think
I can get ahold of him right now.

As with all projects, we LAPACK folks definitely are interested
in better testing and workspace queries.  But it's not the top
priority of the current funding grants, so those are open for
other contributors...

Jason