Problem with cblas_dgemv call in DotInc on certain networks.

[e2crawfo]

Try running a SPA associative memory with 1-dimensional vectors. We get output like:

Transpose A: 111
m: 1
n: 10
leading_dim_A: 1
X.stride1: 1
Y.stride1: 1
A_label: <Connection from <Node "element output"> to <Node "output">>.weights
X_label: <Node "element output">.in[(slice(None, None, None),)]
Y_label: <Connection from <Node "element output"> to <Node "output">>.weighted

lda must be >= MAX(N,1): lda=1 N=10Parameter 7 to routine cblas_dgemv was incorrect

The first block of output is diagnostic info from added print statements. Has something to do with treating a matrix as a vector.