E should really be of size N-1, xSYCONV() is not ready for this and so, in the
comment at the header level, we now require E to be of size N. Also xSYTRS2()
is using E(N) = ZERO, so if we make E of size N-1 in xSYCONV(), then we need to
see the impact on xSYTRS2(). This is long term cleaning, so for now, we revert
back to requesting an array E of size N. Thanks Lawrence for noticing this.
*>
*> \param[out] E
*> \verbatim
-*> E is COMPLEX array, dimension (N-1)
+*> E is COMPLEX array, dimension (N)
*> E stores the supdiagonal/subdiagonal of the symmetric 1-by-1
*> or 2-by-2 block diagonal matrix D in LDLT.
*> \endverbatim
*>
*> \param[out] E
*> \verbatim
-*> E is DOUBLE PRECISION array, dimension (N-1)
+*> E is DOUBLE PRECISION array, dimension (N)
*> E stores the supdiagonal/subdiagonal of the symmetric 1-by-1
*> or 2-by-2 block diagonal matrix D in LDLT.
*> \endverbatim
*>
*> \param[out] E
*> \verbatim
-*> E is REAL array, dimension (N-1)
+*> E is REAL array, dimension (N)
*> E stores the supdiagonal/subdiagonal of the symmetric 1-by-1
*> or 2-by-2 block diagonal matrix D in LDLT.
*> \endverbatim
*>
*> \param[out] E
*> \verbatim
-*> E is COMPLEX*16 array, dimension (N-1)
+*> E is COMPLEX*16 array, dimension (N)
*> E stores the supdiagonal/subdiagonal of the symmetric 1-by-1
*> or 2-by-2 block diagonal matrix D in LDLT.
*> \endverbatim