#882 removes the transposes from

i0 ( p3 p4 h1 h2 )_vt + = 1/2 * Sum ( p5 p6 ) * t ( p5 p6 h1 h2 )_t * v ( p3 p4 p5 p6 )_v

We can do the same to eomccsd_x2_7 and eomicsd_x2_7.

There are also a bunch of other places that don't matter (CISD, QCISD, CC2) where we can probably just replace their routine with ccsd_t2_8 (although check the intorb effects).