If it's close enough that you can make it work without bending the TREs using a shallow sump and relieving the transmission tunnel a bit, I think that's what I'd do, TBH.
Yes ok I see where the description has gone off course. "Bending" no. But....nearly there.
So the inner ball joint HAS to be in the correct place, in line with the wishbone pivot, vertically on the centre line pivot. But that the rack body must be lower to miss the engine. (Ignoring modding the sump for a minute)
So, in order to connect the two components there would have to be a vertical connection added. Say, for argument sake, a one inch square solid metal steel bar 40(?)mm long with a hole drilled in each end. Fitted vertically. So the bottom hole bolts to the rack end, where the tre once was in its lower position, the top hole bolts to the tre inner ball joint end 40mm directly above.
By 40mm I mean whatever the distance has to be. Might be nearer 50mm.
But does that make sense? And, is it structurally string enough? Would the rack shaft bend in corners? Would it then binde up.
The pic in post 12 doesn't appear to make sense to anyone, but it shows what I mean. Except, that pic shows end plates. And they point rearward, rather than vertically as mine would, to allow the rack body to be forward of the engine, but allows the tre ball joints to be in the correct/other/wherever place, as needed.
It's the ball joint position that's key, the tre rods can be whatever shape they want to reach the wheel. Purely bending the rod, means little as Aaron is struggling with I think. By that description one could argue oe omega rods are cranked. But no that won't help in this application, well, it might if the rods four the wheel on full lock of course.
So by cranking tre's I mean, cranking the pivot points, not the rods themselves. Or even crank the rack.
Rod horizontal to ball joint->------o o-------- >hub(ditto opposite<)
40mm con rod -> | ^sump^ |
Bolt to end of rack shaft->o --------rack body------o