I agree that the structural benefit of glass is marginal, probably to the point of insignificance. I was just trying to make the point that you give up nothing in terms of torsional stiffness with a glass canopy. (This is different from saying that a glass roof offers more protection against cabin penetration in a rollover. However, even on that score glass roofs have done very well in testing.)
In the Lucid Air and the new Model S, the glass panels are larger than the openings and sit atop the roof frame members.
I think there are two aspects to this. Think about a piece of sheet metal the same thickness as an exterior car panel. Then think about a piece of tempered, laminated glass such as is used in a car roof. If you apply force to the opposite edges of the sheet metal and apply the same force to the opposite edges of the glass, or you apply equal torsional forces to each piece of material, which will deform first? A: The sheet metal. (The glass will eventually shatter while the sheet metal will only crumple, but the glass will remain relatively less deformed up until the point it shatters than the sheet metal will up until the point it crumples.)
The other aspect is the strength and flexibility of the adhesive. While there may be some give in the adhesion for the reason MikeTz pointed out, the adhesion will still exert a force countering the forces that are trying to pull the glass away from the surface to which it is mounted. To that extent you would get resistance to the twisting forces of a car frame at the points of contact with the glass.
(I'm not a physicist or engineer and would welcome any correction or further elucidation on this issue. This is just the lay understanding I've gleaned from doing some reading a while back on this topic.)