OK, so it's definitely the suspension arms that are bent.
I've dismantled the two lower arms on the right side, which is where the bushing popped out. And then I have measured them, as described in the workshop manual, section C.3, figure 17. The dimension '4' should be 58.7mm (I can't for the moment measure that more precise than 59mm)
As can be seen from the attached photo IMG_Right_Side_Lower_Front_5a.jpg, the arm with the ruined bushing is measuring 48mm, that is 11mm too little! The rearmost arm is measuring 57mm, 2mm too little. Of course these measurements are again not very precise, the arms can wobble a little since they are a little rounded on the flat part, and also the flat area for the trunnion is a little indented so the true measurement point is hard to get to. But it's good enough for government work, as they say.
I lined up the 2 arms (photos IMG_Right_Side_Lower_1a.jpg' & IMG_Right_Side_Lower_1a.jpg) and got to the total of 105mm (48mm+57mm). This is 13mm shorter than the 118mm it should have been, and the 13mm corresponds almost to the 15mm that the bushing has moved.
I take this as a very clear indication, and will take a look at the lower left side also at a later time. The top arms are harder to get to with the rod going into the cabin and all, so I think I will just loosen up the nuts and see if the bushings are jammed in any direction, i.e. if there is a similar problem.
Where did I do wrong?
I think that it was a mistake to assemble the outer ends to the uprights by hand and then believe that the dimensions were correct. It was probably first when I torqued it all up, including the outer ends where the shock bushings has a lot of rubber to compress, that the wrong (bent) dimensions came into play.
So, what to do now?
Since the front arm is definitely bent, measuring it by adding a washer (1.5mm) will bring the measurement to 57mm also, see photo IMG_Right_Side_Lower_Front_6a.jpg. The washer will sit between the shock bushing and the suspension arm,so that when the trunnion and shock bushings are both torqued, the a-arms will keep their proper locations.
Does anyone see a problem in doing this?
Oh, and BTW if any of you who has experienced similar bushing problems have the possibility to measure the dimension '4', it could be interesting to hear what you find.