That's an interesting idea too. However, if you did make 1,2-dibromopropane then whichever bromine formed the Grignard first would almost immediated eliminate the other bromine to form propene. Remember that you can use 1,2-dibromoethane to activate the Mg turnings when you make a Grignard reagent and in that case, essentially the same process occurs (producing ethylene in that case). I have some experience with making multiply charged Grignard reagents (that is, molecules with more than one R-Mg-Br) and it is actually very difficult to get the second one to form while the first forms relatively rapidly.