Aren't there already ballistics tests for this sort of thing, to match up the marks on the bullet with individual guns?
It's not really all that reliable.
First of all: the bullet would have to be intact enough to make out the scratches left by the rifling, which isn't always the case. (Also, shotguns don't leave such markings.)
Second: Sometimes all that the rifling will reveal is the type of the gun used in the crime, sometimes even that isn't clear if several guns use the same cartridge and have similar rifling.
Third: Even if they did reveal the type of gun used, finding out the actual gun used in the crime would mean finding the unique imperfections and wear and tear in the gun which is even harder and these characteristic change as the gun is used. (Besides the criminal might just scrub the barrel really hard to change the barrel enough that it cannot be confirmed with 100% probability.)
As for proposed "fingerprinting" what if someone goes to the gun range to collect casings (some people collect them to reload their own ammo so there are legimite reasons for doing that) and plants them on the crime scene? (after gathering his own casings.)