It's not a matter of the rifle failing, they are damn near impossible to blow up. The MkIII/1910 action is probably only equalled or exceeded by some of the Arisakas for sheer strength.
The problem was the potential to manually rotate the bolt head in such a way that the bolt would slide into battery, but without the bolt head rotating into the locked position. The early M10 .280s had the end of the extractor slot shaped like a ramp making it easy to turn the bolt head out of the correct position when cleaning or fiddling around with the bolt out of the rifle. This was later changed to a square-ended slot so that the extractor would have to be pulled up out of the groove before the bolt head could be so rotated.
Herbert Cox who was an expert machinist and gunsmith and a Ross rifle aficionado claimed to have found a way to make the 1905 action fail to lock also.
The "1907" or "Scotch Deer Stalker" .280 has a receiver that looks like the 1905 with the interrupted thread bolthead and lugs of the 1910 action. It is generally considered that they also could not be wrongly assembled, but I have the remains of one here that blew out its bolt, so that action also is not entirely foolproof.
It may be that in some very unusual circumstances a degree of "self-unlocking" is possible, (the Blish Principle etc), even without incorrect assembly, but the jury is still out on that one and probably always will be.