Interesting code observation/interpretation. Real and in fact many software are built with the same pre-assumptions. That, "ah... 32-bit should be enough for this long". CPU's were the same thing... then eventual 64bit had to arrive to overcome memory addressing problems, etc... then IPv4/6 same thing...

If we would do the same analysis to many software out there, we would probably find many more in the same situation. In fact, in that year (I am probably not going to be here to witness) many thing things will fail if still using 32-bit registers (probably even before, due to offsets considerations in some cases), mostly because of the time like you described, but also not just into the future, also towards the past.

One common example of another failure is trying to go index "old" things... where the current register width does not allow to go far enough. And when you realize, you can't read data from the past! Usually these are not that problematic because you can then adapt the software to overcome that problem. But in conjunction with old hardware based technology it's usually a trickier problem to solve.