Fix build failure

[mikenichols]

As of the new year, we started seeing a test failure that I believe was caused by the intersection of:

1. A test using a relative offset into the past
2. The need to correctly identify a valid 2-digit number as a prefix for a valid 4-digit year (e.g. 20 is a potentially valid expiration year because you could continue typing and input 2030, which is valid)
3. The test helper function yearsFromNow which optionally gives you the last 2 digits of a year (e.g. given the current year is 2025, yearsFromNow(-5, 2) will give you "20" as output)

So, why did the test fail?

• On December 31st, 2024, the code yearsFromNow(-5, 2) returned "19", which is not a potentially valid prefix; the test captured this correctly. However, on January 1st 2025, the code yearsFromNow(-5, 2) returned "20", which is a potentially valid prefix, and the test that was expecting it to not be potentially valid, now fails.

I wanted to make the fix be as unobtrusive as possible, so my preferred solution is to replace the test that was failing with 3 hard-coded tests that show the expected behavior of the year prefixes "19", "20", and "21". This fixes the build with a small diff, and is a robust enough solution to be valid until 2081 when "21" becomes a valid year prefix.

[ibooker]

Does this change warrant a changlog.md entry?

[GoogilyBoogily]

👍 🐸

[GoogilyBoogily]

Does this change warrant a changlog.md entry?

I personally don't think so. I am thinking about it in the same way as version bumps. Source code changes only (excluding tests and docs)