Reminds me of when I wrote an app to track and store tweets about a given hashtag. Worked fine in its first few months, then started throwing encoding errors right around the time Twitter enabled emoji support. I was using MySQL as a data store and was ignorant to the fact that UTF8 encoding in MySQL only uses 3 bytes. I was ignorant about encoding in general but thought that UTF8 always meant 4-bytes. Apparently so did many other MySQL users, and it was bad enough that MySQL questions on Stackoverflow are already filled with erroneous info about all kinds of encoding issues.
The search for the fix -- which is that MySQL has a utf8mb4 encoding for 4-bytes -- was probably the most frustrating debugging exercise in my career. Even worse was the eventual revelation that the MySQL devs were technically right -- the UTF8 spec says nothing about requiring 4 bytes, and 3 bytes probably seemed sensible enough in terms of efficiency at the time of MySQL 4.