My know-it-all AI bot suggests the same... "Character Encoding Differences: The maxlength attribute counts UTF-16 code units, while server-side processing might count bytes or Unicode code points. For example, emojis or certain special characters (like accented letters) may use multiple code units or bytes"

(For perspective on this next bit, this is a big complex app written over many years by many people, mostly me for the last 5.) Apparently the line of code that would have validated the length of that field was there, but was commented out in 2013, by the very very smart person who wrote a great deal of this app. If I put it back, remove maxlenth in the form with browser dev tools, and put in text that's too long, the validation error gets triggered as expected. It's possible that hidden or multibyte characters might have slipped through anyway, hard to say without more testing I don't have time for. There's no indication as to why that validation was removed, grrr. This is a particularly knotty section of code, so I'm a bit shy about putting it back for real. I can hear all the "just rerun your test cases" righteous hollers back there, but alas, there's no such animal, and building them is prohibitive, especially for this bit. And yes I know that makes them an especially good idea to have, but I cannae do it sir.

Ah. Looks like the Safari version is 18.3.1 so it’s not an older version if it’s a human submission