To an Oracle guy like me, this sounds completely weird. A trigger not getting executed just because the table was updated indirectly via some constraint like a Foreign Key update and this being an acceptable bug for over 10 years now is so strange to me that I cannot make any sense of it at all. It shows that some updates happen via a different code path and that scares me to no end.
It is an altogether different matter that coding anything in triggers today sounds anachronistic. This also goes to show that people using mySQL are probably not using triggers to code logic. They are probably more likely to use it as a dumb datastore.