I don't think it's entirely unreasonable to expect someone to use an available library for parsing dates.
It's not a terribly good example.
There are cases where "reinventing the wheel," is simply an ignorant statement. Perhaps the system in question has developed semantics for working with date-based data that the grammar of the existing date-parsing library's API doesn't match well with. This extremely contrived example may force the developers to write a lot of code for working around those inconsistencies which themselves can lead to bugs and maintenence headaches in the future.
What if your lack-wit developer who wrote his own date-parsing library saw these problems and wrote a library whose API resolved them? Or perhaps in another scenario the existing libraries are simply inefficient for some metric of acceptable performance in your system. Maybe this ignorant developer simply writes a better library that allows you to do more.
Either way, there is reason to believe that either scenario is valid. It is likely that writing a date parsing library is not going to be terribly useful. That doesn't mean it's impossible and so one must be willing to evaluate evidence and draw conclusions rather than hide behind blanket statements.
What I'm really talking about is the attitude people have when they think someone is "reinventing the wheel." While some people are helpful and will politely point out that there are existing date-parsing libraries, there are a great many more it seems who will assume that the person who is "reinventing the wheel" is a clueless loser who wouldn't make such a "stupid mistake" if they were worth their salt. That kind of reaction isn't helpful at all.