There used to be a quadratic algorithm in character input in readline, the library that many common shells are using to parse commands. So if you copy-pasted a somewhat long list into something like Postgres or the Python shell, it would take a lot of time because it had to check a couple of things on each input, and that somehow involved checking the existing characters in the input (I've forgotten the details).

I think that's a really classical case of O(n^2). You have a linear algorithm (n), and then run it on each instance you insert (n) so n^2.

It bothered me for several years, until I eventually did some profiling and contacted the readline (and bash) maintainer who fixed it.

This is about 5 years ago.

There's also another readline-related problem in several shells that they don't discover if you resize the terminal window. I can't remember if I raised that too. It sometimes amazes me that very few people appear to ever look at or debug our common infrastructure.