Why did std::set not have a contains function until C++20?

Question

Another design decision that baffled me. Checking if an element is in a set is the entire purpose of a set! Until C++20 I had to write stuff like s.find(x) != s.end() or the nicer s.count() cast as a bool that's almost as good as contains() but somewhat obscures the intent (and runs in O(n) for multisets?). Even if the standard library containers were designed around iterators, count() has existed from the beginning.

The P0458 proposal to add contains is obvious, but there must've been some kind of argument for leaving it out originally and then leaving it out every time it was brought up. Was there some kind of design philosophy change in C++20?

Answer 1

Because they wanted all containers to adhere to the same interface as much as possible, and .count() made sense for more containers, and still worked equivalently for set. They didn't want to break interface uniformity just for clearer semantics in one case.

Later on that decision was re-evaluated.

Also, the lack was not really much of a lack, since you can say if (myset.count(x)) just the same as you'd say if (myset.contains(x)). In modern times, more people run tools that complain about treating ints as bools, so that's become less attractive.