Why is it not necessary to test third-party libraries?

Question

Software testing

I often hear that it is not necessary to test third-party libraries, but I just can’t understand what it actually means ... For example, I put a library in the project that sends specific TCP requests, and accordingly I use several of its methods to implement the request.
Imagine a library is being updated and some of its methods have either changed their name or, even worse, their behavior. As a result, it turns out that the request will be invalid or will not take place at all. Personally, this happened to me, and the worst thing is that it does not immediately open up. In the presence of not a small project and 100500 libraries, I can’t follow everyone.
To resolve this situation, every time I add a new library (or a new method), I write a test for it (or for its API, it may be more correct to call it), as a result, after each update I do a test and see whether it has fallen or not.
Is that how everyone does it or am I cycling?

Answer 1

Imagine a library is being updated and some of its methods have either changed their name

unit test case

their behavior is even worse.

unit test + functional test case

each time a new library is added

do not use libraries not covered by tests