and both the base URL needed to end with a slash, and each path must begin with one.

...so the respective library functions each sanitizes the input meets their requirement, pre- and appending slashes as needed?

...right?

Every reasonable programmer would assume that this is a mistake because the final path would end up with two slashes, but the library actually required that.

For fuck's sake

I'm guessing they have two instances of string validation, the developer for each of which helpfully decided "I'll make sure there's a slash, idk if the other end checks it" and then a third function that trims the slashes from the parts and concatenates them with a slash for a separator.

But, man, is this ever stupid.