If all curl did was "opening a socket and issuing a GET" it would be about a hundred lines long, including input validation and two kinds of errors.

It does a lot more than that, because the Internet isn't actually that simple. You should look at the first paragraph of https://curl.se rather than have me rehash it here.

CURL supports plenty of protocols, but at the end of the day they're all application-layer and should've easily been written as platform-agnostic business logic. If CURL was architected properly, keeping any OS-specific code behind platform-agnostic interfaces - porting it across operating systems should not have been an issue.

When the codebase depends on POSIX and platform-specific quirks everywhere, putting the blame on the operating system when it's time to port things is hilarious.

I don't see posts like this from the people who had to port Portal 2 to run natively on Linux, as an example. If Portal 2 (a codebase you can safely assume is 1000x more complex) made it fine without the drama, there's no reason CURL can't follow suit (unless the architecture is botched).

I wonder what part of "curl has to run on many versions of a single OS" is so hard to uderstand that people keep making dumb example like this.

Porting a codebase to a specific target is way easier compare to porting it to multi version of the same target, especially the older ones.