Advantages of preferring stdin over /dev/tty for reading input

[aschey]

Hi, I have a bit of a design question. The default behavior of this library is to read input from stdin by default and only open /dev/tty if stdin is redirected. I've seen it done this way in a few libraries, but I haven't been able to find much discussion on what the advantage of this is over just opening a tty and avoiding the conditional logic. I can think of a few scenarios where /dev/tty is not available (executing a specific command over ssh, running docker without -it, etc.), but stdin is also not a tty in those scenarios, so it doesn't seem to matter there. The other (very minor) advantage I can think of would be to avoid opening a separate file descriptor, but I'm curious to know if there's anything else that I'm missing.

Thanks!

[meowgorithm]

Off the top of my head, I think that's pretty much it: not having to open another file descriptor unless we have to, which could result in an error. Also note that on Windows we're essentially doing the same thing, but in a Windows specific way via $CONIN, so there's an extra code path there as well.

I don't have a strong opinion about this though but @aymanbagabas, who has rearchitected much of Bubble Tea lately, might.

[aschey]

Thanks for the reply! That makes sense. I should probably clarify that my question was purely out of curiosity rather than any kind of statement about which method is better.