You will want to read the man pages. 'pee' is straightforward to understand if y...

thiht · on April 16, 2022

I know tee but have no idea what pee would do considering this description.

From your description I guess something like « pee a b » would be the same as « a | b »? If so that’s cool, but the one line descriptions definitely need a rework.

oftenwrong · on April 16, 2022

It loosely follows the same metaphor as the T-shaped pipe connector that 'tee' is named for.

    $ seq 3 | tee somefile
    1
    2
    3

            ╔═somefile
    seq═tee═╣    
            ╚═stdout


    $ seq 10 | pee 'head -n 3' 'tail -n 3'
    1
    2
    3
    8 
    9
    10


            ╔═head═╗
    seq═pee═╣      ╠═stdout
            ╚═tail═╝

nerdponx · on April 16, 2022

I believe it's more like `a | pee b x y z | c`. This would run `b x y x` as well as `c` using the output from `a`.

I think Zsh supports this natively with its "multios" option, `a > >(b c y z) > >(c)`. But then you have to write the rest of your pipe inside the ().

somat · on April 16, 2022

I can only guess from that description is that it is the cat command.

somat · on April 16, 2022

The unix shell was designed to optimize for simple pipes. to make pipe networks you usually have to use named pipes, aka fifos.

I once had to use an IBM mainframe shell(cms If I remember correctly) which had pipes however IBM in their infinite wisdom decided to make the pipe network the primary interface, while this made complex pipes a bit less awkward, simple(single output) pipes were wordy compared to the unix equivalent.