13

<(commands ...) in bash/zsh makes the output behavior as a file. Does a POSIX equivalent exist?

Cyrus
  • 77,979
  • 13
  • 71
  • 125
Sandlayth
  • 155
  • 8
  • 4
    [POSIX does not specify process substitution, but you may used named pipes to accomplish the same thing.](http://mywiki.wooledge.org/ProcessSubstitution) – bishop Aug 05 '16 at 19:09
  • Possible duplicate of [Is it possible to mimic process substitution on msys /mingw (with bash 3.x)](http://stackoverflow.com/questions/20866832/is-it-possible-to-mimic-process-substitution-on-msys-mingw-with-bash-3-x) – P.P Aug 05 '16 at 19:16
  • 2
    @P.P., I'm not really fond of that specific potential dupe, inasmuch as it folds in some misconceptions (among them, that bash 3.x *doesn't* support process substitution; that's a msys restriction, as opposed to a version-based one). – Charles Duffy Aug 05 '16 at 19:22

1 Answers1

14
mkfifo foo.fifo

## if your "commands" is multiple commands
# { commands ...; } >foo.fifo &

# otherwise, if it's just one
commands ... >foo.fifo &

something_else foo.fifo

is the closest available equivalent to

something_else <( commands ... )
Charles Duffy
  • 257,635
  • 38
  • 339
  • 400
  • Possible "robustness" improvements would be to use something like `mktemp` to generate a directory for one of more pipes safe from collisions (or just to generate a name to use for the pipe, since `mkfifo` is atomic), etc. In reality, that's all ` – mtraceur Sep 05 '16 at 07:52
  • 2
    @mtraceur, that's what ` – Charles Duffy Sep 05 '16 at 18:54
  • Right. Thanks for the correction - I forgot all about that. – mtraceur Sep 05 '16 at 19:28
  • 1
    There's no POSIX way to do this because POSIX does not define `mktemp`. One either needs to reimplement `mktemp` with using only POSIX shell behavior or have to come up with some kind of hacks. – Mikko Rantalainen Apr 26 '20 at 11:56