Last Updated: 2024-05-01

I wanted to use the diff command to compare a function repeated almost verbatim within a file. I found a way to focus on that repeated function with sed:

First instance: $ sed -n -e '286,541p' resources/js/maps.js

Second: $ sed -n -e '16,266p' resources/js/maps.js

When I tried to diff I ran into a problem: diff could only compare files - whereas I was dealing with process outputs. The solution was process substitution:

diff <(sed -n -e '286,541p' resources/js/maps.js) <(sed -n -e '16,266p' resources/js/maps.js) | nvim

• Process substitution works with <(command) or >(command)
• <(command) feeds the output of a process (or processes) into the stdin of another process.
• There is no space between the "<" or ">" and the parentheses. Space there would give an error message.
• Behind the scenes the OS creates a file, even on macos

echo >(true)
# >(true) turned into fd 12
/dev/fd/12

The name of this file is passed as an argument to the current command as the result of the expansion

Process substitution can compare the output of two different commands, or even the output of different options to the same command:

comm <(ls -l) <(ls -al)
...
total 12
...
total 20

Or it can compare the contents of two directories:

diff <(ls $first_directory) <(ls $second_directory)

• Process substitution has nothing to do with redirection < and in fact can be used in parallel with it: cat < <(date) - this is a fairly useless example but shows data converted to a file with process substitution, then input direction connecting the file to STDIN

Resources

• https://www.tldp.org/LDP/abs/html/process-sub.html
• https://unix.stackexchange.com/questions/17107/process-substitution-and-pipe