[alec]

I'm a big fan of Joey Hess' moreutils - https://joeyh.name/code/moreutils/

Examples:

- sponge: read in all of stdin, then write to the given file. Great for pipelines: sed "s/root/toor/" /etc/passwd | grep -v joey | sponge /etc/passwd

- vipe: easily drop a $EDITOR instance in the middle of a pipe chain

- ts: timestamp all standard input; great for long-running output

[stevekemp]

I hope my own collection of related tools is also useful?

https://github.com/skx/sysadmin-util

More focused on those a sysadmin might prefer than general-use, but contributions or updates are welcome.

[chipb]

Although there's some real gems in there (I can't believe I never ran across vipe before!), its Debian package conflicts with the GNU parallel package: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=718816 Quite annoying.

[wging]

How is sponge different from > filename?

[tmhedberg]

For one thing, it allows your pipeline to both read from and write to the same file, as it defers the output until there is no more data to read from "upstream". For instance, this won't work (it truncates the file for writing before it is ever read):

    grep foo <some_file >some_file

But this will have the intended effect:

    grep foo <some_file | sponge some_file

[loevborg]

It can be used where ">" does not work, e.g. when using sudo. So you can do

  echo data | sudo sponge /etc/a.file

`tee` can be abused in a similar way, but it also prints the data to stdout, which is ugly.

[justincormack]

sed can edit in place though.

[kylek]

indeed!

sed -i -e 's/root/toor/' -e '/joey/d' /etc/passwd