I noticed that the author uses cmd.exe in one of the charts. I wonder if this is...

spudlyo · on Aug 22, 2012

Why would I want a shell script to change my $CWD after the shell script exits?

Maybe you wanted a script that would use shortcuts to take you to cumbersome directory paths?

I enjoyed this article, even though some of it was review. I didn't know that anything in a pipeline got its own separate subshell in bash, and that different shells behaved differently in that regard. I'm always happy to learn interesting uses for process substitution, and got a kick out of:

    read sum < <(echo "2+3" | bc)

Good stuff.

dfc · on Aug 22, 2012

For cumbersome directory paths:

  $ tail -1 ~/.bash_aliases 
  CMBRSM=/opt/var/dont-know-about-FHS/local/etc/cache/config
  $ cd $CMBRSM

Unlike a shell script an alias allows you to:

  $ ls $CMBRSM/subdirectory
  $ cat $CMBRSM/file.txt

dllthomas · on Aug 22, 2012

> Why would I want a shell script to change my $CWD after the shell script exits?

If you called the script with the intent that it change your $CWD, obviously.

The solution, of course, is to source the script rather than executing it. This is mentioned in the article, but requires an understanding (at least a minimal one) of subshells to get why - which is precisely what the article addresses.