The cool thing about programming languages like Ruby is that they’re very conducive to one-liners. You can use them to generate bash statements that you can then pipe back to bash. `ls | ruby -nle 'puts "mv #{$_} #{$_.gsub(/_/, ".")}"' | bash` is a pattern I use all the time, for example. Easy way of generating complex bash statements without having to bother with the bash man page.

Do you need to pipe the ruby output to bash? Ruby could just run the mv directly by wrapping in backticks instead of quotes.

Running it through the pipe lets you preview the command first and make sure it actually looks correct. Then you just re-run the command and append the bash pipe.

ls | while read filename; do mv "$filename" "$(echo "$filename" | sed "s/_/./g")" || break; done

I think there's a variable expansion syntax that does replacement, but I haven't learned it so I use sed.

I find that piping to a while loop is often simpler and more flexible than something like you've written, but maybe not in this specific case with the weird string replacement syntax. However, mine doesn't have a command injection vulnerability (that I know of).

A real scripting language can also do this easily:

for f in os.listdir("."): os.rename(f, f.replace("_","."))

Sometimes I'm tempted to make Python my shell. The fewer different syntaxes you need to learn for the same basic operations, the better!

If you want to use ruby then just use ruby? That is a very fragile way to use bash that could easily lead to bad things.

Seems to be the flavour of the day :)