Discrepancy in Ruby CLI Option `-p` Behavior with and without Pipe

Question

I'm observing different behaviors with the Ruby command-line option -p depending on whether it's used with or without a pipe. Here are the two scenarios:

Without Pipe:

• When I run ruby -pe 'puts "You said #{$_}", it seems to just output the input as it is after what puts prints.

With Pipe:

• However, using a pipe as in echo "hello world" | ruby -pe '$_.upcase! results in the expected transformation, printing an upcased "HELLO WORLD".

Can someone explain why there's a difference in behavior between these two scenarios? Is the -p option interacting differently with the input in each case?

FYI. The results are below.

$ echo "hello world" | ruby -pe '$_.upcase!'
HELLO WORLD

$ ruby -pe 'puts "You said #{$_}"'
1
You said 1
1
...

Answer 1

The -p flag, according to the ruby(1) man page, will "print the value of variable $_ at the each end of the loop.". In your example with the pipe, this is working as you expect: using upcase! will replace the value of $_ and Ruby will print it out afterwards.

However, in your example without a pipe, you're using puts to explicitely print out a string, which happens in addition to the -p behaviour. Using the first example with and without a pipe works the same way:

$ ruby -pe '$_.upcase!'
aa
AA
$ echo "aa" | ruby -pe '$_.upcase!'
AA

You can also use the -n flag, which does the same as -p but won't automatically print out anything. For example:

$ ruby -ne 'puts "You said #{$_}"'
aa
You said aa
bb
You said bb
cc
You said cc