Bash history in script: `!#:*`

Question

I have been reading through the study guides for the LPIC-1.

echo "This is a sentence. " !#:* !#:1->text3

I'm having trouble understanding how the above line of code repeats the echo command multiple times. I know that it is using a feature of bash's history but I can't find any documentation on !#:* or !#:1. Could someone explain this for me?

Answer 1

Yes, this is using history.  !# is a history event designator that refers to the entire command line typed so far.  :* is a word (range) designator that refers to all of the words, except the 0th.  So, after you have typed echo "This is a sentence. ", then !#:* expands to "This is a sentence.  ".  And x-y (where x and y are integers) is a word (range) designator that refers to word number x through word number y.  If y is omitted (x-), this is interpreted to mean word number x through the second to last word.  So, after your “entire command line typed so far” stands as

echo "This is a sentence. " "This is a sentence. "

then !#:1- expands to "This is a sentence. ", because each of the quoted "This is a sentence. " strings counts as one word, and so !#:1- is equivalent to !#:1 (just word number 1).  So you end up with

echo "This is a sentence. " "This is a sentence. " "This is a sentence. " >text3

The fact that the - and the > appear together in the command is just a confusion; they don’t interact.  And the fact that “This is a sentence.” is quoted obscures what is going on; if you said

echo This is a sentence. !#:* !#:1-

it would expand to

echo This is a sentence. This is a sentence. !#:1-

and thence to

echo This is a sentence. This is a sentence. This is a sentence. This is a

(because !#:1- expands to word number 1 through the second to last word.)

Answer 2

A good documentation about bash history cheat sheet by Peter Krumins you can found here.

Since you understood !#, so we just explain :* and :1.

• *: refer to all word but the zeroth.
• x-: refer to the words from x to the second to last word.

So your command:

echo "This is a sentence. " !#:* !#:1-

• !#:* get all but the zeroth, so it get "This is a sentence. ", your command becomes echo "This is a sentence. " This is a sentence. ".

• !#:1- get word from 1 to second to last word, meaning "This is a sentence. " again, your command becomes echo "This is a sentence. " "This is a sentence. " "This is a sentence. "

Answer 3

As @glennjackman points out, this is all covered in the bash manual. Search the man page for HISTORY EXPANSION.

• !# is an event designator meaning "the entire command line typed so far". In your example that's echo "This is a sentence. ", including a trailing space after the final double quotation mark.

• : separates the event designator (!#) from the word designator (*).

• * is a word designator meaning "all of the words in the event except the zeroth"

So echo "This is a sentence. " !#:* expands to

echo "This is a sentence. " "This is a sentence. " $

(the dollar sign is mine, indicating end of white space) which echoes

This is a sentence.  This is a sentence. $

to the terminal (again, the dollar sign is mine, indicating end of white space).

The second event designator:word designator pair (!#:1-) functions in a similar manner (but bear in mind it has to deal with the fact that the command line has been expanded considerably now that the first designator pair has been evaluated).

Answer 4

I have experimented with !#:* and !#:1. I could not find many sources online for this but I am sure that you introduce an echo "Sample text" and then you can use !#:* !#:1 !#:2->text and the !#:* will store the string contents in there temporarily. I am not sure if it is a kind of conditional statement as I was unable to find much resources on it, but it seems this is useful when you are attempting to code the same thing and want to save keystrokes. I have tried this in bash:

echo "test" !#:* !#:1 !#:2 !#:3 !#:4->test1

This is what I get in test1:

test test test test test test

Pretty nifty tool.