Vagrant bash autocomplete

gistfile1.sh

# Autocompletion for Vagrant just put this line in your ~/.profile

complete -W "$(echo `vagrant --help | awk '/box/,/up/ {print $1}'`;)" vagrant

good contribution, thanks. I tried replacing "--help" with "list-commands" but the "rsync" option does not appear. You know why?

@Velaa98 There's a possibility you have a plugin installed which provides a vagrant command with the string "up" in its name, for example something like vagrant-hostsupdater. If this is the case then you might be getting a shorter command list from vagrant list-commands | awk '/box/,/up/ {print $1}' than you were expecting.

I saw a comment on stackoverflow that made me think this might work:

# Vagrant bash completion.
complete -W "$(echo $(vagrant list-commands | sed '1,3d' | awk '{print $1}'))" vagrant

As far as I understand it...

1. We use the complete bash builtin function, pass it a word list -W "..." and tell it to do its magic on the vagrant command.
2. The word list is constructed by using two nested command substitions - $() - "it literally plugs the command output into another context".
3. The outermost command substitution echo will print the results of the innermost one vagrant list-commands | sed '1,3d' | awk '{print $1}.
4. vagrant list-commands will give us a list of commands vagrant knows about, but with 3 lines of descriptive text at the start.
5. The sed command 1,3d will delete the first 3 lines of the output piped to it, this is how the header text is chopped off.
6. The awk program {print $1}prints the first field for each line of output. A field to awk is anything up to the first white-space character for each line, this is how the descriptions are removed, leaving just the vagrant command name.

I think that's how it works.