Whats the meaning of this sed command sed '/^[[:space:]]*$/d'

Question

I came across this sed command in a script to compare a machine's local binaries to GTFObins binaries.

    for i in $(curl -s https://gtfobins.github.io/ | html2text | cut -d" " -f1 | sed '/^[[:space:]]*$/d');
do 
    if grep -q "$i" installed_pkgs.list;
    then 
        echo "Check GTFO for: $i";
    fi;
done

I have tested it and found that it deletes the spaces between lines, but I don't understand its form or how it is written.

Related: How to remove blank lines from a file (including tab and spaces)? — steeldriver, Nov 23 '23 at 15:19
You probably want something like curl -s https://gtfobins.github.io/ | grep -Po 'class="bin-name">\K[^<]+' | grep -Fxf - installed_pkgs.list — Stéphane Chazelas, Nov 23 '23 at 17:19

score 2 · Accepted Answer · answered Nov 23 '23 at 15:17

2

Matches any line that only contains space characters between the beginning (^) and the end ($) of a line.

https://regex101.com/ is your friend.

/…/d instructs sed to delete any match, so this just eliminates empty lines (as in: no content, or only spaces, tabs and similar on the line) from the input).

answered Nov 23 '23 at 15:17

Marcus Müller

28,836

1

Convoluted way to write grep -vx '[[:space:]]*' – Stéphane Chazelas Nov 23 '23 at 17:12
@StéphaneChazelas you mean "convoluted way to write grep '[^[:space:]]'"? – Marcus Müller Nov 24 '23 at 01:45
Just realized you wrote exactly that in the answer linked the question's comments, ha! – Marcus Müller Nov 24 '23 at 01:47
1

Those are not exactly equivalent for input not fully encoded in the user's locale encoding. grep '[^[:space:]]' could also not return lines that contain invalid text in the locale which is why I didn't suggest it here to avoid that argument. – Stéphane Chazelas Nov 24 '23 at 14:58

Whats the meaning of this sed command sed '/^[[:space:]]*$/d'

1 Answers1