Question

I know, don’t parse using curl, grep and sed. But I am looking for an easy approach, not a very safe one.

So I get an HTML file with curl, from which I need a value of a certain attribute from a tag. I use grep to get me the line where it says token. This only occurs once.
This gives me a whole div:

<div class="userlinks">
  <span class="arrow flleft profilesettings">settings</span>
  <form class="logoutform" method="post" action="/logout">
    <input class="logoutbtn arrow flright" type="submit" value="Log out">
    <input type="hidden" name="ltoken" value="a5fc8828a42277538f1352cf9ea27a71">
  </form>
</div>

How can I get just the value attribute (e.g. “a5fc8828a42277538f1352cf9ea27a71”)?

Answer 1

There’s no need to grep:

sed -n '/token/s/.*name="ltoken"\s\+value="\([^"]\+\).*/\1/p' input_file