Question

I’m trying to get the redirect link from a site by using curl -I then grep to “location” and then sed out the location text so that I am left with the URL.

But this doesn’t work. It outputs the URL to screen and doesn’t put it

test=$(curl -I "http://www.redirectURL.com/" 2> /dev/null | grep "location" | sed -E 's/location:[ ]+//g')
echo "1..$test..2"

Which then outputs:

..2http://www.newURLfromRedirect.com/bla

What’s going on?

Answer 1

As @user353852 points out, you have a carriage return character in you output from curl that is only apparent when you try to echo any character after it. The less pager shows this up as ^M

You can use sed to remove “control characters”, like in this example:

% test=$(curl -I "http://www.redirectURL.com/" 2>|/dev/null | awk '/^Location:/ { print $2 }' | sed -e 's/[[:cntrl:]]//') && echo "1..${test}..2"
1..http://www.redirecturl.com..2

Notes:

I used awk rather than your grep [...] | sed approach, saving one process.

For me, curl returns the location in a line starting with ‘Location:’ (with a capital ‘L’), if your version is really reporting it with a lowercase ‘l’, then you may need to change the regular expression accordingly.