[SOLVED] bash if statement not working

trscookie · 06-21-2016, 11:47 PM

Hello all,

I am trying to test the condition of an if statement against a curl response, however for some reason it will not work no matter how many ways I try, please can someone take a look and see what I could be doing wrong?

Code:

#!/bin/bash

HTTP_200=$(curl -sS -I -L http://127.0.0.1/ | /bin/grep '^HTTP/1.1 200 OK' | awk '{printf "%s", $3}')
echo "op: $HTTP_200"
echo "XXX${HTTP_200}XXX"

if [ "$HTTP_200" = "OK" ]; then
        echo "WORKING"
else
        echo "not working"
fi

output:

Code:

op: OK
XXXOK
not working

trscookie · 06-21-2016, 11:54 PM

I've noticed this, is there a way to compare against this as a string

Code:

curl -sS -I -L http://127.0.0.1 | /bin/grep '^HTTP/1.1 200 OK' |od -c
0000000   H   T   T   P   /   1   .   1       2   0   0       O   K  \r
0000020  \n
0000021

Turbocapitalist · 06-22-2016, 12:50 AM

The carriage return is what's causing the problem. You can use a search and replace to eliminated it or just use your own output:

Code:

$(curl -sS -I -L http://127.0.0.1/ | awk '/^HTTP\/1.1 200 OK/ { printf "OK" }');

trscookie · 06-22-2016, 01:04 AM

Awesome, cheers!

grail · 06-22-2016, 01:20 AM

I am curious if there is likely to be the word 'OK' anywhere else in the header? If not, you could simplify a little:

Code:

if curl -sS -I -L http://127.0.0.1 | grep -q OK
then
  echo "WORKING"
else
  echo "not working"
fi

Turbocapitalist · 06-22-2016, 01:30 AM

Quote:

Originally Posted by grail

I am curious if there is likely to be the word 'OK' anywhere else in the header?

It's not likely but still possible. One of the other headers could contain the two letters next to each other. The "grep" solution could still look for the right HTTP response status

Code:

if curl -sS -I -L http://127.0.0.1 | grep -qE '^HTTP\/1.1 200 OK'
then
  echo "WORKING"
else
  echo "not working"
fi

The earlier bit about the hidden carriage return brings up an important point that any and all data downloaded (or otherwise inputted) from external sources, especially the net, should not be trusted until it is processed and either replaced or properly cleaned. perl, for example, has a taint mode to detect and tract the spread of such tainted data within a script. Sanitize and validate externally acquired data before using it.

Habitual · 06-22-2016, 06:09 AM

Whole script looks familiar.

I did this 5 years ago:

Code:

###
#!/bin/bash
# Written by user@domain.tld
# Monitors an Apache server and emails the Admin if the
# server reports anything other than "Apache"
# http://bashscripts.org/forum/viewtopic.php?f=16&t=1522
# Last Edited on Thu Dec 29, 2011 - 10:11:40 PM EST
###
MAIL_ME="user@address.com"
MAIL_SUBJECT="Web_Server_needs_attention"
SERVER_DATE=$(TZ=PST8PDT date "+%a %b %d %I:%M:%S %p %Z")
LOCAL_DATE=$(date +%c)

webserver=$(curl -Is http://address.com | \grep -E '^Server' | cut -c9-21)
echo "$SERVER_DATE" - Web Server Check = "$webserver" >> ~/server.log

if [ "$webserver" = "Apache" ] ; then
exit 0
else
 echo "My_Server_Server_Check - FAILED" on "$LOCAL_DATE" | mail "$MAIL_ME" -s "$MAIL_SUBJECT"
 exit 1
fi
#EOF

Hope that helps. May even serve as a bad example. IDK.