#!/bin/sh scheme=$(printf "%s\n" "$1" | uricut -s) path=$(printf "%s\n" "$1" | uricut -p) qs=$(printf "%s\n" "$1" | uricut -q) domain=$(printf "%s\n" "$1" | uricut -d) port=$(printf "%s\n" "$1" | uricut -P) if [ ! "$port" ];then if [ "$scheme" = "https" ];then port=443 SSL=--ssl fi if [ "$scheme" = "http" ];then port=80 fi fi if [ "$qs" ];then path="${path}?${qs}" fi if [ "$port" ];then UA="Mozilla/5.0 (impersonator)" # content_type="$(printf "HEAD %s HTTP/1.1\r\nHost: %s\r\nUser-Agent: %s\r\n\r\n" "$path" "$domain" "$UA" | ncat -4 $SSL "$domain" "$port" | grep -i '^Content-Type: ' | head -n1 | cut '-d ' -f2 | cut '-d;' -f1 | tr -d '\r\n')" a_header="$(curl -Lsi "$1" | egrep -ai '^Location: |^Content-Type: ' | head -n1 | tr -d '\r\n')" if printf "%s\n" "${a_header}" | grep -i^Content-Type: 2>&1 >/dev/null 2>&1;then content_type="$(printf '%s\n' "${a_header}" | cut '-d ' -f2- | cut -d; -f1)" else content_type="redirect probably" fi if [ "${content_type}" = "text/html" -o "${content_type}" = "application/xhtml+xml" ];then # title="$(printf "GET %s HTTP/1.1\r\nHost: %s\r\nUser-Agent: %s\r\n\r\n" "$path" "$domain" "$UA" | ncat -4 $SSL "$domain" "$port" | head -c 10000 | tr -d '\n' | tr '<' '\n' | grep -A 10 '^title>' | grep -B 10 '^\/title>' | cut '-d>' -f2)" title="$(curl -si "$1" | head -c 10000 | tr -d '\n' | tr '<' '\n' | grep -iA 10 '^title' | grep -iB 10 '^\/title>' | cut '-d>' -f2 | sed 's/^ .//g' | sed 's/ .$//g' | grep .)" printf "title: %s\n" "$title" | html_entities_decode else printf "%s\n" "${a_header}" fi else printf "DONT KNOW HOW TO GET TITLE FOR THIS URL: %s\n" "$1" fi