#!/usr/bin/env bash
uri="$1"

scheme=$(printf "%s\n" "$uri" | uricut -s)
path=$(printf "%s\n" "$uri" | uricut -p)
qs=$(printf "%s\n" "$uri" | uricut -q)
domain=$(printf "%s\n" "$uri" | uricut -d)
port=$(printf "%s\n" "$uri" | uricut -P)
if [ "$scheme" = "http" -o "$scheme" = "https" ];then
  ## dereference ONLY ONCE. give up after that.
  newuri="$(unshorten.sh "$uri")"
  if [ "$newuri" ];then
    uri="$newuri"
  fi
#  newuri="$(unshorten.sh "$uri")"
#  if [ "$newuri" ];then
#    uri="$newuri"
#  fi
fi

if [ ! "$port" ];then
  if [ "$scheme" = "https" ];then
    port=443
    SSL=--ssl
  fi
  if [ "$scheme" = "http" ];then
    port=80
  fi
fi
case "$scheme" in
http*)
  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 -gA "$UA" -Lsi "$uri" || echo curl failed) | head -c 10000 | egrep -ai '^curl failed|^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' "${a_header}" | cut '-d ' -f2- | cut '-d;' -f1)"
    fi
    #if printf "%s\n" "${a_header}" | grep -i '^Location: ' 2>&1 >/dev/null 2>&1;then
    #  content_type="redirect. ${a_header}"
    #fi
    if printf "%s\n" "${a_hreader}" | grep -i '^curl failed' 2>&1 >/dev/null 2>&1;then
      content_type="curl failed. cert expired? dunno yet. TODO: code openssl checker."
    fi
### main.lv doesn't have content-type on some pages, so if the content-type is missing or empty, we're assuming html
    if [ "${content_type}" = "text/html" -o "${content_type}" = "application/xhtml+xml" -o "${content_type}" = "" ];then
#    title="$(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 -gsi "$uri" | head -c 1000000 | tr -d '\n' | tr '<' '\n' | grep -iA 10 '^title' | grep -iB 10 '^\/title>' | cut '-d>' -f2 | tr '\t' ' ' | sed 's/^ *//g' | sed 's/ *$//g' | grep .)"
      if [ "${content_type}" = "" ];then
        printf "WTF: header: %s\n" "${a_header}"
        printf "WARNING: NO CONTENT-TYPE RETURNED FROM SERVER. Assuming text/html. title: %s\n" "$title" | html_entities_decode
      else
        printf "title: %s\n" "$title" | html_entities_decode
      fi
    else
      printf "%s\n" "${a_header}"
    fi
  fi
  ;;
gemini)
  if [ "$qs" ];then
    path="${path}?${qs}"
  fi
  first=1
  gemini-get "$uri" | while read -r line;do
    if [ "$first" ];then
      unset first
      type="$(printf "%s\n" "$line" | tr -s ' ' | cut '-d ' -f2 | tr -d '\r')"
      if [ "$type" != 'text/gemini' ];then
        if [ "$type" = "text/html" ];then
          head -c 1000000 | tr -d '\n' | tr '<' '\n' | grep -iA 10 '^title' | grep -iB 10 '^\/title>' | cut '-d>' -f2 | tr '\t' ' ' | sed 's/^ *//g' | sed 's/ *$//g' | grep .
        elif [ "$type" = "text/plain" ];then
          printf "title: %s\n" "$(head -n1)"
        else
          printf "title: %s\n" "$(printf "%s\n" "$line" | tr '\t' ' ' | tr -s ' ' | cut '-d ' -f2-)"
        fi
      fi
    else
      printf "title: %s\n" "$(printf "%s\n" "$line" | grep '^#' | sed 's/^#* *//g')"
    fi
  done | head -n1
  ;;
magnet)
  printf "title: %s\n" "$(printf "%s\n" "$uri" | tr '&' '\n' | grep ^dn= | cut -d= -f2- | uriunescape)"
  ;;
dns)
  if [ "$qs" ] ; then
    export QUERY_STRING="$(printf "%s\n" "$qs" | tr 'A-Z' 'a-z' | tr ';' '&')"
    if [ "$(query_param type)" ];then
      type="-t $(query_param "type")"
    fi
    if [ "$(query_param class)" ];then
      class="-c $(query_param "class")"
    fi
  fi
  if [ "$domain" ];then
    server="@$domain"
    path="$(printf "%s\n" "$path" | sed "s|^/||")"
  else
    server=""
  fi
  dig $class $type "$path" +short $server | tr '\n' ' '
  echo # the tr above strips out the trailing \n
  ;;
ftp)
  curl -g "$uri" 2>&1 | tail -n1
  ;;
gopher)
  if [ "$qs" ];then
    path="${path}?${qs}"
  fi
  type="$(printf "%s\n" "$uri" | uricut -p | cut -b2- | cut -b1)"
  if [ "$type" = 1 -o "$type" = "" ];then
    printf "title: %s\n" "$(curl -gs "$uri" | grep ^i | head -n1 | cut -f1 | cut -b2-)"
  elif [ "$type" = 0 ];then
    printf "title: %s\n" "$(curl -gs "$uri" | head -n1)"
  elif [ "$type" = "h" ];then
    printf "title: %s\n" "$(curl -gs "$uri" | head -c 1000000 | tr -d '\n' | tr '<' '\n' | grep -iA 10 '^title' | grep -iB 10 '^\/title>' | cut '-d>' -f2 | tr '\t' ' ' | sed 's/^ *//g' | sed 's/ *$//g' | grep .)"
  else
    printf "title: don't know how to get title of non-1 gopher links"
  fi
  ;;
urn)
  nid="$(printf "%s\n" "$uri" | cut -d: -f2)"
  nss="$(printf "%s\n" "$uri" | cut -d: -f3)"
  nss2="$(printf "%s\n" "$uri" | cut -d: -f4)"
  if [ "$nid" = "ietf" ];then
    if [ "$nss" = "rfc" ];then
      cat "/var/db/rfc/rfc${nss2}.json" | jq .title
      exit 0
    fi
  fi
  if [ "$nid" = "phrack" ];then
    if [ ! "${nss2}" ];then
      printf "issue %s of phrack has %d articles\n" "${nss}" "$(grep -c '.' /var/db/phrack/meta/${nss}.tsv)"
      exit 0
    fi
    cat /var/db/phrack/meta/${nss}.tsv | grep "^${nss}/${nss2}"$'\t' | sed 's|/|\t|' | tr '\n\t' '\0\0' | xargs -n4 -0 printf 'issue %2s article %2s %79s by [%s]\n'
    exit 0
  fi
  if [ "$nid" = "mitre" ];then
    if [ "$nss" = "cve" ];then
      urititle "$(urnresolve "$uri")"
      exit 0
    fi
  fi
  urnresolve "$uri"
  ;;
ssh)
  if [ ! "$port" ];then
    port=22
  fi
  printf "title: %s\n" "$(printf "" | nc "$domain" "$port" | head -n1)"
  ;;
file)
  file="$(uriunescape "$path")"
  if [ -e "$file" ];then
    mime-type -v "$file"
  else
    printf "file '%s' doesn't exist.\n" "$file"
  fi
  ;;
*)
  printf "DONT KNOW HOW TO GET TITLE FOR THIS URL: %s\n" "$uri"
esac