--****

-- == HTTP

--

-- <>

namespace http

include std/socket.e as sock

include std/net/url.e

include std/net/dns.e

include std/text.e

include std/get.e as val

include euphoria/info.e

--****

-- === Constants

public constant

sequence

integer slen, spt, flag

sequence parsed, upperc, uppers

end if

else

end if

else

end if

end if

end function

--****

-- === Header management

--

--**

-- Retrieve either the whole sendheader sequence, or just a single

-- field.

--

-- Parameters:

-- # ##field## : an object indicating which part is being requested, see Comments section.

--

-- Returns:

-- An **object**, either:

-- * -1 if the field cannot be found,

-- * ##{{"label","delimiter","value"},...}## for the whole sendheader sequence

-- * a three-element sequence in the form ##{"label","delimiter","value"}## when only a single field is selected.

--

-- Comments:

-- ##field## can be either an HTTP_HEADER_xxx access constant,

-- the number 0 to retrieve the whole sendheader sequence, or

-- a string matching one of the header field labels. The string is

-- not case sensitive.

--

-- if field is 0, return the whole sequence.

-- if field is 1..length(sendheader), return just that field

-- if field is invalid, return -1. -- no, this is a problem , some code needs a sequence

-- if field is a sequence, try to match it to sendheader[x][1].

-- Kat: should i return [1] & [2] as well? Mike: yes

-- most server interfaces return the only value, saves parsing

-- we'll return a {"Name","spacer","value"} format

sequence upperfield

end if

else

end if

end function

--**

-- Sets header elements to default values. The default User Agent

-- is Opera (currently the most standards compliant). Before setting

-- any header option individually, programs must call this procedure.

--

-- See Also:

-- [[:get_sendheader]], [[:set_sendheader]], [[:set_sendheader_useragent_msie]]

-- this sets some defaults, if not previously set to something by the user

-- if a header line was previously set by the user, do not change it here

-- httpversion MUST come before GET in this program: some servers default to 1.0, even if you say 1.1

-- NO spaces around [3] on httpversion

-- POSTDATA MUST come before Content-Length in this program

-- Referer is often used by sites to be sure your fetch was from one of their own pages

-- headers with [3] = "" won't be sent

-- you can add more [1], and modify [3], [2] is the ' ' or ": " (GET and POST have no ": ")

{"httpversion","","HTTP/1.0"}, -- not a legal http headerline, but to append to GET or POST later on

{"GET"," ",""}, -- [3] = the filename you want

{"POST"," ",""}, -- [3] = the filename you want

{"Host",": ",""}, -- the domain. You might think this was obvious, but for vhosting sites it's necessary.

{"Referer",": ",""}, -- i know it's misspelled, but that's official! , the site that sent you to this one

{"User-Agent",": ", whoami},

{"Accept-Charset",": ","utf-8,ISO-8859-1;q=0.7,*;q=0.5"},

{"Accept-Ranges",": ",""},

{"Authorization",": ",""},

{"Date",": ",""}, -- who cares if the server has my time? Except for cookie timeouts, that is.

{"If-Modified-Since",": ",""}, -- for keeping a local cache sync'd

{"POSTDATA","",""}, -- not a legal headerline, but has to go somewhere; put the POST data here, it will be appended to the bottom later

{"Content-Type",": ",""}, -- if POST or PUT transaction

{"Content-Length",": ",""}, -- if POST or PUT transaction

{"From",": ",""}, -- possible in POST or PUT or Authorization

{"Keep-Alive",": ",""}, -- set value depending on Connection

{"Cache-Control",": ",""},

{"Connection",": ","close"} -- this is usually "close", sometimes "keep-alive" for http/1.1 and SSL, even if you set "close" the server may ignore you

}

-- the following not only puts the default header lines,

-- it sorts the already-set lines to match the defaultsendheader order

else

end if

--**

-- Set an individual header field.

--

-- Parameters:

-- # ##whatheader## : an object, either an explicit name string or a HTTP_HEADER_xxx constant

-- # ##whatdata## : a string, the associated data

--

-- Comments:

-- If the requested field is not one of the default header fields,

-- the field MUST be set by string. This will increase the length

-- of the header overall.

--

-- Example 1:

--

-- set_sendheader("Referer","search.yahoo.com")

--

--

-- See Also:

-- [[:get_sendheader]]

end if

end if

-- is this whatheader already in sendheader?

-- then simply set it to this value

end if

-- ok, if we got here, then whatheader isn't in sendheader

-- you better know what you are doing here!

-- ": " is supplied as default, lets hope it's not an aberration like GET or POST

-- this doesn't put it in any correct order

--**

-- Inform listener that user agent is Microsoft (R) Internet Explorer (TM).

--

-- Comments:

-- This is a convenience procedure to tell a website that a Microsoft

-- Internet Explorer (TM) browser is requesting data. Because some

-- websites format their response differently (or simply refuse data)

-- depending on the browser, this procedure provides a quick means

-- around that.

-- For example, see:

-- http://www.missporters.org/podium/nonsupport.aspx

--------------------------------------------------------------------------------

-- this can also be used to flatten the sendheader record, for printing, etc

sequence

not equal("POSTDATA",sendheader[idx][1])

then

-- if the data field is not empty...

case "GET" then

-- remove any trailing nulls

-- append the http version

sendheader[idx][3] & " " & httpversion & "\r\n"

case "POST" then

-- remove any trailing nulls

-- append the http version

sendheader[idx][3] & " " & httpversion & "\r\n"

case else

-- remove any trailing nulls

-- else just flatten the sequence

sendheader[idx][3] & "\r\n"

end switch

end if

end if

-- this is done here because

-- this is where the POSTDATA is moved to the bottom of the header,

-- the Content-length and Content-Type is filled in

-- POSTDATA was set to something

end if

end if

end if

end function

--**

-- Populates the internal sequence recvheader from the flat string header.

--

-- Parameters:

-- # ##header## : a string, the header data

--

-- Comments:

-- This must be called prior to calling [[:get_recvheader]]().

sequence junk

atom place

else

end if

end if

--**

-- Return the value of a named field in the received http header as

-- returned by the most recent call to [[:get_http]].

--

-- Parameters:

-- # ##field## : an object, either a string holding a field name (case insensitive),

-- 0 to return the whole header, or a numerical index.

--

-- Returns:

-- An **object**,

-- * -1 on error

-- * a sequence in the form, ##{field name, field value}## on success.

sequence upperfield

-- recvheader was parsed out previously in parse_recvheader()

-- if field is 0, return the whole sequence.

-- if field is 1..length(recvheader), return just that field

-- if field is invalid, return -1.

-- if field is a sequence, try to match it to recvheader[x][1].

-- we'll NOT return a {"Name","value"} format

-- because that leads to using a junk seq to get the [2] from

-- --> And yet, that's exactly what we're doing. -- Mike.

end if

end if

end if

end if

end function

--****

-- === Web interface

--

--**

-- Returns data from an http internet site.

--

-- Parameters:

-- # ##inet_addr## : a sequence holding an address

-- # ##hostname## : a string, the name for the host

-- # ##file## : a file name to transmit

--

-- Returns:

-- A **sequence**, empty sequence on error, of length 2 on success,

-- like ##{sequence header, sequence data}##.

object junk, junk2, header

sock:socket sock

atom success, last_data_len, gotheader, contentlen, last

sequence data

-- Notes for future additions:

-- HUGE differences in HTTP/1.1 vs HTTP/1.0

-- GET /index.html HTTP/1.1

end if

end if

-- was the POSTDATA set?

-- if no, assume it's a GET

else

-- if so, then it's definitely a POST (err, or PUT, but we don't do PUT)

end if

-- This is required for virtual shared hosting. On dedicated boxes on fixed ip,

-- you can http to the ip, and GET/POST is enough to deal with it. Setting it is

-- safe, either way.

-- eunet_format_sendheader sets up the header to sent,

-- putting the POST data at the end,

-- filling in the CONTENT-LENGTH,

-- and avoiding sending empty fields for any field

-- } end version 1.3.0 mod

then

end if

end if

end if

entry

-- Do we have readable data?

else

-- assume server has hung, abort

end if

else

end if -- if success -- sock:send

else

end if -- if success = 1 then -- sock:connect

-- clear any POSTDATA

end function

--**

-- Works the same as [[:get_url]](), but maintains an internal

-- state register based on cookies received.

--

-- Warning:

-- This function is not yet implemented.

--

-- Parameters:

-- # ##inet_addr## : a sequence holding an address

-- # ##hostname## : a string, the name for the host

-- # ##file## : a file name to transmit

--

-- Returns:

-- A **sequence**, {header, body} on success, or an empty sequence on error.

--

-- Example 1:

--

-- addrinfo = getaddrinfo("www.yahoo.com","http",0)

-- if atom(addrinfo) or length(addrinfo) < 1 or

-- length(addrinfo[1]) < 5 then

-- puts(1,"Uh, oh")

-- return {}

-- else

-- inet_addr = addrinfo[1][5]

-- end if

-- data = get_http_use_cookie(inet_addr,"www.yahoo.com","")

--

--

-- See also:

-- [[:get_url]]

/*

atom socket, success, last_data_len, cpos, offset

sequence header, header2, body, data, updata

sequence cookielist, request, cookie

object junk -- a general throwaway temp var

cookielist = {}

request = ""

-- cookie = {name,domain,path,expires,encrypted,version}

if length(inet_addr)=0 then

return {"",""}

end if

if length(file)=0 or file[1]!='/' then file = '/'&file end if

-- was the POSTDATA set?

junk = get_sendheader("POSTDATA")

if equal(junk[3],"") then

-- if no, assume it's a GET

set_sendheader("GET",file)

else

-- if so, then it's definitely a POST (err, or PUT, but we don't do PUT)

set_sendheader("POST",file)

end if

-- This is required for virtual shared hosting.

-- On dedicated boxes on fixed ip,

-- you can http to the ip, and GET/POST is enough to deal with it.

-- Setting it is safe, either way.

set_sendheader("HOST",hostname)

for ctr = 1 to length(this_cookiejar) do

if sequence(this_cookiejar[ctr]) and length(this_cookiejar[ctr])>=2 and

sequence(this_cookiejar[ctr][1]) and

(match(hostname,this_cookiejar[ctr][2])>0 or match(this_cookiejar[ctr][2],hostname)>0) and

(length(file)=0 or match(this_cookiejar[ctr][3],file)>0)

then

cookielist = append(cookielist,this_cookiejar[ctr])

end if

end for

-- TO DO: Sort cookielist by domain, path (longer path before shorter path)

-- request = sprintf("GET /%s HTTP/1.0\nHost: %s\n",{file,hostname})

for idx = 1 to length(cookielist) do

-- if idx = 1 then

-- request = request & "Cookie: "&cookielist[idx][1]

-- else

-- request = request & " "&cookielist[idx][1]

-- end if

request = request & cookielist[idx][1]

if length(cookielist[idx][3])>0 then

request = request & "; $Path=" & cookielist[idx][3]

end if

if idx < length(cookielist) then

--request = request & ";\n"

request = request & ";"

else

--request = request & "\n"

end if

end for

-- request = request & "\n"

set_sendheader("Cookie",request)

data = {}

last_data_len = 0

socket = sock:create(AF_INET,SOCK_STREAM,0)

success = sock:connect(AF_INET, socket,inet_addr)

if success = 0 then

-- success = sock:send(socket,request,0)

success = sock:send(socket,eunet_format_sendheader(),0)

-- } end version 1.3.0 modification

if success > 0 then

junk = sock:receive(sock, 0)

while sequence(junk) do

data = data & junk

if gotheader and equal(contentlen,length(data)) then

exit -- we got all the server said it had

end if

if not gotheader and match({13,10,13,10},data) then -- we got the header in there

header = data[1..match({13,10,13,10},data)-1] -- split off the header

data = data[match({13,10,13,10},data)+4..$] -- and the data is what's left, we keep using data in the sock loop

parse_recvheader(header) -- sets up recvheader -- global var

junk = get_recvheader("Content-Length")

if not equal(junk,-1) then

junk = val:value(junk[2])

contentlen = junk[2]

if equal(contentlen,0) then exit end if -- there's no more

if equal(contentlen,length(data)) then exit end if -- there's no more

end if

gotheader = 1 -- we got what we came for here

end if

junk = sock:receive(sock, 0)

end while

end if

end if

if close_socket(socket) then end if

header2 = header

cpos = match("SET-COOKIE",upper(header2)) -- this should be using get_recvheader() etc

while cpos > 0 do

header2 = header2[cpos+10 .. $]

data = header2

cpos = find(':',data)

if cpos > 0 then

data = data[cpos+1..$]

end if

offset = 0

cpos = match(13&10,data)

while cpos > 1 and data[offset+cpos-1]=';' do

offset = offset + cpos + 2

cpos = match(13&10,data[offset..$])

end while

offset = offset + cpos - 1

data = data[1..offset]

updata = upper(data)

cookie = {"","","","","N",""}

offset = match("PATH=",updata)

if offset > 0 then

cpos = find(';',data[offset .. $])

if cpos = 0 then cpos = length(data)-offset+2 end if

cookie[3] = data[offset+5..offset+cpos-2]

end if

cpos = find(';',data)

if cpos = 0 then cpos = length(data)+1 end if

cookie[1] = _socket_trim(data[1..cpos-1])

if cpos > length(data) then

data = ""

updata = ""

else

data = data[cpos+1 .. $]

updata = updata[cpos+1..length(data)]

end if

offset = match("DOMAIN=",updata)

if offset > 0 then

cpos = find(';',data[offset .. $])

if cpos = 0 then cpos = length(data)-offset+2 end if

cookie[2] = data[offset+7..offset+cpos-2]

-- Offset is base 1. If the semicolon is in the first position, cpos

-- is also 1. Since we don't want to include the semicolon, we need

-- to subtract 1 for offset's base and 1 to go to the char before

-- cpos, thus the subtracting of two. In the case of end of string

-- (cpos = 0), we need to add those two back to compensate for the

-- different scenario (+offset-offset = 0 and +2-2 = 0, therefore

-- cpos = length(data), which is what we want).

end if

offset = match("EXPIRES=",updata)

if offset > 0 then

cpos = find(';',data[offset .. $])

if cpos = 0 then cpos = length(data)-offset+2 end if

cookie[4] = data[offset+8..offset+cpos-2]

end if

offset = match("VERSION=",updata)

if offset > 0 then

cpos = find(';',data[offset..$])

if cpos = 0 then cpos = length(data)-offset+2 end if

cookie[6] = data[offset+8..offset+cpos-2]

end if

offset = match("MAX-AGE=",updata)

if offset > 0 then

cpos = find(';',data[offset..$])

if cpos = 0 then cpos = length(data)-offset+2 end if

cookie[4] = data[offset+8..offset+cpos-2]

end if

offset = match("SECURE",updata)

if offset > 0 then

cookie[5] = "Y"

end if

cpos = find('=',cookie[1])

if cpos > 0 then

request = cookie[1][1..cpos]

else

request = "="

end if

cpos = 0

for ctr = 1 to length(this_cookiejar) do

if sequence(this_cookiejar[ctr]) and length(this_cookiejar[ctr])>=2 and

match(cookie[1],this_cookiejar[ctr][1])>0 and

eu:compare(cookie[2],this_cookiejar[ctr][2])=0 and

eu:compare(this_cookiejar[ctr][3],cookie[3])=0 then

this_cookiejar[ctr] = cookie

cpos = ctr

exit

end if

end for

if cpos = 0 then

this_cookiejar = append(this_cookiejar,cookie)

end if

cpos = match("SET-COOKIE",upper(header2))

end while

-- clear any POSTDATA

set_sendheader("POSTDATA", "")

set_sendheader("POST", "")

set_sendheader("GET", "")

set_sendheader("Content-Type", "")

set_sendheader("Content-Length", "0")

set_sendheader_default()

return {header,body}

*/

end function

--**

-- Returns data from an http internet site.

--

-- Parameters:

-- # ##url##: URL to access

-- # ##post_data##: Optional post data

--

-- Returns:

-- A **sequence** {header, body} on success, or an empty sequence on error.

--

-- Comments:

-- If ##post_data## is empty, then a normal GET request is done. If ##post_data## is non-empty

-- then ##get_url## will perform a POST request and supply ##post_data## during the request.

--

-- Example 1:

--

-- url = "http://banners.wunderground.com/weathersticker/mini" &

-- "Weather2_metric_cond/language/www/US/PA/Philadelphia.gif"

--

-- temp = get_url(url)

-- if length(temp)>=2 and length(temp[2])>0 then

-- tempfp = open(TEMPDIR&"current_weather.gif","wb")

-- puts(tempfp,temp[2])

-- close(tempfp)

-- end if

--

object addrinfo, url_data

end if

-- url didn't specify a port, so default to 80

end if

url_data[URL_PATH] & url_data[URL_QUERY_STRING], , url_data[URL_PORT])

end if

end function

-- set the lines in the "proper" order for sending, not that the defaults will get sent.