comparison src/http.lua @ 0:4b915342e2a8

LuaSocket 2.0.2 + CMake build description.
author Eric Wing <ewing . public |-at-| gmail . com>
date Tue, 26 Aug 2008 18:40:01 -0700
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:4b915342e2a8
1 -----------------------------------------------------------------------------
2 -- HTTP/1.1 client support for the Lua language.
3 -- LuaSocket toolkit.
4 -- Author: Diego Nehab
5 -- RCS ID: $Id: http.lua,v 1.71 2007/10/13 23:55:20 diego Exp $
6 -----------------------------------------------------------------------------
7
8 -----------------------------------------------------------------------------
9 -- Declare module and import dependencies
10 -------------------------------------------------------------------------------
11 local socket = require("socket")
12 local url = require("socket.url")
13 local ltn12 = require("ltn12")
14 local mime = require("mime")
15 local string = require("string")
16 local base = _G
17 local table = require("table")
18 module("socket.http")
19
20 -----------------------------------------------------------------------------
21 -- Program constants
22 -----------------------------------------------------------------------------
23 -- connection timeout in seconds
24 TIMEOUT = 60
25 -- default port for document retrieval
26 PORT = 80
27 -- user agent field sent in request
28 USERAGENT = socket._VERSION
29
30 -----------------------------------------------------------------------------
31 -- Reads MIME headers from a connection, unfolding where needed
32 -----------------------------------------------------------------------------
33 local function receiveheaders(sock, headers)
34 local line, name, value, err
35 headers = headers or {}
36 -- get first line
37 line, err = sock:receive()
38 if err then return nil, err end
39 -- headers go until a blank line is found
40 while line ~= "" do
41 -- get field-name and value
42 name, value = socket.skip(2, string.find(line, "^(.-):%s*(.*)"))
43 if not (name and value) then return nil, "malformed reponse headers" end
44 name = string.lower(name)
45 -- get next line (value might be folded)
46 line, err = sock:receive()
47 if err then return nil, err end
48 -- unfold any folded values
49 while string.find(line, "^%s") do
50 value = value .. line
51 line = sock:receive()
52 if err then return nil, err end
53 end
54 -- save pair in table
55 if headers[name] then headers[name] = headers[name] .. ", " .. value
56 else headers[name] = value end
57 end
58 return headers
59 end
60
61 -----------------------------------------------------------------------------
62 -- Extra sources and sinks
63 -----------------------------------------------------------------------------
64 socket.sourcet["http-chunked"] = function(sock, headers)
65 return base.setmetatable({
66 getfd = function() return sock:getfd() end,
67 dirty = function() return sock:dirty() end
68 }, {
69 __call = function()
70 -- get chunk size, skip extention
71 local line, err = sock:receive()
72 if err then return nil, err end
73 local size = base.tonumber(string.gsub(line, ";.*", ""), 16)
74 if not size then return nil, "invalid chunk size" end
75 -- was it the last chunk?
76 if size > 0 then
77 -- if not, get chunk and skip terminating CRLF
78 local chunk, err, part = sock:receive(size)
79 if chunk then sock:receive() end
80 return chunk, err
81 else
82 -- if it was, read trailers into headers table
83 headers, err = receiveheaders(sock, headers)
84 if not headers then return nil, err end
85 end
86 end
87 })
88 end
89
90 socket.sinkt["http-chunked"] = function(sock)
91 return base.setmetatable({
92 getfd = function() return sock:getfd() end,
93 dirty = function() return sock:dirty() end
94 }, {
95 __call = function(self, chunk, err)
96 if not chunk then return sock:send("0\r\n\r\n") end
97 local size = string.format("%X\r\n", string.len(chunk))
98 return sock:send(size .. chunk .. "\r\n")
99 end
100 })
101 end
102
103 -----------------------------------------------------------------------------
104 -- Low level HTTP API
105 -----------------------------------------------------------------------------
106 local metat = { __index = {} }
107
108 function open(host, port, create)
109 -- create socket with user connect function, or with default
110 local c = socket.try((create or socket.tcp)())
111 local h = base.setmetatable({ c = c }, metat)
112 -- create finalized try
113 h.try = socket.newtry(function() h:close() end)
114 -- set timeout before connecting
115 h.try(c:settimeout(TIMEOUT))
116 h.try(c:connect(host, port or PORT))
117 -- here everything worked
118 return h
119 end
120
121 function metat.__index:sendrequestline(method, uri)
122 local reqline = string.format("%s %s HTTP/1.1\r\n", method or "GET", uri)
123 return self.try(self.c:send(reqline))
124 end
125
126 function metat.__index:sendheaders(headers)
127 local h = "\r\n"
128 for i, v in base.pairs(headers) do
129 h = i .. ": " .. v .. "\r\n" .. h
130 end
131 self.try(self.c:send(h))
132 return 1
133 end
134
135 function metat.__index:sendbody(headers, source, step)
136 source = source or ltn12.source.empty()
137 step = step or ltn12.pump.step
138 -- if we don't know the size in advance, send chunked and hope for the best
139 local mode = "http-chunked"
140 if headers["content-length"] then mode = "keep-open" end
141 return self.try(ltn12.pump.all(source, socket.sink(mode, self.c), step))
142 end
143
144 function metat.__index:receivestatusline()
145 local status = self.try(self.c:receive(5))
146 -- identify HTTP/0.9 responses, which do not contain a status line
147 -- this is just a heuristic, but is what the RFC recommends
148 if status ~= "HTTP/" then return nil, status end
149 -- otherwise proceed reading a status line
150 status = self.try(self.c:receive("*l", status))
151 local code = socket.skip(2, string.find(status, "HTTP/%d*%.%d* (%d%d%d)"))
152 return self.try(base.tonumber(code), status)
153 end
154
155 function metat.__index:receiveheaders()
156 return self.try(receiveheaders(self.c))
157 end
158
159 function metat.__index:receivebody(headers, sink, step)
160 sink = sink or ltn12.sink.null()
161 step = step or ltn12.pump.step
162 local length = base.tonumber(headers["content-length"])
163 local t = headers["transfer-encoding"] -- shortcut
164 local mode = "default" -- connection close
165 if t and t ~= "identity" then mode = "http-chunked"
166 elseif base.tonumber(headers["content-length"]) then mode = "by-length" end
167 return self.try(ltn12.pump.all(socket.source(mode, self.c, length),
168 sink, step))
169 end
170
171 function metat.__index:receive09body(status, sink, step)
172 local source = ltn12.source.rewind(socket.source("until-closed", self.c))
173 source(status)
174 return self.try(ltn12.pump.all(source, sink, step))
175 end
176
177 function metat.__index:close()
178 return self.c:close()
179 end
180
181 -----------------------------------------------------------------------------
182 -- High level HTTP API
183 -----------------------------------------------------------------------------
184 local function adjusturi(reqt)
185 local u = reqt
186 -- if there is a proxy, we need the full url. otherwise, just a part.
187 if not reqt.proxy and not PROXY then
188 u = {
189 path = socket.try(reqt.path, "invalid path 'nil'"),
190 params = reqt.params,
191 query = reqt.query,
192 fragment = reqt.fragment
193 }
194 end
195 return url.build(u)
196 end
197
198 local function adjustproxy(reqt)
199 local proxy = reqt.proxy or PROXY
200 if proxy then
201 proxy = url.parse(proxy)
202 return proxy.host, proxy.port or 3128
203 else
204 return reqt.host, reqt.port
205 end
206 end
207
208 local function adjustheaders(reqt)
209 -- default headers
210 local lower = {
211 ["user-agent"] = USERAGENT,
212 ["host"] = reqt.host,
213 ["connection"] = "close, TE",
214 ["te"] = "trailers"
215 }
216 -- if we have authentication information, pass it along
217 if reqt.user and reqt.password then
218 lower["authorization"] =
219 "Basic " .. (mime.b64(reqt.user .. ":" .. reqt.password))
220 end
221 -- override with user headers
222 for i,v in base.pairs(reqt.headers or lower) do
223 lower[string.lower(i)] = v
224 end
225 return lower
226 end
227
228 -- default url parts
229 local default = {
230 host = "",
231 port = PORT,
232 path ="/",
233 scheme = "http"
234 }
235
236 local function adjustrequest(reqt)
237 -- parse url if provided
238 local nreqt = reqt.url and url.parse(reqt.url, default) or {}
239 -- explicit components override url
240 for i,v in base.pairs(reqt) do nreqt[i] = v end
241 if nreqt.port == "" then nreqt.port = 80 end
242 socket.try(nreqt.host and nreqt.host ~= "",
243 "invalid host '" .. base.tostring(nreqt.host) .. "'")
244 -- compute uri if user hasn't overriden
245 nreqt.uri = reqt.uri or adjusturi(nreqt)
246 -- ajust host and port if there is a proxy
247 nreqt.host, nreqt.port = adjustproxy(nreqt)
248 -- adjust headers in request
249 nreqt.headers = adjustheaders(nreqt)
250 return nreqt
251 end
252
253 local function shouldredirect(reqt, code, headers)
254 return headers.location and
255 string.gsub(headers.location, "%s", "") ~= "" and
256 (reqt.redirect ~= false) and
257 (code == 301 or code == 302) and
258 (not reqt.method or reqt.method == "GET" or reqt.method == "HEAD")
259 and (not reqt.nredirects or reqt.nredirects < 5)
260 end
261
262 local function shouldreceivebody(reqt, code)
263 if reqt.method == "HEAD" then return nil end
264 if code == 204 or code == 304 then return nil end
265 if code >= 100 and code < 200 then return nil end
266 return 1
267 end
268
269 -- forward declarations
270 local trequest, tredirect
271
272 function tredirect(reqt, location)
273 local result, code, headers, status = trequest {
274 -- the RFC says the redirect URL has to be absolute, but some
275 -- servers do not respect that
276 url = url.absolute(reqt.url, location),
277 source = reqt.source,
278 sink = reqt.sink,
279 headers = reqt.headers,
280 proxy = reqt.proxy,
281 nredirects = (reqt.nredirects or 0) + 1,
282 create = reqt.create
283 }
284 -- pass location header back as a hint we redirected
285 headers = headers or {}
286 headers.location = headers.location or location
287 return result, code, headers, status
288 end
289
290 function trequest(reqt)
291 -- we loop until we get what we want, or
292 -- until we are sure there is no way to get it
293 local nreqt = adjustrequest(reqt)
294 local h = open(nreqt.host, nreqt.port, nreqt.create)
295 -- send request line and headers
296 h:sendrequestline(nreqt.method, nreqt.uri)
297 h:sendheaders(nreqt.headers)
298 -- if there is a body, send it
299 if nreqt.source then
300 h:sendbody(nreqt.headers, nreqt.source, nreqt.step)
301 end
302 local code, status = h:receivestatusline()
303 -- if it is an HTTP/0.9 server, simply get the body and we are done
304 if not code then
305 h:receive09body(status, nreqt.sink, nreqt.step)
306 return 1, 200
307 end
308 local headers
309 -- ignore any 100-continue messages
310 while code == 100 do
311 headers = h:receiveheaders()
312 code, status = h:receivestatusline()
313 end
314 headers = h:receiveheaders()
315 -- at this point we should have a honest reply from the server
316 -- we can't redirect if we already used the source, so we report the error
317 if shouldredirect(nreqt, code, headers) and not nreqt.source then
318 h:close()
319 return tredirect(reqt, headers.location)
320 end
321 -- here we are finally done
322 if shouldreceivebody(nreqt, code) then
323 h:receivebody(headers, nreqt.sink, nreqt.step)
324 end
325 h:close()
326 return 1, code, headers, status
327 end
328
329 local function srequest(u, b)
330 local t = {}
331 local reqt = {
332 url = u,
333 sink = ltn12.sink.table(t)
334 }
335 if b then
336 reqt.source = ltn12.source.string(b)
337 reqt.headers = {
338 ["content-length"] = string.len(b),
339 ["content-type"] = "application/x-www-form-urlencoded"
340 }
341 reqt.method = "POST"
342 end
343 local code, headers, status = socket.skip(1, trequest(reqt))
344 return table.concat(t), code, headers, status
345 end
346
347 request = socket.protect(function(reqt, body)
348 if base.type(reqt) == "string" then return srequest(reqt, body)
349 else return trequest(reqt) end
350 end)