more (web http) exports for parsing request components
[bpt/guile.git] / module / web / request.scm
CommitLineData
ad05d4e8
AW
1;;; HTTP request objects
2
3;; Copyright (C) 2010 Free Software Foundation, Inc.
4
5;; This library is free software; you can redistribute it and/or
6;; modify it under the terms of the GNU Lesser General Public
7;; License as published by the Free Software Foundation; either
8;; version 3 of the License, or (at your option) any later version.
9;;
10;; This library is distributed in the hope that it will be useful,
11;; but WITHOUT ANY WARRANTY; without even the implied warranty of
12;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13;; Lesser General Public License for more details.
14;;
15;; You should have received a copy of the GNU Lesser General Public
16;; License along with this library; if not, write to the Free Software
17;; Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
18;; 02110-1301 USA
19
20;;; Code:
21
22(define-module (web request)
23 #:use-module (rnrs bytevectors)
24 #:use-module (rnrs io ports)
25 #:use-module (ice-9 rdelim)
26 #:use-module (srfi srfi-9)
27 #:use-module (web uri)
28 #:use-module (web http)
29 #:export (request?
30 request-method
31 request-uri
32 request-version
33 request-headers
34 request-port
35
36 read-request
37 build-request
38 write-request
39
40 read-request-body/latin-1
41 write-request-body/latin-1
42
43 read-request-body/bytevector
44 write-request-body/bytevector
45
46 ;; General headers
47 ;;
48 request-cache-control
49 request-connection
50 request-date
51 request-pragma
52 request-trailer
53 request-transfer-encoding
54 request-upgrade
55 request-via
56 request-warning
57
58 ;; Entity headers
59 ;;
60 request-allow
61 request-content-encoding
62 request-content-language
63 request-content-length
64 request-content-location
65 request-content-md5
66 request-content-range
67 request-content-type
68 request-expires
69 request-last-modified
70
71 ;; Request headers
72 ;;
73 request-accept
74 request-accept-charset
75 request-accept-encoding
76 request-accept-language
77 request-authorization
78 request-expect
79 request-from
80 request-host
81 request-if-match
82 request-if-modified-since
83 request-if-none-match
84 request-if-range
85 request-if-unmodified-since
86 request-max-forwards
87 request-proxy-authorization
88 request-range
89 request-referer
90 request-te
91 request-user-agent
92
93 ;; Misc
94 request-absolute-uri))
95
96
97;;; {Character Encodings, Strings, and Bytevectors}
98;;;
99;;; Requests are read from over the wire, and as such have to be treated
100;;; very carefully.
101;;;
102;;; The header portion of the message is defined to be in a subset of
103;;; ASCII, and may be processed either byte-wise (using bytevectors and
104;;; binary I/O) or as characters in a single-byte ASCII-compatible
105;;; encoding.
106;;;
107;;; We choose the latter, processing as strings in the latin-1
108;;; encoding. This allows us to use all the read-delimited machinery,
109;;; character sets, and regular expressions, shared substrings, etc.
110;;;
111;;; The characters in the header values may themselves encode other
112;;; bytes or characters -- basically each header has its own parser. We
113;;; leave that as a header-specific topic.
114;;;
115;;; The body is present if the content-length header is present. Its
116;;; format and, if textual, encoding is determined by the headers, but
117;;; its length is encoded in bytes. So we just slurp that number of
118;;; characters in latin-1, knowing that the number of characters
119;;; corresponds to the number of bytes, and then convert to a
120;;; bytevector, perhaps for later decoding.
121;;;
122
123(define-record-type <request>
124 (make-request method uri version headers port)
125 request?
126 (method request-method)
127 (uri request-uri)
128 (version request-version)
129 (headers request-headers)
130 (port request-port))
131
132(define (bad-request message . args)
133 (throw 'bad-request message args))
134
135(define (non-negative-integer? n)
136 (and (number? n) (>= n 0) (exact? n) (integer? n)))
137
138(define (validate-headers headers)
139 (if (pair? headers)
140 (let ((h (car headers)))
141 (if (pair? h)
142 (let ((k (car h)) (v (cdr h)))
143 (if (symbol? k)
144 (if (not (valid-header? k v))
145 (bad-request "Bad value for header ~a: ~s" k v))
146 (if (not (and (string? k) (string? v)))
147 (bad-request "Unknown header not a pair of strings: ~s"
148 h)))
149 (validate-headers (cdr headers)))
150 (bad-request "Header not a pair: ~a" h)))
151 (if (not (null? headers))
152 (bad-request "Headers not a list: ~a" headers))))
153
154(define* (build-request #:key (method 'GET) uri (version '(1 . 1))
155 (headers '()) port (validate-headers? #t))
156 (cond
157 ((not (and (pair? version)
158 (non-negative-integer? (car version))
159 (non-negative-integer? (cdr version))))
160 (bad-request "Bad version: ~a" version))
161 ((not (uri? uri))
162 (bad-request "Bad uri: ~a" uri))
163 ((and (not port) (memq method '(POST PUT)))
164 (bad-request "Missing port for message ~a" method))
165 (else
166 (if validate-headers?
167 (validate-headers headers))))
168 (make-request method uri version headers port))
169
170(define (read-request port)
171 (set-port-encoding! port "ISO-8859-1")
172 (call-with-values (lambda () (read-request-line port))
173 (lambda (method uri version)
174 (make-request method uri version (read-headers port) port))))
175
176(define (write-request r port)
177 (write-request-line (request-method r) (request-uri r)
178 (request-version r) port)
179 (write-headers (request-headers r) port)
180 (display "\r\n" port)
181 (if (eq? port (request-port r))
182 r
183 (make-request (request-method r) (request-uri r) (request-version r)
184 (request-headers r) port)))
185
186;; Probably not what you want to use "in production". Relies on one byte
187;; per char because we are in latin-1 encoding.
188;;
189(define (read-request-body/latin-1 r)
190 (let ((nbytes (request-content-length r)))
191 (and nbytes
192 (let ((buf (make-string nbytes)))
193 (read-delimited! "" buf (request-port r))
194 buf))))
195
196;; Likewise, assumes that body can be written in the latin-1 encoding,
197;; and that the latin-1 encoding is what is expected by the server.
198;;
199(define (write-request-body/latin-1 r body)
200 (display body (request-port r)))
201
202(define (read-request-body/bytevector r)
203 (let ((nbytes (request-content-length r)))
204 (and nbytes
205 (let ((bv (get-bytevector-n (request-port r) nbytes)))
206 (if (= (bytevector-length bv) nbytes)
207 bv
208 (bad-request "EOF while reading request body: ~a bytes of ~a"
209 (bytevector-length bv) nbytes))))))
210
211(define (write-request-body/bytevector r bv)
212 (put-bytevector (request-port r) bv))
213
214(define-syntax define-request-accessor
215 (lambda (x)
216 (syntax-case x ()
217 ((_ field)
218 #'(define-request-accessor field #f))
219 ((_ field def) (identifier? #'field)
220 #`(define* (#,(datum->syntax
221 #'field
222 (symbol-append 'request- (syntax->datum #'field)))
223 request
224 #:optional (default def))
225 (cond
226 ((assq 'field (request-headers request)) => cdr)
227 (else default)))))))
228
229;; General headers
230;;
231(define-request-accessor cache-control '())
232(define-request-accessor connection '())
233(define-request-accessor date #f)
234(define-request-accessor pragma '())
235(define-request-accessor trailer '())
236(define-request-accessor transfer-encoding '())
237(define-request-accessor upgrade '())
238(define-request-accessor via '())
239(define-request-accessor warning '())
240
241;; Entity headers
242;;
243(define-request-accessor allow '())
244(define-request-accessor content-encoding '())
245(define-request-accessor content-language '())
246(define-request-accessor content-length #f)
247(define-request-accessor content-location #f)
248(define-request-accessor content-md5 #f)
249(define-request-accessor content-range #f)
250(define-request-accessor content-type #f)
251(define-request-accessor expires #f)
252(define-request-accessor last-modified #f)
253
254;; Request headers
255;;
256(define-request-accessor accept '())
257(define-request-accessor accept-charset '())
258(define-request-accessor accept-encoding '())
259(define-request-accessor accept-language '())
260(define-request-accessor authorization #f)
261(define-request-accessor expect '())
262(define-request-accessor from #f)
263(define-request-accessor host #f)
264;; Absence of an if-directive appears to be different from `*'.
265(define-request-accessor if-match #f)
266(define-request-accessor if-modified-since #f)
267(define-request-accessor if-none-match #f)
268(define-request-accessor if-range #f)
269(define-request-accessor if-unmodified-since #f)
270(define-request-accessor max-forwards #f)
271(define-request-accessor proxy-authorization #f)
272(define-request-accessor range #f)
273(define-request-accessor referer #f)
274(define-request-accessor te '())
275(define-request-accessor user-agent #f)
276
277;; Misc accessors
278(define* (request-absolute-uri r #:optional default-host default-port)
279 (let ((uri (request-uri r)))
280 (if (uri-host uri)
281 uri
282 (let ((host
283 (or (request-host r)
284 (if default-host
285 (cons default-host default-port)
286 (bad-request
287 "URI not absolute, no Host header, and no default: ~s"
288 uri)))))
289 (build-uri (uri-scheme uri)
290 #:host (car host)
291 #:port (cdr host)
292 #:path (uri-path uri)
293 #:query (uri-query uri)
294 #:fragment (uri-fragment uri))))))