Commit | Line | Data |
---|---|---|
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 | |
d4b6200a | 34 | request-meta |
ad05d4e8 AW |
35 | request-port |
36 | ||
37 | read-request | |
38 | build-request | |
39 | write-request | |
40 | ||
41 | read-request-body/latin-1 | |
42 | write-request-body/latin-1 | |
43 | ||
44 | read-request-body/bytevector | |
45 | write-request-body/bytevector | |
46 | ||
47 | ;; General headers | |
48 | ;; | |
49 | request-cache-control | |
50 | request-connection | |
51 | request-date | |
52 | request-pragma | |
53 | request-trailer | |
54 | request-transfer-encoding | |
55 | request-upgrade | |
56 | request-via | |
57 | request-warning | |
58 | ||
59 | ;; Entity headers | |
60 | ;; | |
61 | request-allow | |
62 | request-content-encoding | |
63 | request-content-language | |
64 | request-content-length | |
65 | request-content-location | |
66 | request-content-md5 | |
67 | request-content-range | |
68 | request-content-type | |
69 | request-expires | |
70 | request-last-modified | |
71 | ||
72 | ;; Request headers | |
73 | ;; | |
74 | request-accept | |
75 | request-accept-charset | |
76 | request-accept-encoding | |
77 | request-accept-language | |
78 | request-authorization | |
79 | request-expect | |
80 | request-from | |
81 | request-host | |
82 | request-if-match | |
83 | request-if-modified-since | |
84 | request-if-none-match | |
85 | request-if-range | |
86 | request-if-unmodified-since | |
87 | request-max-forwards | |
88 | request-proxy-authorization | |
89 | request-range | |
90 | request-referer | |
91 | request-te | |
92 | request-user-agent | |
93 | ||
94 | ;; Misc | |
95 | request-absolute-uri)) | |
96 | ||
97 | ||
98 | ;;; {Character Encodings, Strings, and Bytevectors} | |
99 | ;;; | |
100 | ;;; Requests are read from over the wire, and as such have to be treated | |
101 | ;;; very carefully. | |
102 | ;;; | |
103 | ;;; The header portion of the message is defined to be in a subset of | |
104 | ;;; ASCII, and may be processed either byte-wise (using bytevectors and | |
105 | ;;; binary I/O) or as characters in a single-byte ASCII-compatible | |
106 | ;;; encoding. | |
107 | ;;; | |
108 | ;;; We choose the latter, processing as strings in the latin-1 | |
109 | ;;; encoding. This allows us to use all the read-delimited machinery, | |
110 | ;;; character sets, and regular expressions, shared substrings, etc. | |
111 | ;;; | |
112 | ;;; The characters in the header values may themselves encode other | |
113 | ;;; bytes or characters -- basically each header has its own parser. We | |
114 | ;;; leave that as a header-specific topic. | |
115 | ;;; | |
116 | ;;; The body is present if the content-length header is present. Its | |
117 | ;;; format and, if textual, encoding is determined by the headers, but | |
118 | ;;; its length is encoded in bytes. So we just slurp that number of | |
119 | ;;; characters in latin-1, knowing that the number of characters | |
120 | ;;; corresponds to the number of bytes, and then convert to a | |
121 | ;;; bytevector, perhaps for later decoding. | |
122 | ;;; | |
123 | ||
124 | (define-record-type <request> | |
d4b6200a | 125 | (make-request method uri version headers meta port) |
ad05d4e8 AW |
126 | request? |
127 | (method request-method) | |
128 | (uri request-uri) | |
129 | (version request-version) | |
130 | (headers request-headers) | |
d4b6200a | 131 | (meta request-meta) |
ad05d4e8 AW |
132 | (port request-port)) |
133 | ||
134 | (define (bad-request message . args) | |
135 | (throw 'bad-request message args)) | |
136 | ||
137 | (define (non-negative-integer? n) | |
138 | (and (number? n) (>= n 0) (exact? n) (integer? n))) | |
139 | ||
140 | (define (validate-headers headers) | |
141 | (if (pair? headers) | |
142 | (let ((h (car headers))) | |
143 | (if (pair? h) | |
144 | (let ((k (car h)) (v (cdr h))) | |
145 | (if (symbol? k) | |
146 | (if (not (valid-header? k v)) | |
147 | (bad-request "Bad value for header ~a: ~s" k v)) | |
148 | (if (not (and (string? k) (string? v))) | |
149 | (bad-request "Unknown header not a pair of strings: ~s" | |
150 | h))) | |
151 | (validate-headers (cdr headers))) | |
152 | (bad-request "Header not a pair: ~a" h))) | |
153 | (if (not (null? headers)) | |
154 | (bad-request "Headers not a list: ~a" headers)))) | |
155 | ||
156 | (define* (build-request #:key (method 'GET) uri (version '(1 . 1)) | |
d4b6200a AW |
157 | (headers '()) port (meta '()) |
158 | (validate-headers? #t)) | |
ad05d4e8 AW |
159 | (cond |
160 | ((not (and (pair? version) | |
161 | (non-negative-integer? (car version)) | |
162 | (non-negative-integer? (cdr version)))) | |
163 | (bad-request "Bad version: ~a" version)) | |
164 | ((not (uri? uri)) | |
165 | (bad-request "Bad uri: ~a" uri)) | |
166 | ((and (not port) (memq method '(POST PUT))) | |
167 | (bad-request "Missing port for message ~a" method)) | |
d4b6200a AW |
168 | ((not (list? meta)) |
169 | (bad-request "Bad metadata alist" meta)) | |
ad05d4e8 AW |
170 | (else |
171 | (if validate-headers? | |
172 | (validate-headers headers)))) | |
d4b6200a | 173 | (make-request method uri version headers meta port)) |
ad05d4e8 | 174 | |
d4b6200a | 175 | (define* (read-request port #:optional (meta '())) |
ad05d4e8 AW |
176 | (set-port-encoding! port "ISO-8859-1") |
177 | (call-with-values (lambda () (read-request-line port)) | |
178 | (lambda (method uri version) | |
d4b6200a | 179 | (make-request method uri version (read-headers port) meta port)))) |
ad05d4e8 | 180 | |
d4b6200a | 181 | ;; FIXME: really return a new request? |
ad05d4e8 AW |
182 | (define (write-request r port) |
183 | (write-request-line (request-method r) (request-uri r) | |
184 | (request-version r) port) | |
185 | (write-headers (request-headers r) port) | |
186 | (display "\r\n" port) | |
187 | (if (eq? port (request-port r)) | |
188 | r | |
189 | (make-request (request-method r) (request-uri r) (request-version r) | |
d4b6200a | 190 | (request-headers r) (request-meta r) port))) |
ad05d4e8 AW |
191 | |
192 | ;; Probably not what you want to use "in production". Relies on one byte | |
193 | ;; per char because we are in latin-1 encoding. | |
194 | ;; | |
195 | (define (read-request-body/latin-1 r) | |
196 | (let ((nbytes (request-content-length r))) | |
197 | (and nbytes | |
198 | (let ((buf (make-string nbytes))) | |
199 | (read-delimited! "" buf (request-port r)) | |
200 | buf)))) | |
201 | ||
202 | ;; Likewise, assumes that body can be written in the latin-1 encoding, | |
203 | ;; and that the latin-1 encoding is what is expected by the server. | |
204 | ;; | |
205 | (define (write-request-body/latin-1 r body) | |
206 | (display body (request-port r))) | |
207 | ||
208 | (define (read-request-body/bytevector r) | |
209 | (let ((nbytes (request-content-length r))) | |
210 | (and nbytes | |
211 | (let ((bv (get-bytevector-n (request-port r) nbytes))) | |
212 | (if (= (bytevector-length bv) nbytes) | |
213 | bv | |
214 | (bad-request "EOF while reading request body: ~a bytes of ~a" | |
215 | (bytevector-length bv) nbytes)))))) | |
216 | ||
217 | (define (write-request-body/bytevector r bv) | |
218 | (put-bytevector (request-port r) bv)) | |
219 | ||
220 | (define-syntax define-request-accessor | |
221 | (lambda (x) | |
222 | (syntax-case x () | |
223 | ((_ field) | |
224 | #'(define-request-accessor field #f)) | |
225 | ((_ field def) (identifier? #'field) | |
226 | #`(define* (#,(datum->syntax | |
227 | #'field | |
228 | (symbol-append 'request- (syntax->datum #'field))) | |
229 | request | |
230 | #:optional (default def)) | |
231 | (cond | |
232 | ((assq 'field (request-headers request)) => cdr) | |
233 | (else default))))))) | |
234 | ||
235 | ;; General headers | |
236 | ;; | |
237 | (define-request-accessor cache-control '()) | |
238 | (define-request-accessor connection '()) | |
239 | (define-request-accessor date #f) | |
240 | (define-request-accessor pragma '()) | |
241 | (define-request-accessor trailer '()) | |
242 | (define-request-accessor transfer-encoding '()) | |
243 | (define-request-accessor upgrade '()) | |
244 | (define-request-accessor via '()) | |
245 | (define-request-accessor warning '()) | |
246 | ||
247 | ;; Entity headers | |
248 | ;; | |
249 | (define-request-accessor allow '()) | |
250 | (define-request-accessor content-encoding '()) | |
251 | (define-request-accessor content-language '()) | |
252 | (define-request-accessor content-length #f) | |
253 | (define-request-accessor content-location #f) | |
254 | (define-request-accessor content-md5 #f) | |
255 | (define-request-accessor content-range #f) | |
256 | (define-request-accessor content-type #f) | |
257 | (define-request-accessor expires #f) | |
258 | (define-request-accessor last-modified #f) | |
259 | ||
260 | ;; Request headers | |
261 | ;; | |
262 | (define-request-accessor accept '()) | |
263 | (define-request-accessor accept-charset '()) | |
264 | (define-request-accessor accept-encoding '()) | |
265 | (define-request-accessor accept-language '()) | |
266 | (define-request-accessor authorization #f) | |
267 | (define-request-accessor expect '()) | |
268 | (define-request-accessor from #f) | |
269 | (define-request-accessor host #f) | |
270 | ;; Absence of an if-directive appears to be different from `*'. | |
271 | (define-request-accessor if-match #f) | |
272 | (define-request-accessor if-modified-since #f) | |
273 | (define-request-accessor if-none-match #f) | |
274 | (define-request-accessor if-range #f) | |
275 | (define-request-accessor if-unmodified-since #f) | |
276 | (define-request-accessor max-forwards #f) | |
277 | (define-request-accessor proxy-authorization #f) | |
278 | (define-request-accessor range #f) | |
279 | (define-request-accessor referer #f) | |
280 | (define-request-accessor te '()) | |
281 | (define-request-accessor user-agent #f) | |
282 | ||
283 | ;; Misc accessors | |
284 | (define* (request-absolute-uri r #:optional default-host default-port) | |
285 | (let ((uri (request-uri r))) | |
286 | (if (uri-host uri) | |
287 | uri | |
288 | (let ((host | |
289 | (or (request-host r) | |
290 | (if default-host | |
291 | (cons default-host default-port) | |
292 | (bad-request | |
293 | "URI not absolute, no Host header, and no default: ~s" | |
294 | uri))))) | |
295 | (build-uri (uri-scheme uri) | |
296 | #:host (car host) | |
297 | #:port (cdr host) | |
298 | #:path (uri-path uri) | |
299 | #:query (uri-query uri) | |
300 | #:fragment (uri-fragment uri)))))) |