1 | /********************************************************************\ |
---|
2 | * BitlBee -- An IRC to other IM-networks gateway * |
---|
3 | * * |
---|
4 | * Copyright 2002-2005 Wilmer van der Gaast and others * |
---|
5 | \********************************************************************/ |
---|
6 | |
---|
7 | /* HTTP(S) module */ |
---|
8 | |
---|
9 | /* |
---|
10 | This program is free software; you can redistribute it and/or modify |
---|
11 | it under the terms of the GNU General Public License as published by |
---|
12 | the Free Software Foundation; either version 2 of the License, or |
---|
13 | (at your option) any later version. |
---|
14 | |
---|
15 | This program is distributed in the hope that it will be useful, |
---|
16 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
---|
17 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
---|
18 | GNU General Public License for more details. |
---|
19 | |
---|
20 | You should have received a copy of the GNU General Public License with |
---|
21 | the Debian GNU/Linux distribution in /usr/share/common-licenses/GPL; |
---|
22 | if not, write to the Free Software Foundation, Inc., 59 Temple Place, |
---|
23 | Suite 330, Boston, MA 02111-1307 USA |
---|
24 | */ |
---|
25 | |
---|
26 | #include <string.h> |
---|
27 | #include <stdio.h> |
---|
28 | |
---|
29 | #include "http_client.h" |
---|
30 | #include "url.h" |
---|
31 | #include "sock.h" |
---|
32 | |
---|
33 | |
---|
34 | static gboolean http_connected( gpointer data, int source, b_input_condition cond ); |
---|
35 | static gboolean http_ssl_connected( gpointer data, void *source, b_input_condition cond ); |
---|
36 | static gboolean http_incoming_data( gpointer data, int source, b_input_condition cond ); |
---|
37 | |
---|
38 | |
---|
39 | void *http_dorequest( char *host, int port, int ssl, char *request, http_input_function func, gpointer data ) |
---|
40 | { |
---|
41 | struct http_request *req; |
---|
42 | int error = 0; |
---|
43 | |
---|
44 | req = g_new0( struct http_request, 1 ); |
---|
45 | |
---|
46 | if( ssl ) |
---|
47 | { |
---|
48 | req->ssl = ssl_connect( host, port, http_ssl_connected, req ); |
---|
49 | if( req->ssl == NULL ) |
---|
50 | error = 1; |
---|
51 | } |
---|
52 | else |
---|
53 | { |
---|
54 | req->fd = proxy_connect( host, port, http_connected, req ); |
---|
55 | if( req->fd < 0 ) |
---|
56 | error = 1; |
---|
57 | } |
---|
58 | |
---|
59 | if( error ) |
---|
60 | { |
---|
61 | http_free( req ); |
---|
62 | return NULL; |
---|
63 | } |
---|
64 | |
---|
65 | req->func = func; |
---|
66 | req->data = data; |
---|
67 | req->request = g_strdup( request ); |
---|
68 | req->request_length = strlen( request ); |
---|
69 | |
---|
70 | return( req ); |
---|
71 | } |
---|
72 | |
---|
73 | void *http_dorequest_url( char *url_string, http_input_function func, gpointer data ) |
---|
74 | { |
---|
75 | url_t *url = g_new0( url_t, 1 ); |
---|
76 | char *request; |
---|
77 | void *ret; |
---|
78 | |
---|
79 | if( !url_set( url, url_string ) ) |
---|
80 | { |
---|
81 | g_free( url ); |
---|
82 | return NULL; |
---|
83 | } |
---|
84 | |
---|
85 | if( url->proto != PROTO_HTTP && url->proto != PROTO_HTTPS ) |
---|
86 | { |
---|
87 | g_free( url ); |
---|
88 | return NULL; |
---|
89 | } |
---|
90 | |
---|
91 | request = g_strdup_printf( "GET %s HTTP/1.0\r\n" |
---|
92 | "Host: %s\r\n" |
---|
93 | "User-Agent: BitlBee " BITLBEE_VERSION " " ARCH "/" CPU "\r\n" |
---|
94 | "\r\n", url->file, url->host ); |
---|
95 | |
---|
96 | ret = http_dorequest( url->host, url->port, |
---|
97 | url->proto == PROTO_HTTPS, request, func, data ); |
---|
98 | |
---|
99 | g_free( url ); |
---|
100 | g_free( request ); |
---|
101 | return ret; |
---|
102 | } |
---|
103 | |
---|
104 | /* This one is actually pretty simple... Might get more calls if we can't write |
---|
105 | the whole request at once. */ |
---|
106 | static gboolean http_connected( gpointer data, int source, b_input_condition cond ) |
---|
107 | { |
---|
108 | struct http_request *req = data; |
---|
109 | int st; |
---|
110 | |
---|
111 | if( source < 0 ) |
---|
112 | goto error; |
---|
113 | |
---|
114 | if( req->inpa > 0 ) |
---|
115 | b_event_remove( req->inpa ); |
---|
116 | |
---|
117 | sock_make_nonblocking( req->fd ); |
---|
118 | |
---|
119 | if( req->ssl ) |
---|
120 | { |
---|
121 | st = ssl_write( req->ssl, req->request + req->bytes_written, |
---|
122 | req->request_length - req->bytes_written ); |
---|
123 | if( st < 0 ) |
---|
124 | { |
---|
125 | if( ssl_errno != SSL_AGAIN ) |
---|
126 | { |
---|
127 | ssl_disconnect( req->ssl ); |
---|
128 | goto error; |
---|
129 | } |
---|
130 | } |
---|
131 | } |
---|
132 | else |
---|
133 | { |
---|
134 | st = write( source, req->request + req->bytes_written, |
---|
135 | req->request_length - req->bytes_written ); |
---|
136 | if( st < 0 ) |
---|
137 | { |
---|
138 | if( !sockerr_again() ) |
---|
139 | { |
---|
140 | closesocket( req->fd ); |
---|
141 | goto error; |
---|
142 | } |
---|
143 | } |
---|
144 | } |
---|
145 | |
---|
146 | if( st > 0 ) |
---|
147 | req->bytes_written += st; |
---|
148 | |
---|
149 | if( req->bytes_written < req->request_length ) |
---|
150 | req->inpa = b_input_add( source, |
---|
151 | req->ssl ? ssl_getdirection( req->ssl ) : B_EV_IO_WRITE, |
---|
152 | http_connected, req ); |
---|
153 | else |
---|
154 | req->inpa = b_input_add( source, B_EV_IO_READ, http_incoming_data, req ); |
---|
155 | |
---|
156 | return FALSE; |
---|
157 | |
---|
158 | error: |
---|
159 | req->status_string = g_strdup( "Error while writing HTTP request" ); |
---|
160 | |
---|
161 | req->func( req ); |
---|
162 | http_free( req ); |
---|
163 | return FALSE; |
---|
164 | } |
---|
165 | |
---|
166 | static gboolean http_ssl_connected( gpointer data, void *source, b_input_condition cond ) |
---|
167 | { |
---|
168 | struct http_request *req = data; |
---|
169 | |
---|
170 | if( source == NULL ) |
---|
171 | return http_connected( data, -1, cond ); |
---|
172 | |
---|
173 | req->fd = ssl_getfd( source ); |
---|
174 | |
---|
175 | return http_connected( data, req->fd, cond ); |
---|
176 | } |
---|
177 | |
---|
178 | static gboolean http_incoming_data( gpointer data, int source, b_input_condition cond ) |
---|
179 | { |
---|
180 | struct http_request *req = data; |
---|
181 | int evil_server = 0; |
---|
182 | char buffer[2048]; |
---|
183 | char *end1, *end2; |
---|
184 | int st; |
---|
185 | |
---|
186 | if( req->inpa > 0 ) |
---|
187 | b_event_remove( req->inpa ); |
---|
188 | |
---|
189 | if( req->ssl ) |
---|
190 | { |
---|
191 | st = ssl_read( req->ssl, buffer, sizeof( buffer ) ); |
---|
192 | if( st < 0 ) |
---|
193 | { |
---|
194 | if( ssl_errno != SSL_AGAIN ) |
---|
195 | { |
---|
196 | /* goto cleanup; */ |
---|
197 | |
---|
198 | /* YAY! We have to deal with crappy Microsoft |
---|
199 | servers that LOVE to send invalid TLS |
---|
200 | packets that abort connections! \o/ */ |
---|
201 | |
---|
202 | goto got_reply; |
---|
203 | } |
---|
204 | } |
---|
205 | else if( st == 0 ) |
---|
206 | { |
---|
207 | goto got_reply; |
---|
208 | } |
---|
209 | } |
---|
210 | else |
---|
211 | { |
---|
212 | st = read( req->fd, buffer, sizeof( buffer ) ); |
---|
213 | if( st < 0 ) |
---|
214 | { |
---|
215 | if( !sockerr_again() ) |
---|
216 | { |
---|
217 | req->status_string = g_strdup( strerror( errno ) ); |
---|
218 | goto cleanup; |
---|
219 | } |
---|
220 | } |
---|
221 | else if( st == 0 ) |
---|
222 | { |
---|
223 | goto got_reply; |
---|
224 | } |
---|
225 | } |
---|
226 | |
---|
227 | if( st > 0 ) |
---|
228 | { |
---|
229 | req->reply_headers = g_realloc( req->reply_headers, req->bytes_read + st + 1 ); |
---|
230 | memcpy( req->reply_headers + req->bytes_read, buffer, st ); |
---|
231 | req->bytes_read += st; |
---|
232 | } |
---|
233 | |
---|
234 | /* There will be more! */ |
---|
235 | req->inpa = b_input_add( req->fd, |
---|
236 | req->ssl ? ssl_getdirection( req->ssl ) : B_EV_IO_READ, |
---|
237 | http_incoming_data, req ); |
---|
238 | |
---|
239 | return FALSE; |
---|
240 | |
---|
241 | got_reply: |
---|
242 | /* Maybe if the webserver is overloaded, or when there's bad SSL |
---|
243 | support... */ |
---|
244 | if( req->bytes_read == 0 ) |
---|
245 | { |
---|
246 | req->status_string = g_strdup( "Empty HTTP reply" ); |
---|
247 | goto cleanup; |
---|
248 | } |
---|
249 | |
---|
250 | /* Zero termination is very convenient. */ |
---|
251 | req->reply_headers[req->bytes_read] = 0; |
---|
252 | |
---|
253 | /* Find the separation between headers and body, and keep stupid |
---|
254 | webservers in mind. */ |
---|
255 | end1 = strstr( req->reply_headers, "\r\n\r\n" ); |
---|
256 | end2 = strstr( req->reply_headers, "\n\n" ); |
---|
257 | |
---|
258 | if( end2 && end2 < end1 ) |
---|
259 | { |
---|
260 | end1 = end2 + 1; |
---|
261 | evil_server = 1; |
---|
262 | } |
---|
263 | else if( end1 ) |
---|
264 | { |
---|
265 | end1 += 2; |
---|
266 | } |
---|
267 | else |
---|
268 | { |
---|
269 | req->status_string = g_strdup( "Malformed HTTP reply" ); |
---|
270 | goto cleanup; |
---|
271 | } |
---|
272 | |
---|
273 | *end1 = 0; |
---|
274 | |
---|
275 | if( evil_server ) |
---|
276 | req->reply_body = end1 + 1; |
---|
277 | else |
---|
278 | req->reply_body = end1 + 2; |
---|
279 | |
---|
280 | req->body_size = req->reply_headers + req->bytes_read - req->reply_body; |
---|
281 | |
---|
282 | if( ( end1 = strchr( req->reply_headers, ' ' ) ) != NULL ) |
---|
283 | { |
---|
284 | if( sscanf( end1 + 1, "%d", &req->status_code ) != 1 ) |
---|
285 | { |
---|
286 | req->status_string = g_strdup( "Can't parse status code" ); |
---|
287 | req->status_code = -1; |
---|
288 | } |
---|
289 | else |
---|
290 | { |
---|
291 | char *eol; |
---|
292 | |
---|
293 | if( evil_server ) |
---|
294 | eol = strchr( end1, '\n' ); |
---|
295 | else |
---|
296 | eol = strchr( end1, '\r' ); |
---|
297 | |
---|
298 | req->status_string = g_strndup( end1 + 1, eol - end1 - 1 ); |
---|
299 | |
---|
300 | /* Just to be sure... */ |
---|
301 | if( ( eol = strchr( req->status_string, '\r' ) ) ) |
---|
302 | *eol = 0; |
---|
303 | if( ( eol = strchr( req->status_string, '\n' ) ) ) |
---|
304 | *eol = 0; |
---|
305 | } |
---|
306 | } |
---|
307 | else |
---|
308 | { |
---|
309 | req->status_string = g_strdup( "Can't locate status code" ); |
---|
310 | req->status_code = -1; |
---|
311 | } |
---|
312 | |
---|
313 | if( req->status_code == 301 || req->status_code == 302 ) |
---|
314 | { |
---|
315 | char *loc, *new_request, *new_host; |
---|
316 | int error = 0, new_port, new_proto; |
---|
317 | |
---|
318 | /* We might fill it again, so let's not leak any memory. */ |
---|
319 | g_free( req->status_string ); |
---|
320 | req->status_string = NULL; |
---|
321 | |
---|
322 | loc = strstr( req->reply_headers, "\nLocation: " ); |
---|
323 | if( loc == NULL ) /* We can't handle this redirect... */ |
---|
324 | { |
---|
325 | req->status_string = g_strdup( "Can't locate Location: header" ); |
---|
326 | goto cleanup; |
---|
327 | } |
---|
328 | |
---|
329 | loc += 11; |
---|
330 | while( *loc == ' ' ) |
---|
331 | loc ++; |
---|
332 | |
---|
333 | /* TODO/FIXME: Possibly have to handle relative redirections, |
---|
334 | and rewrite Host: headers. Not necessary for now, it's |
---|
335 | enough for passport authentication like this. */ |
---|
336 | |
---|
337 | if( *loc == '/' ) |
---|
338 | { |
---|
339 | /* Just a different pathname... */ |
---|
340 | |
---|
341 | /* Since we don't cache the servername, and since we |
---|
342 | don't need this yet anyway, I won't implement it. */ |
---|
343 | |
---|
344 | req->status_string = g_strdup( "Can't handle recursive redirects" ); |
---|
345 | |
---|
346 | goto cleanup; |
---|
347 | } |
---|
348 | else |
---|
349 | { |
---|
350 | /* A whole URL */ |
---|
351 | url_t *url; |
---|
352 | char *s; |
---|
353 | |
---|
354 | s = strstr( loc, "\r\n" ); |
---|
355 | if( s == NULL ) |
---|
356 | goto cleanup; |
---|
357 | |
---|
358 | url = g_new0( url_t, 1 ); |
---|
359 | *s = 0; |
---|
360 | |
---|
361 | if( !url_set( url, loc ) ) |
---|
362 | { |
---|
363 | req->status_string = g_strdup( "Malformed redirect URL" ); |
---|
364 | g_free( url ); |
---|
365 | goto cleanup; |
---|
366 | } |
---|
367 | |
---|
368 | /* Okay, this isn't fun! We have to rebuild the request... :-( */ |
---|
369 | new_request = g_malloc( req->request_length + strlen( url->file ) ); |
---|
370 | |
---|
371 | /* So, now I just allocated enough memory, so I'm |
---|
372 | going to use strcat(), whether you like it or not. :-) */ |
---|
373 | |
---|
374 | sprintf( new_request, "GET %s HTTP/1.0", url->file ); |
---|
375 | |
---|
376 | s = strstr( req->request, "\r\n" ); |
---|
377 | if( s == NULL ) |
---|
378 | { |
---|
379 | req->status_string = g_strdup( "Error while rebuilding request string" ); |
---|
380 | g_free( new_request ); |
---|
381 | g_free( url ); |
---|
382 | goto cleanup; |
---|
383 | } |
---|
384 | |
---|
385 | strcat( new_request, s ); |
---|
386 | new_host = g_strdup( url->host ); |
---|
387 | new_port = url->port; |
---|
388 | new_proto = url->proto; |
---|
389 | |
---|
390 | g_free( url ); |
---|
391 | } |
---|
392 | |
---|
393 | if( req->ssl ) |
---|
394 | ssl_disconnect( req->ssl ); |
---|
395 | else |
---|
396 | closesocket( req->fd ); |
---|
397 | |
---|
398 | req->fd = -1; |
---|
399 | req->ssl = NULL; |
---|
400 | |
---|
401 | if( new_proto == PROTO_HTTPS ) |
---|
402 | { |
---|
403 | req->ssl = ssl_connect( new_host, new_port, http_ssl_connected, req ); |
---|
404 | if( req->ssl == NULL ) |
---|
405 | error = 1; |
---|
406 | } |
---|
407 | else |
---|
408 | { |
---|
409 | req->fd = proxy_connect( new_host, new_port, http_connected, req ); |
---|
410 | if( req->fd < 0 ) |
---|
411 | error = 1; |
---|
412 | } |
---|
413 | g_free( new_host ); |
---|
414 | |
---|
415 | if( error ) |
---|
416 | { |
---|
417 | req->status_string = g_strdup( "Connection problem during redirect" ); |
---|
418 | g_free( new_request ); |
---|
419 | goto cleanup; |
---|
420 | } |
---|
421 | |
---|
422 | g_free( req->request ); |
---|
423 | g_free( req->reply_headers ); |
---|
424 | req->request = new_request; |
---|
425 | req->request_length = strlen( new_request ); |
---|
426 | req->bytes_read = req->bytes_written = req->inpa = 0; |
---|
427 | req->reply_headers = req->reply_body = NULL; |
---|
428 | |
---|
429 | return FALSE; |
---|
430 | } |
---|
431 | |
---|
432 | /* Assume that a closed connection means we're finished, this indeed |
---|
433 | breaks with keep-alive connections and faulty connections. */ |
---|
434 | req->finished = 1; |
---|
435 | |
---|
436 | cleanup: |
---|
437 | if( req->ssl ) |
---|
438 | ssl_disconnect( req->ssl ); |
---|
439 | else |
---|
440 | closesocket( req->fd ); |
---|
441 | |
---|
442 | req->func( req ); |
---|
443 | http_free( req ); |
---|
444 | return FALSE; |
---|
445 | } |
---|
446 | |
---|
447 | void http_free( struct http_request *req ) |
---|
448 | { |
---|
449 | g_free( req->request ); |
---|
450 | g_free( req->reply_headers ); |
---|
451 | g_free( req->status_string ); |
---|
452 | g_free( req ); |
---|
453 | } |
---|
454 | |
---|