Context Navigation

source: lib/http_client.c @ 4bc66ae

Last change on this file since 4bc66ae was 2423c93, checked in by Wilmer van der Gaast <wilmer@…>, at 2011-04-18T14:14:28Z
Send Connection: close headers so webservers won't try keepalive connections which http_client really can't deal with.
Property mode set to `100644`
File size: 10.6 KB

Line
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	static void http_free( struct http_request *req );
38
39
40	struct http_request http_dorequest( char host, int port, int ssl, char *request, http_input_function func, gpointer data )
41	{
42	struct http_request *req;
43	int error = 0;
44
45	req = g_new0( struct http_request, 1 );
46
47	if( ssl )
48	{
49	req->ssl = ssl_connect( host, port, http_ssl_connected, req );
50	if( req->ssl == NULL )
51	error = 1;
52	}
53	else
54	{
55	req->fd = proxy_connect( host, port, http_connected, req );
56	if( req->fd < 0 )
57	error = 1;
58	}
59
60	if( error )
61	{
62	http_free( req );
63	return NULL;
64	}
65
66	req->func = func;
67	req->data = data;
68	req->request = g_strdup( request );
69	req->request_length = strlen( request );
70	req->redir_ttl = 3;
71
72	return( req );
73	}
74
75	struct http_request http_dorequest_url( char url_string, http_input_function func, gpointer data )
76	{
77	url_t *url = g_new0( url_t, 1 );
78	char *request;
79	void *ret;
80
81	if( !url_set( url, url_string ) )
82	{
83	g_free( url );
84	return NULL;
85	}
86
87	if( url->proto != PROTO_HTTP && url->proto != PROTO_HTTPS )
88	{
89	g_free( url );
90	return NULL;
91	}
92
93	request = g_strdup_printf( "GET %s HTTP/1.0\r\n"
94	"Host: %s\r\n"
95	"Connection: close\r\n"
96	"User-Agent: BitlBee " BITLBEE_VERSION " " ARCH "/" CPU "\r\n"
97	"\r\n", url->file, url->host );
98
99	ret = http_dorequest( url->host, url->port,
100	url->proto == PROTO_HTTPS, request, func, data );
101
102	g_free( url );
103	g_free( request );
104	return ret;
105	}
106
107	/* This one is actually pretty simple... Might get more calls if we can't write
108	the whole request at once. */
109	static gboolean http_connected( gpointer data, int source, b_input_condition cond )
110	{
111	struct http_request *req = data;
112	int st;
113
114	if( source < 0 )
115	goto error;
116
117	if( req->inpa > 0 )
118	b_event_remove( req->inpa );
119
120	sock_make_nonblocking( req->fd );
121
122	if( req->ssl )
123	{
124	st = ssl_write( req->ssl, req->request + req->bytes_written,
125	req->request_length - req->bytes_written );
126	if( st < 0 )
127	{
128	if( ssl_errno != SSL_AGAIN )
129	{
130	ssl_disconnect( req->ssl );
131	goto error;
132	}
133	}
134	}
135	else
136	{
137	st = write( source, req->request + req->bytes_written,
138	req->request_length - req->bytes_written );
139	if( st < 0 )
140	{
141	if( !sockerr_again() )
142	{
143	closesocket( req->fd );
144	goto error;
145	}
146	}
147	}
148
149	if( st > 0 )
150	req->bytes_written += st;
151
152	if( req->bytes_written < req->request_length )
153	req->inpa = b_input_add( source,
154	req->ssl ? ssl_getdirection( req->ssl ) : B_EV_IO_WRITE,
155	http_connected, req );
156	else
157	req->inpa = b_input_add( source, B_EV_IO_READ, http_incoming_data, req );
158
159	return FALSE;
160
161	error:
162	req->status_string = g_strdup( "Error while writing HTTP request" );
163
164	req->func( req );
165	http_free( req );
166	return FALSE;
167	}
168
169	static gboolean http_ssl_connected( gpointer data, void *source, b_input_condition cond )
170	{
171	struct http_request *req = data;
172
173	if( source == NULL )
174	return http_connected( data, -1, cond );
175
176	req->fd = ssl_getfd( source );
177
178	return http_connected( data, req->fd, cond );
179	}
180
181	static gboolean http_incoming_data( gpointer data, int source, b_input_condition cond )
182	{
183	struct http_request *req = data;
184	int evil_server = 0;
185	char buffer[2048];
186	char end1, end2;
187	int st;
188
189	if( req->inpa > 0 )
190	b_event_remove( req->inpa );
191
192	if( req->ssl )
193	{
194	st = ssl_read( req->ssl, buffer, sizeof( buffer ) );
195	if( st < 0 )
196	{
197	if( ssl_errno != SSL_AGAIN )
198	{
199	/* goto cleanup; */
200
201	/* YAY! We have to deal with crappy Microsoft
202	servers that LOVE to send invalid TLS
203	packets that abort connections! \o/ */
204
205	goto got_reply;
206	}
207	}
208	else if( st == 0 )
209	{
210	goto got_reply;
211	}
212	}
213	else
214	{
215	st = read( req->fd, buffer, sizeof( buffer ) );
216	if( st < 0 )
217	{
218	if( !sockerr_again() )
219	{
220	req->status_string = g_strdup( strerror( errno ) );
221	goto cleanup;
222	}
223	}
224	else if( st == 0 )
225	{
226	goto got_reply;
227	}
228	}
229
230	if( st > 0 )
231	{
232	req->reply_headers = g_realloc( req->reply_headers, req->bytes_read + st + 1 );
233	memcpy( req->reply_headers + req->bytes_read, buffer, st );
234	req->bytes_read += st;
235	}
236
237	/* There will be more! */
238	req->inpa = b_input_add( req->fd,
239	req->ssl ? ssl_getdirection( req->ssl ) : B_EV_IO_READ,
240	http_incoming_data, req );
241
242	return FALSE;
243
244	got_reply:
245	/* Maybe if the webserver is overloaded, or when there's bad SSL
246	support... */
247	if( req->bytes_read == 0 )
248	{
249	req->status_string = g_strdup( "Empty HTTP reply" );
250	goto cleanup;
251	}
252
253	/* Zero termination is very convenient. */
254	req->reply_headers[req->bytes_read] = 0;
255
256	/* Find the separation between headers and body, and keep stupid
257	webservers in mind. */
258	end1 = strstr( req->reply_headers, "\r\n\r\n" );
259	end2 = strstr( req->reply_headers, "\n\n" );
260
261	if( end2 && end2 < end1 )
262	{
263	end1 = end2 + 1;
264	evil_server = 1;
265	}
266	else if( end1 )
267	{
268	end1 += 2;
269	}
270	else
271	{
272	req->status_string = g_strdup( "Malformed HTTP reply" );
273	goto cleanup;
274	}
275
276	*end1 = 0;
277
278	if( evil_server )
279	req->reply_body = end1 + 1;
280	else
281	req->reply_body = end1 + 2;
282
283	req->body_size = req->reply_headers + req->bytes_read - req->reply_body;
284
285	if( ( end1 = strchr( req->reply_headers, ' ' ) ) != NULL )
286	{
287	if( sscanf( end1 + 1, "%d", &req->status_code ) != 1 )
288	{
289	req->status_string = g_strdup( "Can't parse status code" );
290	req->status_code = -1;
291	}
292	else
293	{
294	char *eol;
295
296	if( evil_server )
297	eol = strchr( end1, '\n' );
298	else
299	eol = strchr( end1, '\r' );
300
301	req->status_string = g_strndup( end1 + 1, eol - end1 - 1 );
302
303	/* Just to be sure... */
304	if( ( eol = strchr( req->status_string, '\r' ) ) )
305	*eol = 0;
306	if( ( eol = strchr( req->status_string, '\n' ) ) )
307	*eol = 0;
308	}
309	}
310	else
311	{
312	req->status_string = g_strdup( "Can't locate status code" );
313	req->status_code = -1;
314	}
315
316	if( ( req->status_code == 301 \|\| req->status_code == 302 ) && req->redir_ttl-- > 0 )
317	{
318	char loc, new_request, *new_host;
319	int error = 0, new_port, new_proto;
320
321	/* We might fill it again, so let's not leak any memory. */
322	g_free( req->status_string );
323	req->status_string = NULL;
324
325	loc = strstr( req->reply_headers, "\nLocation: " );
326	if( loc == NULL ) /* We can't handle this redirect... */
327	{
328	req->status_string = g_strdup( "Can't locate Location: header" );
329	goto cleanup;
330	}
331
332	loc += 11;
333	while( *loc == ' ' )
334	loc ++;
335
336	/* TODO/FIXME: Possibly have to handle relative redirections,
337	and rewrite Host: headers. Not necessary for now, it's
338	enough for passport authentication like this. */
339
340	if( *loc == '/' )
341	{
342	/* Just a different pathname... */
343
344	/* Since we don't cache the servername, and since we
345	don't need this yet anyway, I won't implement it. */
346
347	req->status_string = g_strdup( "Can't handle recursive redirects" );
348
349	goto cleanup;
350	}
351	else
352	{
353	/* A whole URL */
354	url_t *url;
355	char *s;
356
357	s = strstr( loc, "\r\n" );
358	if( s == NULL )
359	goto cleanup;
360
361	url = g_new0( url_t, 1 );
362	*s = 0;
363
364	if( !url_set( url, loc ) )
365	{
366	req->status_string = g_strdup( "Malformed redirect URL" );
367	g_free( url );
368	goto cleanup;
369	}
370
371	/* Okay, this isn't fun! We have to rebuild the request... :-( */
372	new_request = g_malloc( req->request_length + strlen( url->file ) );
373
374	/* So, now I just allocated enough memory, so I'm
375	going to use strcat(), whether you like it or not. :-) */
376
377	sprintf( new_request, "GET %s HTTP/1.0", url->file );
378
379	s = strstr( req->request, "\r\n" );
380	if( s == NULL )
381	{
382	req->status_string = g_strdup( "Error while rebuilding request string" );
383	g_free( new_request );
384	g_free( url );
385	goto cleanup;
386	}
387
388	strcat( new_request, s );
389	new_host = g_strdup( url->host );
390	new_port = url->port;
391	new_proto = url->proto;
392
393	g_free( url );
394	}
395
396	if( req->ssl )
397	ssl_disconnect( req->ssl );
398	else
399	closesocket( req->fd );
400
401	req->fd = -1;
402	req->ssl = NULL;
403
404	if( new_proto == PROTO_HTTPS )
405	{
406	req->ssl = ssl_connect( new_host, new_port, http_ssl_connected, req );
407	if( req->ssl == NULL )
408	error = 1;
409	}
410	else
411	{
412	req->fd = proxy_connect( new_host, new_port, http_connected, req );
413	if( req->fd < 0 )
414	error = 1;
415	}
416	g_free( new_host );
417
418	if( error )
419	{
420	req->status_string = g_strdup( "Connection problem during redirect" );
421	g_free( new_request );
422	goto cleanup;
423	}
424
425	g_free( req->request );
426	g_free( req->reply_headers );
427	req->request = new_request;
428	req->request_length = strlen( new_request );
429	req->bytes_read = req->bytes_written = req->inpa = 0;
430	req->reply_headers = req->reply_body = NULL;
431
432	return FALSE;
433	}
434
435	/* Assume that a closed connection means we're finished, this indeed
436	breaks with keep-alive connections and faulty connections. */
437	req->finished = 1;
438
439	cleanup:
440	if( req->ssl )
441	ssl_disconnect( req->ssl );
442	else
443	closesocket( req->fd );
444
445	req->func( req );
446	http_free( req );
447	return FALSE;
448	}
449
450	static void http_free( struct http_request *req )
451	{
452	g_free( req->request );
453	g_free( req->reply_headers );
454	g_free( req->status_string );
455	g_free( req );
456	}
457

Note: See TracBrowser for help on using the repository browser.

Download in other formats: