1 /* libxml2 - Library for parsing XML documents
2 * Copyright (C) 2006-2019 Free Software Foundation, Inc.
3 *
4 * This file is not part of the GNU gettext program, but is used with
5 * GNU gettext.
6 *
7 * The original copyright notice is as follows:
8 */
9
10 /*
11 * Copyright (C) 1998-2012 Daniel Veillard. All Rights Reserved.
12 *
13 * Permission is hereby granted, free of charge, to any person obtaining a copy
14 * of this software and associated documentation files (the "Software"), to deal
15 * in the Software without restriction, including without limitation the rights
16 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
17 * copies of the Software, and to permit persons to whom the Software is fur-
18 * nished to do so, subject to the following conditions:
19 *
20 * The above copyright notice and this permission notice shall be included in
21 * all copies or substantial portions of the Software.
22 *
23 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
24 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FIT-
25 * NESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
26 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
27 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
28 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
29 * THE SOFTWARE.
30 *
31 * daniel@veillard.com
32 */
33
34 /*
35 * nanohttp.c: minimalist HTTP GET implementation to fetch external subsets.
36 * focuses on size, streamability, reentrancy and portability
37 *
38 * This is clearly not a general purpose HTTP implementation
39 * If you look for one, check:
40 * http://www.w3.org/Library/
41 */
42
43 #define IN_LIBXML
44 #include "libxml.h"
45
46 #ifdef LIBXML_HTTP_ENABLED
47 #include <string.h>
48
49 #ifdef HAVE_STDLIB_H
50 #include <stdlib.h>
51 #endif
52 #ifdef HAVE_UNISTD_H
53 #include <unistd.h>
54 #endif
55 #ifdef HAVE_SYS_TYPES_H
56 #include <sys/types.h>
57 #endif
58 #ifdef HAVE_SYS_SOCKET_H
59 #include <sys/socket.h>
60 #endif
61 #ifdef HAVE_NETINET_IN_H
62 #include <netinet/in.h>
63 #endif
64 #ifdef HAVE_ARPA_INET_H
65 #include <arpa/inet.h>
66 #endif
67 #ifdef HAVE_NETDB_H
68 #include <netdb.h>
69 #endif
70 #ifdef HAVE_RESOLV_H
71 #ifdef HAVE_ARPA_NAMESER_H
72 #include <arpa/nameser.h>
73 #endif
74 #include <resolv.h>
75 #endif
76 #ifdef HAVE_FCNTL_H
77 #include <fcntl.h>
78 #endif
79 #ifdef HAVE_ERRNO_H
80 #include <errno.h>
81 #endif
82 #ifdef HAVE_SYS_TIME_H
83 #include <sys/time.h>
84 #endif
85 #ifndef HAVE_POLL_H
86 #ifdef HAVE_SYS_SELECT_H
87 #include <sys/select.h>
88 #endif
89 #else
90 #include <poll.h>
91 #endif
92 #ifdef HAVE_STRINGS_H
93 #include <strings.h>
94 #endif
95 #ifdef LIBXML_ZLIB_ENABLED
96 #include <zlib.h>
97 #endif
98
99
100 #ifdef VMS
101 #include <stropts>
102 #define XML_SOCKLEN_T unsigned int
103 #endif
104
105 #if defined(_WIN32) && !defined(__CYGWIN__)
106 #include <wsockcompat.h>
107 #endif
108
109 #include <libxml/globals.h>
110 #include <libxml/xmlerror.h>
111 #include <libxml/xmlmemory.h>
112 #include <libxml/parser.h> /* for xmlStr(n)casecmp() */
113 #include <libxml/nanohttp.h>
114 #include <libxml/globals.h>
115 #include <libxml/uri.h>
116
117 /**
118 * A couple portability macros
119 */
120 #ifndef _WINSOCKAPI_
121 #if !defined(__BEOS__) || defined(__HAIKU__)
122 #define closesocket(s) close(s)
123 #endif
124 #define SOCKET int
125 #define INVALID_SOCKET (-1)
126 #endif
127
128 #ifdef __BEOS__
129 #ifndef PF_INET
130 #define PF_INET AF_INET
131 #endif
132 #endif
133
134 #ifndef XML_SOCKLEN_T
135 #define XML_SOCKLEN_T unsigned int
136 #endif
137
138 #ifdef STANDALONE
139 #define DEBUG_HTTP
140 #define xmlStrncasecmp(a, b, n) strncasecmp((char *)a, (char *)b, n)
141 #define xmlStrcasecmpi(a, b) strcasecmp((char *)a, (char *)b)
142 #endif
143
144 #define XML_NANO_HTTP_MAX_REDIR 10
145
146 #define XML_NANO_HTTP_CHUNK 4096
147
148 #define XML_NANO_HTTP_CLOSED 0
149 #define XML_NANO_HTTP_WRITE 1
150 #define XML_NANO_HTTP_READ 2
151 #define XML_NANO_HTTP_NONE 4
152
153 typedef struct xmlNanoHTTPCtxt {
154 char *protocol; /* the protocol name */
155 char *hostname; /* the host name */
156 int port; /* the port */
157 char *path; /* the path within the URL */
158 char *query; /* the query string */
159 SOCKET fd; /* the file descriptor for the socket */
160 int state; /* WRITE / READ / CLOSED */
161 char *out; /* buffer sent (zero terminated) */
162 char *outptr; /* index within the buffer sent */
163 char *in; /* the receiving buffer */
164 char *content; /* the start of the content */
165 char *inptr; /* the next byte to read from network */
166 char *inrptr; /* the next byte to give back to the client */
167 int inlen; /* len of the input buffer */
168 int last; /* return code for last operation */
169 int returnValue; /* the protocol return value */
170 int version; /* the protocol version */
171 int ContentLength; /* specified content length from HTTP header */
172 char *contentType; /* the MIME type for the input */
173 char *location; /* the new URL in case of redirect */
174 char *authHeader; /* contents of {WWW,Proxy}-Authenticate header */
175 char *encoding; /* encoding extracted from the contentType */
176 char *mimeType; /* Mime-Type extracted from the contentType */
177 #ifdef LIBXML_ZLIB_ENABLED
178 z_stream *strm; /* Zlib stream object */
179 int usesGzip; /* "Content-Encoding: gzip" was detected */
180 #endif
181 } xmlNanoHTTPCtxt, *xmlNanoHTTPCtxtPtr;
182
183 static int initialized = 0;
184 static char *proxy = NULL; /* the proxy name if any */
185 static int proxyPort; /* the proxy port if any */
186 static unsigned int timeout = 60;/* the select() timeout in seconds */
187
188 static int xmlNanoHTTPFetchContent( void * ctx, char ** ptr, int * len );
189
190 /**
191 * xmlHTTPErrMemory:
192 * @extra: extra informations
193 *
194 * Handle an out of memory condition
195 */
196 static void
xmlHTTPErrMemory(const char * extra)197 xmlHTTPErrMemory(const char *extra)
198 {
199 __xmlSimpleError(XML_FROM_HTTP, XML_ERR_NO_MEMORY, NULL, NULL, extra);
200 }
201
202 /**
203 * A portability function
204 */
socket_errno(void)205 static int socket_errno(void) {
206 #ifdef _WINSOCKAPI_
207 int err = WSAGetLastError();
208 switch(err) {
209 case WSAECONNRESET:
210 return(ECONNRESET);
211 case WSAEINPROGRESS:
212 return(EINPROGRESS);
213 case WSAEINTR:
214 return(EINTR);
215 case WSAESHUTDOWN:
216 return(ESHUTDOWN);
217 case WSAEWOULDBLOCK:
218 return(EWOULDBLOCK);
219 default:
220 return(err);
221 }
222 #else
223 return(errno);
224 #endif
225 }
226
227 #ifdef SUPPORT_IP6
228 static
have_ipv6(void)229 int have_ipv6(void) {
230 SOCKET s;
231
232 s = socket (AF_INET6, SOCK_STREAM, 0);
233 if (s != INVALID_SOCKET) {
234 close (s);
235 return (1);
236 }
237 return (0);
238 }
239 #endif
240
241 /**
242 * xmlNanoHTTPInit:
243 *
244 * Initialize the HTTP protocol layer.
245 * Currently it just checks for proxy informations
246 */
247
248 void
xmlNanoHTTPInit(void)249 xmlNanoHTTPInit(void) {
250 const char *env;
251 #ifdef _WINSOCKAPI_
252 WSADATA wsaData;
253 #endif
254
255 if (initialized)
256 return;
257
258 #ifdef _WINSOCKAPI_
259 if (WSAStartup(MAKEWORD(1, 1), &wsaData) != 0)
260 return;
261 #endif
262
263 if (proxy == NULL) {
264 proxyPort = 80;
265 env = getenv("no_proxy");
266 if (env && ((env[0] == '*') && (env[1] == 0)))
267 goto done;
268 env = getenv("http_proxy");
269 if (env != NULL) {
270 xmlNanoHTTPScanProxy(env);
271 goto done;
272 }
273 env = getenv("HTTP_PROXY");
274 if (env != NULL) {
275 xmlNanoHTTPScanProxy(env);
276 goto done;
277 }
278 }
279 done:
280 initialized = 1;
281 }
282
283 /**
284 * xmlNanoHTTPCleanup:
285 *
286 * Cleanup the HTTP protocol layer.
287 */
288
289 void
xmlNanoHTTPCleanup(void)290 xmlNanoHTTPCleanup(void) {
291 if (proxy != NULL) {
292 xmlFree(proxy);
293 proxy = NULL;
294 }
295 #ifdef _WINSOCKAPI_
296 if (initialized)
297 WSACleanup();
298 #endif
299 initialized = 0;
300 return;
301 }
302
303 /**
304 * xmlNanoHTTPScanURL:
305 * @ctxt: an HTTP context
306 * @URL: The URL used to initialize the context
307 *
308 * (Re)Initialize an HTTP context by parsing the URL and finding
309 * the protocol host port and path it indicates.
310 */
311
312 static void
xmlNanoHTTPScanURL(xmlNanoHTTPCtxtPtr ctxt,const char * URL)313 xmlNanoHTTPScanURL(xmlNanoHTTPCtxtPtr ctxt, const char *URL) {
314 xmlURIPtr uri;
315 int len;
316
317 /*
318 * Clear any existing data from the context
319 */
320 if (ctxt->protocol != NULL) {
321 xmlFree(ctxt->protocol);
322 ctxt->protocol = NULL;
323 }
324 if (ctxt->hostname != NULL) {
325 xmlFree(ctxt->hostname);
326 ctxt->hostname = NULL;
327 }
328 if (ctxt->path != NULL) {
329 xmlFree(ctxt->path);
330 ctxt->path = NULL;
331 }
332 if (ctxt->query != NULL) {
333 xmlFree(ctxt->query);
334 ctxt->query = NULL;
335 }
336 if (URL == NULL) return;
337
338 uri = xmlParseURIRaw(URL, 1);
339 if (uri == NULL)
340 return;
341
342 if ((uri->scheme == NULL) || (uri->server == NULL)) {
343 xmlFreeURI(uri);
344 return;
345 }
346
347 ctxt->protocol = xmlMemStrdup(uri->scheme);
348 /* special case of IPv6 addresses, the [] need to be removed */
349 if ((uri->server != NULL) && (*uri->server == '[')) {
350 len = strlen(uri->server);
351 if ((len > 2) && (uri->server[len - 1] == ']')) {
352 ctxt->hostname = (char *) xmlCharStrndup(uri->server + 1, len -2);
353 } else
354 ctxt->hostname = xmlMemStrdup(uri->server);
355 } else
356 ctxt->hostname = xmlMemStrdup(uri->server);
357 if (uri->path != NULL)
358 ctxt->path = xmlMemStrdup(uri->path);
359 else
360 ctxt->path = xmlMemStrdup("/");
361 if (uri->query != NULL)
362 ctxt->query = xmlMemStrdup(uri->query);
363 if (uri->port != 0)
364 ctxt->port = uri->port;
365
366 xmlFreeURI(uri);
367 }
368
369 /**
370 * xmlNanoHTTPScanProxy:
371 * @URL: The proxy URL used to initialize the proxy context
372 *
373 * (Re)Initialize the HTTP Proxy context by parsing the URL and finding
374 * the protocol host port it indicates.
375 * Should be like http://myproxy/ or http://myproxy:3128/
376 * A NULL URL cleans up proxy informations.
377 */
378
379 void
xmlNanoHTTPScanProxy(const char * URL)380 xmlNanoHTTPScanProxy(const char *URL) {
381 xmlURIPtr uri;
382
383 if (proxy != NULL) {
384 xmlFree(proxy);
385 proxy = NULL;
386 }
387 proxyPort = 0;
388
389 #ifdef DEBUG_HTTP
390 if (URL == NULL)
391 xmlGenericError(xmlGenericErrorContext,
392 "Removing HTTP proxy info\n");
393 else
394 xmlGenericError(xmlGenericErrorContext,
395 "Using HTTP proxy %s\n", URL);
396 #endif
397 if (URL == NULL) return;
398
399 uri = xmlParseURIRaw(URL, 1);
400 if ((uri == NULL) || (uri->scheme == NULL) ||
401 (strcmp(uri->scheme, "http")) || (uri->server == NULL)) {
402 __xmlIOErr(XML_FROM_HTTP, XML_HTTP_URL_SYNTAX, "Syntax Error\n");
403 if (uri != NULL)
404 xmlFreeURI(uri);
405 return;
406 }
407
408 proxy = xmlMemStrdup(uri->server);
409 if (uri->port != 0)
410 proxyPort = uri->port;
411
412 xmlFreeURI(uri);
413 }
414
415 /**
416 * xmlNanoHTTPNewCtxt:
417 * @URL: The URL used to initialize the context
418 *
419 * Allocate and initialize a new HTTP context.
420 *
421 * Returns an HTTP context or NULL in case of error.
422 */
423
424 static xmlNanoHTTPCtxtPtr
xmlNanoHTTPNewCtxt(const char * URL)425 xmlNanoHTTPNewCtxt(const char *URL) {
426 xmlNanoHTTPCtxtPtr ret;
427
428 ret = (xmlNanoHTTPCtxtPtr) xmlMalloc(sizeof(xmlNanoHTTPCtxt));
429 if (ret == NULL) {
430 xmlHTTPErrMemory("allocating context");
431 return(NULL);
432 }
433
434 memset(ret, 0, sizeof(xmlNanoHTTPCtxt));
435 ret->port = 80;
436 ret->returnValue = 0;
437 ret->fd = INVALID_SOCKET;
438 ret->ContentLength = -1;
439
440 xmlNanoHTTPScanURL(ret, URL);
441
442 return(ret);
443 }
444
445 /**
446 * xmlNanoHTTPFreeCtxt:
447 * @ctxt: an HTTP context
448 *
449 * Frees the context after closing the connection.
450 */
451
452 static void
xmlNanoHTTPFreeCtxt(xmlNanoHTTPCtxtPtr ctxt)453 xmlNanoHTTPFreeCtxt(xmlNanoHTTPCtxtPtr ctxt) {
454 if (ctxt == NULL) return;
455 if (ctxt->hostname != NULL) xmlFree(ctxt->hostname);
456 if (ctxt->protocol != NULL) xmlFree(ctxt->protocol);
457 if (ctxt->path != NULL) xmlFree(ctxt->path);
458 if (ctxt->query != NULL) xmlFree(ctxt->query);
459 if (ctxt->out != NULL) xmlFree(ctxt->out);
460 if (ctxt->in != NULL) xmlFree(ctxt->in);
461 if (ctxt->contentType != NULL) xmlFree(ctxt->contentType);
462 if (ctxt->encoding != NULL) xmlFree(ctxt->encoding);
463 if (ctxt->mimeType != NULL) xmlFree(ctxt->mimeType);
464 if (ctxt->location != NULL) xmlFree(ctxt->location);
465 if (ctxt->authHeader != NULL) xmlFree(ctxt->authHeader);
466 #ifdef LIBXML_ZLIB_ENABLED
467 if (ctxt->strm != NULL) {
468 inflateEnd(ctxt->strm);
469 xmlFree(ctxt->strm);
470 }
471 #endif
472
473 ctxt->state = XML_NANO_HTTP_NONE;
474 if (ctxt->fd != INVALID_SOCKET) closesocket(ctxt->fd);
475 ctxt->fd = INVALID_SOCKET;
476 xmlFree(ctxt);
477 }
478
479 /**
480 * xmlNanoHTTPSend:
481 * @ctxt: an HTTP context
482 *
483 * Send the input needed to initiate the processing on the server side
484 * Returns number of bytes sent or -1 on error.
485 */
486
487 static int
xmlNanoHTTPSend(xmlNanoHTTPCtxtPtr ctxt,const char * xmt_ptr,int outlen)488 xmlNanoHTTPSend(xmlNanoHTTPCtxtPtr ctxt, const char *xmt_ptr, int outlen)
489 {
490 int total_sent = 0;
491 #ifdef HAVE_POLL_H
492 struct pollfd p;
493 #else
494 struct timeval tv;
495 fd_set wfd;
496 #endif
497
498 if ((ctxt->state & XML_NANO_HTTP_WRITE) && (xmt_ptr != NULL)) {
499 while (total_sent < outlen) {
500 int nsent = send(ctxt->fd, SEND_ARG2_CAST (xmt_ptr + total_sent),
501 outlen - total_sent, 0);
502
503 if (nsent > 0)
504 total_sent += nsent;
505 else if ((nsent == -1) &&
506 #if defined(EAGAIN) && EAGAIN != EWOULDBLOCK
507 (socket_errno() != EAGAIN) &&
508 #endif
509 (socket_errno() != EWOULDBLOCK)) {
510 __xmlIOErr(XML_FROM_HTTP, 0, "send failed\n");
511 if (total_sent == 0)
512 total_sent = -1;
513 break;
514 } else {
515 /*
516 * No data sent
517 * Since non-blocking sockets are used, wait for
518 * socket to be writable or default timeout prior
519 * to retrying.
520 */
521 #ifndef HAVE_POLL_H
522 #ifndef _WINSOCKAPI_
523 if (ctxt->fd > FD_SETSIZE)
524 return -1;
525 #endif
526
527 tv.tv_sec = timeout;
528 tv.tv_usec = 0;
529 FD_ZERO(&wfd);
530 #ifdef _MSC_VER
531 #pragma warning(push)
532 #pragma warning(disable: 4018)
533 #endif
534 FD_SET(ctxt->fd, &wfd);
535 #ifdef _MSC_VER
536 #pragma warning(pop)
537 #endif
538 (void) select(ctxt->fd + 1, NULL, &wfd, NULL, &tv);
539 #else
540 p.fd = ctxt->fd;
541 p.events = POLLOUT;
542 (void) poll(&p, 1, timeout * 1000);
543 #endif /* !HAVE_POLL_H */
544 }
545 }
546 }
547
548 return total_sent;
549 }
550
551 /**
552 * xmlNanoHTTPRecv:
553 * @ctxt: an HTTP context
554 *
555 * Read information coming from the HTTP connection.
556 * This is a blocking call (but it blocks in select(), not read()).
557 *
558 * Returns the number of byte read or -1 in case of error.
559 */
560
561 static int
xmlNanoHTTPRecv(xmlNanoHTTPCtxtPtr ctxt)562 xmlNanoHTTPRecv(xmlNanoHTTPCtxtPtr ctxt)
563 {
564 #ifdef HAVE_POLL_H
565 struct pollfd p;
566 #else
567 fd_set rfd;
568 struct timeval tv;
569 #endif
570
571
572 while (ctxt->state & XML_NANO_HTTP_READ) {
573 if (ctxt->in == NULL) {
574 ctxt->in = (char *) xmlMallocAtomic(65000 * sizeof(char));
575 if (ctxt->in == NULL) {
576 xmlHTTPErrMemory("allocating input");
577 ctxt->last = -1;
578 return (-1);
579 }
580 ctxt->inlen = 65000;
581 ctxt->inptr = ctxt->content = ctxt->inrptr = ctxt->in;
582 }
583 if (ctxt->inrptr > ctxt->in + XML_NANO_HTTP_CHUNK) {
584 int delta = ctxt->inrptr - ctxt->in;
585 int len = ctxt->inptr - ctxt->inrptr;
586
587 memmove(ctxt->in, ctxt->inrptr, len);
588 ctxt->inrptr -= delta;
589 ctxt->content -= delta;
590 ctxt->inptr -= delta;
591 }
592 if ((ctxt->in + ctxt->inlen) < (ctxt->inptr + XML_NANO_HTTP_CHUNK)) {
593 int d_inptr = ctxt->inptr - ctxt->in;
594 int d_content = ctxt->content - ctxt->in;
595 int d_inrptr = ctxt->inrptr - ctxt->in;
596 char *tmp_ptr = ctxt->in;
597
598 ctxt->inlen *= 2;
599 ctxt->in = (char *) xmlRealloc(tmp_ptr, ctxt->inlen);
600 if (ctxt->in == NULL) {
601 xmlHTTPErrMemory("allocating input buffer");
602 xmlFree(tmp_ptr);
603 ctxt->last = -1;
604 return (-1);
605 }
606 ctxt->inptr = ctxt->in + d_inptr;
607 ctxt->content = ctxt->in + d_content;
608 ctxt->inrptr = ctxt->in + d_inrptr;
609 }
610 ctxt->last = recv(ctxt->fd, ctxt->inptr, XML_NANO_HTTP_CHUNK, 0);
611 if (ctxt->last > 0) {
612 ctxt->inptr += ctxt->last;
613 return (ctxt->last);
614 }
615 if (ctxt->last == 0) {
616 return (0);
617 }
618 if (ctxt->last == -1) {
619 switch (socket_errno()) {
620 case EINPROGRESS:
621 case EWOULDBLOCK:
622 #if defined(EAGAIN) && EAGAIN != EWOULDBLOCK
623 case EAGAIN:
624 #endif
625 break;
626
627 case ECONNRESET:
628 case ESHUTDOWN:
629 return (0);
630
631 default:
632 __xmlIOErr(XML_FROM_HTTP, 0, "recv failed\n");
633 return (-1);
634 }
635 }
636 #ifdef HAVE_POLL_H
637 p.fd = ctxt->fd;
638 p.events = POLLIN;
639 if ((poll(&p, 1, timeout * 1000) < 1)
640 #if defined(EINTR)
641 && (errno != EINTR)
642 #endif
643 )
644 return (0);
645 #else /* !HAVE_POLL_H */
646 #ifndef _WINSOCKAPI_
647 if (ctxt->fd > FD_SETSIZE)
648 return 0;
649 #endif
650
651 tv.tv_sec = timeout;
652 tv.tv_usec = 0;
653 FD_ZERO(&rfd);
654
655 #ifdef _MSC_VER
656 #pragma warning(push)
657 #pragma warning(disable: 4018)
658 #endif
659
660 FD_SET(ctxt->fd, &rfd);
661
662 #ifdef _MSC_VER
663 #pragma warning(pop)
664 #endif
665
666 if ((select(ctxt->fd + 1, &rfd, NULL, NULL, &tv) < 1)
667 #if defined(EINTR)
668 && (socket_errno() != EINTR)
669 #endif
670 )
671 return (0);
672 #endif /* !HAVE_POLL_H */
673 }
674 return (0);
675 }
676
677 /**
678 * xmlNanoHTTPReadLine:
679 * @ctxt: an HTTP context
680 *
681 * Read one line in the HTTP server output, usually for extracting
682 * the HTTP protocol informations from the answer header.
683 *
684 * Returns a newly allocated string with a copy of the line, or NULL
685 * which indicate the end of the input.
686 */
687
688 static char *
xmlNanoHTTPReadLine(xmlNanoHTTPCtxtPtr ctxt)689 xmlNanoHTTPReadLine(xmlNanoHTTPCtxtPtr ctxt) {
690 char buf[4096];
691 char *bp = buf;
692 int rc;
693
694 while (bp - buf < 4095) {
695 if (ctxt->inrptr == ctxt->inptr) {
696 if ( (rc = xmlNanoHTTPRecv(ctxt)) == 0) {
697 if (bp == buf)
698 return(NULL);
699 else
700 *bp = 0;
701 return(xmlMemStrdup(buf));
702 }
703 else if ( rc == -1 ) {
704 return ( NULL );
705 }
706 }
707 *bp = *ctxt->inrptr++;
708 if (*bp == '\n') {
709 *bp = 0;
710 return(xmlMemStrdup(buf));
711 }
712 if (*bp != '\r')
713 bp++;
714 }
715 buf[4095] = 0;
716 return(xmlMemStrdup(buf));
717 }
718
719
720 /**
721 * xmlNanoHTTPScanAnswer:
722 * @ctxt: an HTTP context
723 * @line: an HTTP header line
724 *
725 * Try to extract useful informations from the server answer.
726 * We currently parse and process:
727 * - The HTTP revision/ return code
728 * - The Content-Type, Mime-Type and charset used
729 * - The Location for redirect processing.
730 *
731 * Returns -1 in case of failure, the file descriptor number otherwise
732 */
733
734 static void
xmlNanoHTTPScanAnswer(xmlNanoHTTPCtxtPtr ctxt,const char * line)735 xmlNanoHTTPScanAnswer(xmlNanoHTTPCtxtPtr ctxt, const char *line) {
736 const char *cur = line;
737
738 if (line == NULL) return;
739
740 if (!strncmp(line, "HTTP/", 5)) {
741 int version = 0;
742 int ret = 0;
743
744 cur += 5;
745 while ((*cur >= '0') && (*cur <= '9')) {
746 version *= 10;
747 version += *cur - '0';
748 cur++;
749 }
750 if (*cur == '.') {
751 cur++;
752 if ((*cur >= '0') && (*cur <= '9')) {
753 version *= 10;
754 version += *cur - '0';
755 cur++;
756 }
757 while ((*cur >= '0') && (*cur <= '9'))
758 cur++;
759 } else
760 version *= 10;
761 if ((*cur != ' ') && (*cur != '\t')) return;
762 while ((*cur == ' ') || (*cur == '\t')) cur++;
763 if ((*cur < '0') || (*cur > '9')) return;
764 while ((*cur >= '0') && (*cur <= '9')) {
765 ret *= 10;
766 ret += *cur - '0';
767 cur++;
768 }
769 if ((*cur != 0) && (*cur != ' ') && (*cur != '\t')) return;
770 ctxt->returnValue = ret;
771 ctxt->version = version;
772 } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"Content-Type:", 13)) {
773 const xmlChar *charset, *last, *mime;
774 cur += 13;
775 while ((*cur == ' ') || (*cur == '\t')) cur++;
776 if (ctxt->contentType != NULL)
777 xmlFree(ctxt->contentType);
778 ctxt->contentType = xmlMemStrdup(cur);
779 mime = (const xmlChar *) cur;
780 last = mime;
781 while ((*last != 0) && (*last != ' ') && (*last != '\t') &&
782 (*last != ';') && (*last != ','))
783 last++;
784 if (ctxt->mimeType != NULL)
785 xmlFree(ctxt->mimeType);
786 ctxt->mimeType = (char *) xmlStrndup(mime, last - mime);
787 charset = xmlStrstr(BAD_CAST ctxt->contentType, BAD_CAST "charset=");
788 if (charset != NULL) {
789 charset += 8;
790 last = charset;
791 while ((*last != 0) && (*last != ' ') && (*last != '\t') &&
792 (*last != ';') && (*last != ','))
793 last++;
794 if (ctxt->encoding != NULL)
795 xmlFree(ctxt->encoding);
796 ctxt->encoding = (char *) xmlStrndup(charset, last - charset);
797 }
798 } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"ContentType:", 12)) {
799 const xmlChar *charset, *last, *mime;
800 cur += 12;
801 if (ctxt->contentType != NULL) return;
802 while ((*cur == ' ') || (*cur == '\t')) cur++;
803 ctxt->contentType = xmlMemStrdup(cur);
804 mime = (const xmlChar *) cur;
805 last = mime;
806 while ((*last != 0) && (*last != ' ') && (*last != '\t') &&
807 (*last != ';') && (*last != ','))
808 last++;
809 if (ctxt->mimeType != NULL)
810 xmlFree(ctxt->mimeType);
811 ctxt->mimeType = (char *) xmlStrndup(mime, last - mime);
812 charset = xmlStrstr(BAD_CAST ctxt->contentType, BAD_CAST "charset=");
813 if (charset != NULL) {
814 charset += 8;
815 last = charset;
816 while ((*last != 0) && (*last != ' ') && (*last != '\t') &&
817 (*last != ';') && (*last != ','))
818 last++;
819 if (ctxt->encoding != NULL)
820 xmlFree(ctxt->encoding);
821 ctxt->encoding = (char *) xmlStrndup(charset, last - charset);
822 }
823 } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"Location:", 9)) {
824 cur += 9;
825 while ((*cur == ' ') || (*cur == '\t')) cur++;
826 if (ctxt->location != NULL)
827 xmlFree(ctxt->location);
828 if (*cur == '/') {
829 xmlChar *tmp_http = xmlStrdup(BAD_CAST "http://");
830 xmlChar *tmp_loc =
831 xmlStrcat(tmp_http, (const xmlChar *) ctxt->hostname);
832 ctxt->location =
833 (char *) xmlStrcat (tmp_loc, (const xmlChar *) cur);
834 } else {
835 ctxt->location = xmlMemStrdup(cur);
836 }
837 } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"WWW-Authenticate:", 17)) {
838 cur += 17;
839 while ((*cur == ' ') || (*cur == '\t')) cur++;
840 if (ctxt->authHeader != NULL)
841 xmlFree(ctxt->authHeader);
842 ctxt->authHeader = xmlMemStrdup(cur);
843 } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"Proxy-Authenticate:", 19)) {
844 cur += 19;
845 while ((*cur == ' ') || (*cur == '\t')) cur++;
846 if (ctxt->authHeader != NULL)
847 xmlFree(ctxt->authHeader);
848 ctxt->authHeader = xmlMemStrdup(cur);
849 #ifdef LIBXML_ZLIB_ENABLED
850 } else if ( !xmlStrncasecmp( BAD_CAST line, BAD_CAST"Content-Encoding:", 17) ) {
851 cur += 17;
852 while ((*cur == ' ') || (*cur == '\t')) cur++;
853 if ( !xmlStrncasecmp( BAD_CAST cur, BAD_CAST"gzip", 4) ) {
854 ctxt->usesGzip = 1;
855
856 ctxt->strm = xmlMalloc(sizeof(z_stream));
857
858 if (ctxt->strm != NULL) {
859 ctxt->strm->zalloc = Z_NULL;
860 ctxt->strm->zfree = Z_NULL;
861 ctxt->strm->opaque = Z_NULL;
862 ctxt->strm->avail_in = 0;
863 ctxt->strm->next_in = Z_NULL;
864
865 inflateInit2( ctxt->strm, 31 );
866 }
867 }
868 #endif
869 } else if ( !xmlStrncasecmp( BAD_CAST line, BAD_CAST"Content-Length:", 15) ) {
870 cur += 15;
871 ctxt->ContentLength = strtol( cur, NULL, 10 );
872 }
873 }
874
875 /**
876 * xmlNanoHTTPConnectAttempt:
877 * @addr: a socket address structure
878 *
879 * Attempt a connection to the given IP:port endpoint. It forces
880 * non-blocking semantic on the socket, and allow 60 seconds for
881 * the host to answer.
882 *
883 * Returns -1 in case of failure, the file descriptor number otherwise
884 */
885
886 static SOCKET
xmlNanoHTTPConnectAttempt(struct sockaddr * addr)887 xmlNanoHTTPConnectAttempt(struct sockaddr *addr)
888 {
889 #ifndef HAVE_POLL_H
890 fd_set wfd;
891 #ifdef _WINSOCKAPI_
892 fd_set xfd;
893 #endif
894 struct timeval tv;
895 #else /* !HAVE_POLL_H */
896 struct pollfd p;
897 #endif /* !HAVE_POLL_H */
898 int status;
899
900 int addrlen;
901
902 SOCKET s;
903
904 #ifdef SUPPORT_IP6
905 if (addr->sa_family == AF_INET6) {
906 s = socket(PF_INET6, SOCK_STREAM, IPPROTO_TCP);
907 addrlen = sizeof(struct sockaddr_in6);
908 } else
909 #endif
910 {
911 s = socket(PF_INET, SOCK_STREAM, IPPROTO_TCP);
912 addrlen = sizeof(struct sockaddr_in);
913 }
914 if (s == INVALID_SOCKET) {
915 #ifdef DEBUG_HTTP
916 perror("socket");
917 #endif
918 __xmlIOErr(XML_FROM_HTTP, 0, "socket failed\n");
919 return INVALID_SOCKET;
920 }
921 #ifdef _WINSOCKAPI_
922 {
923 u_long one = 1;
924
925 status = ioctlsocket(s, FIONBIO, &one) == SOCKET_ERROR ? -1 : 0;
926 }
927 #else /* _WINSOCKAPI_ */
928 #if defined(VMS)
929 {
930 int enable = 1;
931
932 status = ioctl(s, FIONBIO, &enable);
933 }
934 #else /* VMS */
935 #if defined(__BEOS__) && !defined(__HAIKU__)
936 {
937 bool noblock = true;
938
939 status =
940 setsockopt(s, SOL_SOCKET, SO_NONBLOCK, &noblock,
941 sizeof(noblock));
942 }
943 #else /* __BEOS__ */
944 if ((status = fcntl(s, F_GETFL, 0)) != -1) {
945 #ifdef O_NONBLOCK
946 status |= O_NONBLOCK;
947 #else /* O_NONBLOCK */
948 #ifdef F_NDELAY
949 status |= F_NDELAY;
950 #endif /* F_NDELAY */
951 #endif /* !O_NONBLOCK */
952 status = fcntl(s, F_SETFL, status);
953 }
954 if (status < 0) {
955 #ifdef DEBUG_HTTP
956 perror("nonblocking");
957 #endif
958 __xmlIOErr(XML_FROM_HTTP, 0, "error setting non-blocking IO\n");
959 closesocket(s);
960 return INVALID_SOCKET;
961 }
962 #endif /* !__BEOS__ */
963 #endif /* !VMS */
964 #endif /* !_WINSOCKAPI_ */
965
966 if (connect(s, addr, addrlen) == -1) {
967 switch (socket_errno()) {
968 case EINPROGRESS:
969 case EWOULDBLOCK:
970 break;
971 default:
972 __xmlIOErr(XML_FROM_HTTP, 0,
973 "error connecting to HTTP server");
974 closesocket(s);
975 return INVALID_SOCKET;
976 }
977 }
978 #ifndef HAVE_POLL_H
979 tv.tv_sec = timeout;
980 tv.tv_usec = 0;
981
982 #ifdef _MSC_VER
983 #pragma warning(push)
984 #pragma warning(disable: 4018)
985 #endif
986 #ifndef _WINSOCKAPI_
987 if (s > FD_SETSIZE)
988 return INVALID_SOCKET;
989 #endif
990 FD_ZERO(&wfd);
991 FD_SET(s, &wfd);
992
993 #ifdef _WINSOCKAPI_
994 FD_ZERO(&xfd);
995 FD_SET(s, &xfd);
996
997 switch (select(s + 1, NULL, &wfd, &xfd, &tv))
998 #else
999 switch (select(s + 1, NULL, &wfd, NULL, &tv))
1000 #endif
1001 #ifdef _MSC_VER
1002 #pragma warning(pop)
1003 #endif
1004
1005 #else /* !HAVE_POLL_H */
1006 p.fd = s;
1007 p.events = POLLOUT;
1008 switch (poll(&p, 1, timeout * 1000))
1009 #endif /* !HAVE_POLL_H */
1010
1011 {
1012 case 0:
1013 /* Time out */
1014 __xmlIOErr(XML_FROM_HTTP, 0, "Connect attempt timed out");
1015 closesocket(s);
1016 return INVALID_SOCKET;
1017 case -1:
1018 /* Ermm.. ?? */
1019 __xmlIOErr(XML_FROM_HTTP, 0, "Connect failed");
1020 closesocket(s);
1021 return INVALID_SOCKET;
1022 }
1023
1024 #ifndef HAVE_POLL_H
1025 if (FD_ISSET(s, &wfd)
1026 #ifdef _WINSOCKAPI_
1027 || FD_ISSET(s, &xfd)
1028 #endif
1029 )
1030 #else /* !HAVE_POLL_H */
1031 if (p.revents == POLLOUT)
1032 #endif /* !HAVE_POLL_H */
1033 {
1034 XML_SOCKLEN_T len;
1035
1036 len = sizeof(status);
1037 #ifdef SO_ERROR
1038 if (getsockopt(s, SOL_SOCKET, SO_ERROR, (char *) &status, &len) <
1039 0) {
1040 /* Solaris error code */
1041 __xmlIOErr(XML_FROM_HTTP, 0, "getsockopt failed\n");
1042 closesocket(s);
1043 return INVALID_SOCKET;
1044 }
1045 #endif
1046 if (status) {
1047 __xmlIOErr(XML_FROM_HTTP, 0,
1048 "Error connecting to remote host");
1049 closesocket(s);
1050 errno = status;
1051 return INVALID_SOCKET;
1052 }
1053 } else {
1054 /* pbm */
1055 __xmlIOErr(XML_FROM_HTTP, 0, "select failed\n");
1056 closesocket(s);
1057 return INVALID_SOCKET;
1058 }
1059
1060 return (s);
1061 }
1062
1063 /**
1064 * xmlNanoHTTPConnectHost:
1065 * @host: the host name
1066 * @port: the port number
1067 *
1068 * Attempt a connection to the given host:port endpoint. It tries
1069 * the multiple IP provided by the DNS if available.
1070 *
1071 * Returns -1 in case of failure, the file descriptor number otherwise
1072 */
1073
1074 static SOCKET
xmlNanoHTTPConnectHost(const char * host,int port)1075 xmlNanoHTTPConnectHost(const char *host, int port)
1076 {
1077 struct sockaddr *addr = NULL;
1078 struct sockaddr_in sockin;
1079
1080 #ifdef SUPPORT_IP6
1081 struct in6_addr ia6;
1082 struct sockaddr_in6 sockin6;
1083 #endif
1084 SOCKET s;
1085
1086 memset (&sockin, 0, sizeof(sockin));
1087 #ifdef SUPPORT_IP6
1088 memset (&sockin6, 0, sizeof(sockin6));
1089 #endif
1090
1091 #if !defined(HAVE_GETADDRINFO) && defined(SUPPORT_IP6) && defined(RES_USE_INET6)
1092 if (have_ipv6 ())
1093 {
1094 if (!(_res.options & RES_INIT))
1095 res_init();
1096 _res.options |= RES_USE_INET6;
1097 }
1098 #endif
1099
1100 #if defined(HAVE_GETADDRINFO) && defined(SUPPORT_IP6) && !defined(_WIN32)
1101 if (have_ipv6 ())
1102 #endif
1103 #if defined(HAVE_GETADDRINFO) && (defined(SUPPORT_IP6) || defined(_WIN32))
1104 {
1105 int status;
1106 struct addrinfo hints, *res, *result;
1107
1108 result = NULL;
1109 memset (&hints, 0,sizeof(hints));
1110 hints.ai_socktype = SOCK_STREAM;
1111
1112 status = getaddrinfo (host, NULL, &hints, &result);
1113 if (status) {
1114 __xmlIOErr(XML_FROM_HTTP, 0, "getaddrinfo failed\n");
1115 return INVALID_SOCKET;
1116 }
1117
1118 for (res = result; res; res = res->ai_next) {
1119 if (res->ai_family == AF_INET) {
1120 if ((size_t)res->ai_addrlen > sizeof(sockin)) {
1121 __xmlIOErr(XML_FROM_HTTP, 0, "address size mismatch\n");
1122 freeaddrinfo (result);
1123 return INVALID_SOCKET;
1124 }
1125 memcpy (&sockin, res->ai_addr, res->ai_addrlen);
1126 sockin.sin_port = htons (port);
1127 addr = (struct sockaddr *)&sockin;
1128 #ifdef SUPPORT_IP6
1129 } else if (have_ipv6 () && (res->ai_family == AF_INET6)) {
1130 if ((size_t)res->ai_addrlen > sizeof(sockin6)) {
1131 __xmlIOErr(XML_FROM_HTTP, 0, "address size mismatch\n");
1132 freeaddrinfo (result);
1133 return INVALID_SOCKET;
1134 }
1135 memcpy (&sockin6, res->ai_addr, res->ai_addrlen);
1136 sockin6.sin6_port = htons (port);
1137 addr = (struct sockaddr *)&sockin6;
1138 #endif
1139 } else
1140 continue; /* for */
1141
1142 s = xmlNanoHTTPConnectAttempt (addr);
1143 if (s != INVALID_SOCKET) {
1144 freeaddrinfo (result);
1145 return (s);
1146 }
1147 }
1148
1149 if (result)
1150 freeaddrinfo (result);
1151 }
1152 #endif
1153 #if defined(HAVE_GETADDRINFO) && defined(SUPPORT_IP6) && !defined(_WIN32)
1154 else
1155 #endif
1156 #if !defined(HAVE_GETADDRINFO) || !defined(_WIN32)
1157 {
1158 struct hostent *h;
1159 struct in_addr ia;
1160 int i;
1161
1162 h = gethostbyname (GETHOSTBYNAME_ARG_CAST host);
1163 if (h == NULL) {
1164
1165 /*
1166 * Okay, I got fed up by the non-portability of this error message
1167 * extraction code. it work on Linux, if it work on your platform
1168 * and one want to enable it, send me the defined(foobar) needed
1169 */
1170 #if defined(HAVE_NETDB_H) && defined(HOST_NOT_FOUND) && defined(__linux__)
1171 const char *h_err_txt = "";
1172
1173 switch (h_errno) {
1174 case HOST_NOT_FOUND:
1175 h_err_txt = "Authoritive host not found";
1176 break;
1177
1178 case TRY_AGAIN:
1179 h_err_txt =
1180 "Non-authoritive host not found or server failure.";
1181 break;
1182
1183 case NO_RECOVERY:
1184 h_err_txt =
1185 "Non-recoverable errors: FORMERR, REFUSED, or NOTIMP.";
1186 break;
1187
1188 #ifdef NO_ADDRESS
1189 case NO_ADDRESS:
1190 h_err_txt =
1191 "Valid name, no data record of requested type.";
1192 break;
1193 #endif
1194
1195 default:
1196 h_err_txt = "No error text defined.";
1197 break;
1198 }
1199 __xmlIOErr(XML_FROM_HTTP, 0, h_err_txt);
1200 #else
1201 __xmlIOErr(XML_FROM_HTTP, 0, "Failed to resolve host");
1202 #endif
1203 return INVALID_SOCKET;
1204 }
1205
1206 for (i = 0; h->h_addr_list[i]; i++) {
1207 if (h->h_addrtype == AF_INET) {
1208 /* A records (IPv4) */
1209 if ((unsigned int) h->h_length > sizeof(ia)) {
1210 __xmlIOErr(XML_FROM_HTTP, 0, "address size mismatch\n");
1211 return INVALID_SOCKET;
1212 }
1213 memcpy (&ia, h->h_addr_list[i], h->h_length);
1214 sockin.sin_family = h->h_addrtype;
1215 sockin.sin_addr = ia;
1216 sockin.sin_port = (unsigned short)htons ((unsigned short)port);
1217 addr = (struct sockaddr *) &sockin;
1218 #ifdef SUPPORT_IP6
1219 } else if (have_ipv6 () && (h->h_addrtype == AF_INET6)) {
1220 /* AAAA records (IPv6) */
1221 if ((unsigned int) h->h_length > sizeof(ia6)) {
1222 __xmlIOErr(XML_FROM_HTTP, 0, "address size mismatch\n");
1223 return INVALID_SOCKET;
1224 }
1225 memcpy (&ia6, h->h_addr_list[i], h->h_length);
1226 sockin6.sin6_family = h->h_addrtype;
1227 sockin6.sin6_addr = ia6;
1228 sockin6.sin6_port = htons (port);
1229 addr = (struct sockaddr *) &sockin6;
1230 #endif
1231 } else
1232 break; /* for */
1233
1234 s = xmlNanoHTTPConnectAttempt (addr);
1235 if (s != INVALID_SOCKET)
1236 return (s);
1237 }
1238 }
1239 #endif
1240
1241 #ifdef DEBUG_HTTP
1242 xmlGenericError(xmlGenericErrorContext,
1243 "xmlNanoHTTPConnectHost: unable to connect to '%s'.\n",
1244 host);
1245 #endif
1246 return INVALID_SOCKET;
1247 }
1248
1249
1250 /**
1251 * xmlNanoHTTPOpen:
1252 * @URL: The URL to load
1253 * @contentType: if available the Content-Type information will be
1254 * returned at that location
1255 *
1256 * This function try to open a connection to the indicated resource
1257 * via HTTP GET.
1258 *
1259 * Returns NULL in case of failure, otherwise a request handler.
1260 * The contentType, if provided must be freed by the caller
1261 */
1262
1263 void*
xmlNanoHTTPOpen(const char * URL,char ** contentType)1264 xmlNanoHTTPOpen(const char *URL, char **contentType) {
1265 if (contentType != NULL) *contentType = NULL;
1266 return(xmlNanoHTTPMethod(URL, NULL, NULL, contentType, NULL, 0));
1267 }
1268
1269 /**
1270 * xmlNanoHTTPOpenRedir:
1271 * @URL: The URL to load
1272 * @contentType: if available the Content-Type information will be
1273 * returned at that location
1274 * @redir: if available the redirected URL will be returned
1275 *
1276 * This function try to open a connection to the indicated resource
1277 * via HTTP GET.
1278 *
1279 * Returns NULL in case of failure, otherwise a request handler.
1280 * The contentType, if provided must be freed by the caller
1281 */
1282
1283 void*
xmlNanoHTTPOpenRedir(const char * URL,char ** contentType,char ** redir)1284 xmlNanoHTTPOpenRedir(const char *URL, char **contentType, char **redir) {
1285 if (contentType != NULL) *contentType = NULL;
1286 if (redir != NULL) *redir = NULL;
1287 return(xmlNanoHTTPMethodRedir(URL, NULL, NULL, contentType, redir, NULL,0));
1288 }
1289
1290 /**
1291 * xmlNanoHTTPRead:
1292 * @ctx: the HTTP context
1293 * @dest: a buffer
1294 * @len: the buffer length
1295 *
1296 * This function tries to read @len bytes from the existing HTTP connection
1297 * and saves them in @dest. This is a blocking call.
1298 *
1299 * Returns the number of byte read. 0 is an indication of an end of connection.
1300 * -1 indicates a parameter error.
1301 */
1302 int
xmlNanoHTTPRead(void * ctx,void * dest,int len)1303 xmlNanoHTTPRead(void *ctx, void *dest, int len) {
1304 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr) ctx;
1305 #ifdef LIBXML_ZLIB_ENABLED
1306 int bytes_read = 0;
1307 int orig_avail_in;
1308 int z_ret;
1309 #endif
1310
1311 if (ctx == NULL) return(-1);
1312 if (dest == NULL) return(-1);
1313 if (len <= 0) return(0);
1314
1315 #ifdef LIBXML_ZLIB_ENABLED
1316 if (ctxt->usesGzip == 1) {
1317 if (ctxt->strm == NULL) return(0);
1318
1319 ctxt->strm->next_out = dest;
1320 ctxt->strm->avail_out = len;
1321 ctxt->strm->avail_in = ctxt->inptr - ctxt->inrptr;
1322
1323 while (ctxt->strm->avail_out > 0 &&
1324 (ctxt->strm->avail_in > 0 || xmlNanoHTTPRecv(ctxt) > 0)) {
1325 orig_avail_in = ctxt->strm->avail_in =
1326 ctxt->inptr - ctxt->inrptr - bytes_read;
1327 ctxt->strm->next_in = BAD_CAST (ctxt->inrptr + bytes_read);
1328
1329 z_ret = inflate(ctxt->strm, Z_NO_FLUSH);
1330 bytes_read += orig_avail_in - ctxt->strm->avail_in;
1331
1332 if (z_ret != Z_OK) break;
1333 }
1334
1335 ctxt->inrptr += bytes_read;
1336 return(len - ctxt->strm->avail_out);
1337 }
1338 #endif
1339
1340 while (ctxt->inptr - ctxt->inrptr < len) {
1341 if (xmlNanoHTTPRecv(ctxt) <= 0) break;
1342 }
1343 if (ctxt->inptr - ctxt->inrptr < len)
1344 len = ctxt->inptr - ctxt->inrptr;
1345 memcpy(dest, ctxt->inrptr, len);
1346 ctxt->inrptr += len;
1347 return(len);
1348 }
1349
1350 /**
1351 * xmlNanoHTTPClose:
1352 * @ctx: the HTTP context
1353 *
1354 * This function closes an HTTP context, it ends up the connection and
1355 * free all data related to it.
1356 */
1357 void
xmlNanoHTTPClose(void * ctx)1358 xmlNanoHTTPClose(void *ctx) {
1359 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr) ctx;
1360
1361 if (ctx == NULL) return;
1362
1363 xmlNanoHTTPFreeCtxt(ctxt);
1364 }
1365
1366 /**
1367 * xmlNanoHTTPMethodRedir:
1368 * @URL: The URL to load
1369 * @method: the HTTP method to use
1370 * @input: the input string if any
1371 * @contentType: the Content-Type information IN and OUT
1372 * @redir: the redirected URL OUT
1373 * @headers: the extra headers
1374 * @ilen: input length
1375 *
1376 * This function try to open a connection to the indicated resource
1377 * via HTTP using the given @method, adding the given extra headers
1378 * and the input buffer for the request content.
1379 *
1380 * Returns NULL in case of failure, otherwise a request handler.
1381 * The contentType, or redir, if provided must be freed by the caller
1382 */
1383
1384 void*
xmlNanoHTTPMethodRedir(const char * URL,const char * method,const char * input,char ** contentType,char ** redir,const char * headers,int ilen)1385 xmlNanoHTTPMethodRedir(const char *URL, const char *method, const char *input,
1386 char **contentType, char **redir,
1387 const char *headers, int ilen ) {
1388 xmlNanoHTTPCtxtPtr ctxt;
1389 char *bp, *p;
1390 int blen;
1391 SOCKET ret;
1392 int nbRedirects = 0;
1393 char *redirURL = NULL;
1394 #ifdef DEBUG_HTTP
1395 int xmt_bytes;
1396 #endif
1397
1398 if (URL == NULL) return(NULL);
1399 if (method == NULL) method = "GET";
1400 xmlNanoHTTPInit();
1401
1402 retry:
1403 if (redirURL == NULL) {
1404 ctxt = xmlNanoHTTPNewCtxt(URL);
1405 if (ctxt == NULL)
1406 return(NULL);
1407 } else {
1408 ctxt = xmlNanoHTTPNewCtxt(redirURL);
1409 if (ctxt == NULL)
1410 return(NULL);
1411 ctxt->location = xmlMemStrdup(redirURL);
1412 }
1413
1414 if ((ctxt->protocol == NULL) || (strcmp(ctxt->protocol, "http"))) {
1415 __xmlIOErr(XML_FROM_HTTP, XML_HTTP_URL_SYNTAX, "Not a valid HTTP URI");
1416 xmlNanoHTTPFreeCtxt(ctxt);
1417 if (redirURL != NULL) xmlFree(redirURL);
1418 return(NULL);
1419 }
1420 if (ctxt->hostname == NULL) {
1421 __xmlIOErr(XML_FROM_HTTP, XML_HTTP_UNKNOWN_HOST,
1422 "Failed to identify host in URI");
1423 xmlNanoHTTPFreeCtxt(ctxt);
1424 if (redirURL != NULL) xmlFree(redirURL);
1425 return(NULL);
1426 }
1427 if (proxy) {
1428 blen = strlen(ctxt->hostname) * 2 + 16;
1429 ret = xmlNanoHTTPConnectHost(proxy, proxyPort);
1430 }
1431 else {
1432 blen = strlen(ctxt->hostname);
1433 ret = xmlNanoHTTPConnectHost(ctxt->hostname, ctxt->port);
1434 }
1435 if (ret == INVALID_SOCKET) {
1436 xmlNanoHTTPFreeCtxt(ctxt);
1437 if (redirURL != NULL) xmlFree(redirURL);
1438 return(NULL);
1439 }
1440 ctxt->fd = ret;
1441
1442 if (input == NULL)
1443 ilen = 0;
1444 else
1445 blen += 36;
1446
1447 if (headers != NULL)
1448 blen += strlen(headers) + 2;
1449 if (contentType && *contentType)
1450 /* reserve for string plus 'Content-Type: \r\n" */
1451 blen += strlen(*contentType) + 16;
1452 if (ctxt->query != NULL)
1453 /* 1 for '?' */
1454 blen += strlen(ctxt->query) + 1;
1455 blen += strlen(method) + strlen(ctxt->path) + 24;
1456 #ifdef LIBXML_ZLIB_ENABLED
1457 /* reserve for possible 'Accept-Encoding: gzip' string */
1458 blen += 23;
1459 #endif
1460 if (ctxt->port != 80) {
1461 /* reserve space for ':xxxxx', incl. potential proxy */
1462 if (proxy)
1463 blen += 17;
1464 else
1465 blen += 11;
1466 }
1467 bp = (char*)xmlMallocAtomic(blen);
1468 if ( bp == NULL ) {
1469 xmlNanoHTTPFreeCtxt( ctxt );
1470 xmlHTTPErrMemory("allocating header buffer");
1471 return ( NULL );
1472 }
1473
1474 p = bp;
1475
1476 if (proxy) {
1477 if (ctxt->port != 80) {
1478 p += snprintf( p, blen - (p - bp), "%s http://%s:%d%s",
1479 method, ctxt->hostname,
1480 ctxt->port, ctxt->path );
1481 }
1482 else
1483 p += snprintf( p, blen - (p - bp), "%s http://%s%s", method,
1484 ctxt->hostname, ctxt->path);
1485 }
1486 else
1487 p += snprintf( p, blen - (p - bp), "%s %s", method, ctxt->path);
1488
1489 if (ctxt->query != NULL)
1490 p += snprintf( p, blen - (p - bp), "?%s", ctxt->query);
1491
1492 if (ctxt->port == 80) {
1493 p += snprintf( p, blen - (p - bp), " HTTP/1.0\r\nHost: %s\r\n",
1494 ctxt->hostname);
1495 } else {
1496 p += snprintf( p, blen - (p - bp), " HTTP/1.0\r\nHost: %s:%d\r\n",
1497 ctxt->hostname, ctxt->port);
1498 }
1499
1500 #ifdef LIBXML_ZLIB_ENABLED
1501 p += snprintf(p, blen - (p - bp), "Accept-Encoding: gzip\r\n");
1502 #endif
1503
1504 if (contentType != NULL && *contentType)
1505 p += snprintf(p, blen - (p - bp), "Content-Type: %s\r\n", *contentType);
1506
1507 if (headers != NULL)
1508 p += snprintf( p, blen - (p - bp), "%s", headers );
1509
1510 if (input != NULL)
1511 snprintf(p, blen - (p - bp), "Content-Length: %d\r\n\r\n", ilen );
1512 else
1513 snprintf(p, blen - (p - bp), "\r\n");
1514
1515 #ifdef DEBUG_HTTP
1516 xmlGenericError(xmlGenericErrorContext,
1517 "-> %s%s", proxy? "(Proxy) " : "", bp);
1518 if ((blen -= strlen(bp)+1) < 0)
1519 xmlGenericError(xmlGenericErrorContext,
1520 "ERROR: overflowed buffer by %d bytes\n", -blen);
1521 #endif
1522 ctxt->outptr = ctxt->out = bp;
1523 ctxt->state = XML_NANO_HTTP_WRITE;
1524 blen = strlen( ctxt->out );
1525 #ifdef DEBUG_HTTP
1526 xmt_bytes = xmlNanoHTTPSend(ctxt, ctxt->out, blen );
1527 if ( xmt_bytes != blen )
1528 xmlGenericError( xmlGenericErrorContext,
1529 "xmlNanoHTTPMethodRedir: Only %d of %d %s %s\n",
1530 xmt_bytes, blen,
1531 "bytes of HTTP headers sent to host",
1532 ctxt->hostname );
1533 #else
1534 xmlNanoHTTPSend(ctxt, ctxt->out, blen );
1535 #endif
1536
1537 if ( input != NULL ) {
1538 #ifdef DEBUG_HTTP
1539 xmt_bytes = xmlNanoHTTPSend( ctxt, input, ilen );
1540
1541 if ( xmt_bytes != ilen )
1542 xmlGenericError( xmlGenericErrorContext,
1543 "xmlNanoHTTPMethodRedir: Only %d of %d %s %s\n",
1544 xmt_bytes, ilen,
1545 "bytes of HTTP content sent to host",
1546 ctxt->hostname );
1547 #else
1548 xmlNanoHTTPSend( ctxt, input, ilen );
1549 #endif
1550 }
1551
1552 ctxt->state = XML_NANO_HTTP_READ;
1553
1554 while ((p = xmlNanoHTTPReadLine(ctxt)) != NULL) {
1555 if (*p == 0) {
1556 ctxt->content = ctxt->inrptr;
1557 xmlFree(p);
1558 break;
1559 }
1560 xmlNanoHTTPScanAnswer(ctxt, p);
1561
1562 #ifdef DEBUG_HTTP
1563 xmlGenericError(xmlGenericErrorContext, "<- %s\n", p);
1564 #endif
1565 xmlFree(p);
1566 }
1567
1568 if ((ctxt->location != NULL) && (ctxt->returnValue >= 300) &&
1569 (ctxt->returnValue < 400)) {
1570 #ifdef DEBUG_HTTP
1571 xmlGenericError(xmlGenericErrorContext,
1572 "\nRedirect to: %s\n", ctxt->location);
1573 #endif
1574 while ( xmlNanoHTTPRecv(ctxt) > 0 )
1575 ;
1576 if (nbRedirects < XML_NANO_HTTP_MAX_REDIR) {
1577 nbRedirects++;
1578 if (redirURL != NULL)
1579 xmlFree(redirURL);
1580 redirURL = xmlMemStrdup(ctxt->location);
1581 xmlNanoHTTPFreeCtxt(ctxt);
1582 goto retry;
1583 }
1584 xmlNanoHTTPFreeCtxt(ctxt);
1585 if (redirURL != NULL) xmlFree(redirURL);
1586 #ifdef DEBUG_HTTP
1587 xmlGenericError(xmlGenericErrorContext,
1588 "xmlNanoHTTPMethodRedir: Too many redirects, aborting ...\n");
1589 #endif
1590 return(NULL);
1591 }
1592
1593 if (contentType != NULL) {
1594 if (ctxt->contentType != NULL)
1595 *contentType = xmlMemStrdup(ctxt->contentType);
1596 else
1597 *contentType = NULL;
1598 }
1599
1600 if ((redir != NULL) && (redirURL != NULL)) {
1601 *redir = redirURL;
1602 } else {
1603 if (redirURL != NULL)
1604 xmlFree(redirURL);
1605 if (redir != NULL)
1606 *redir = NULL;
1607 }
1608
1609 #ifdef DEBUG_HTTP
1610 if (ctxt->contentType != NULL)
1611 xmlGenericError(xmlGenericErrorContext,
1612 "\nCode %d, content-type '%s'\n\n",
1613 ctxt->returnValue, ctxt->contentType);
1614 else
1615 xmlGenericError(xmlGenericErrorContext,
1616 "\nCode %d, no content-type\n\n",
1617 ctxt->returnValue);
1618 #endif
1619
1620 return((void *) ctxt);
1621 }
1622
1623 /**
1624 * xmlNanoHTTPMethod:
1625 * @URL: The URL to load
1626 * @method: the HTTP method to use
1627 * @input: the input string if any
1628 * @contentType: the Content-Type information IN and OUT
1629 * @headers: the extra headers
1630 * @ilen: input length
1631 *
1632 * This function try to open a connection to the indicated resource
1633 * via HTTP using the given @method, adding the given extra headers
1634 * and the input buffer for the request content.
1635 *
1636 * Returns NULL in case of failure, otherwise a request handler.
1637 * The contentType, if provided must be freed by the caller
1638 */
1639
1640 void*
xmlNanoHTTPMethod(const char * URL,const char * method,const char * input,char ** contentType,const char * headers,int ilen)1641 xmlNanoHTTPMethod(const char *URL, const char *method, const char *input,
1642 char **contentType, const char *headers, int ilen) {
1643 return(xmlNanoHTTPMethodRedir(URL, method, input, contentType,
1644 NULL, headers, ilen));
1645 }
1646
1647 /**
1648 * xmlNanoHTTPFetch:
1649 * @URL: The URL to load
1650 * @filename: the filename where the content should be saved
1651 * @contentType: if available the Content-Type information will be
1652 * returned at that location
1653 *
1654 * This function try to fetch the indicated resource via HTTP GET
1655 * and save it's content in the file.
1656 *
1657 * Returns -1 in case of failure, 0 incase of success. The contentType,
1658 * if provided must be freed by the caller
1659 */
1660 int
xmlNanoHTTPFetch(const char * URL,const char * filename,char ** contentType)1661 xmlNanoHTTPFetch(const char *URL, const char *filename, char **contentType) {
1662 void *ctxt = NULL;
1663 char *buf = NULL;
1664 int fd;
1665 int len;
1666 int ret = 0;
1667
1668 if (filename == NULL) return(-1);
1669 ctxt = xmlNanoHTTPOpen(URL, contentType);
1670 if (ctxt == NULL) return(-1);
1671
1672 if (!strcmp(filename, "-"))
1673 fd = 0;
1674 else {
1675 fd = open(filename, O_CREAT | O_WRONLY, 00644);
1676 if (fd < 0) {
1677 xmlNanoHTTPClose(ctxt);
1678 if ((contentType != NULL) && (*contentType != NULL)) {
1679 xmlFree(*contentType);
1680 *contentType = NULL;
1681 }
1682 return(-1);
1683 }
1684 }
1685
1686 xmlNanoHTTPFetchContent( ctxt, &buf, &len );
1687 if ( len > 0 ) {
1688 if (write(fd, buf, len) == -1) {
1689 ret = -1;
1690 }
1691 }
1692
1693 xmlNanoHTTPClose(ctxt);
1694 close(fd);
1695 return(ret);
1696 }
1697
1698 #ifdef LIBXML_OUTPUT_ENABLED
1699 /**
1700 * xmlNanoHTTPSave:
1701 * @ctxt: the HTTP context
1702 * @filename: the filename where the content should be saved
1703 *
1704 * This function saves the output of the HTTP transaction to a file
1705 * It closes and free the context at the end
1706 *
1707 * Returns -1 in case of failure, 0 incase of success.
1708 */
1709 int
xmlNanoHTTPSave(void * ctxt,const char * filename)1710 xmlNanoHTTPSave(void *ctxt, const char *filename) {
1711 char *buf = NULL;
1712 int fd;
1713 int len;
1714 int ret = 0;
1715
1716 if ((ctxt == NULL) || (filename == NULL)) return(-1);
1717
1718 if (!strcmp(filename, "-"))
1719 fd = 0;
1720 else {
1721 fd = open(filename, O_CREAT | O_WRONLY, 0666);
1722 if (fd < 0) {
1723 xmlNanoHTTPClose(ctxt);
1724 return(-1);
1725 }
1726 }
1727
1728 xmlNanoHTTPFetchContent( ctxt, &buf, &len );
1729 if ( len > 0 ) {
1730 if (write(fd, buf, len) == -1) {
1731 ret = -1;
1732 }
1733 }
1734
1735 xmlNanoHTTPClose(ctxt);
1736 close(fd);
1737 return(ret);
1738 }
1739 #endif /* LIBXML_OUTPUT_ENABLED */
1740
1741 /**
1742 * xmlNanoHTTPReturnCode:
1743 * @ctx: the HTTP context
1744 *
1745 * Get the latest HTTP return code received
1746 *
1747 * Returns the HTTP return code for the request.
1748 */
1749 int
xmlNanoHTTPReturnCode(void * ctx)1750 xmlNanoHTTPReturnCode(void *ctx) {
1751 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr) ctx;
1752
1753 if (ctxt == NULL) return(-1);
1754
1755 return(ctxt->returnValue);
1756 }
1757
1758 /**
1759 * xmlNanoHTTPAuthHeader:
1760 * @ctx: the HTTP context
1761 *
1762 * Get the authentication header of an HTTP context
1763 *
1764 * Returns the stashed value of the WWW-Authenticate or Proxy-Authenticate
1765 * header.
1766 */
1767 const char *
xmlNanoHTTPAuthHeader(void * ctx)1768 xmlNanoHTTPAuthHeader(void *ctx) {
1769 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr) ctx;
1770
1771 if (ctxt == NULL) return(NULL);
1772
1773 return(ctxt->authHeader);
1774 }
1775
1776 /**
1777 * xmlNanoHTTPContentLength:
1778 * @ctx: the HTTP context
1779 *
1780 * Provides the specified content length from the HTTP header.
1781 *
1782 * Return the specified content length from the HTTP header. Note that
1783 * a value of -1 indicates that the content length element was not included in
1784 * the response header.
1785 */
1786 int
xmlNanoHTTPContentLength(void * ctx)1787 xmlNanoHTTPContentLength( void * ctx ) {
1788 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr)ctx;
1789
1790 return ( ( ctxt == NULL ) ? -1 : ctxt->ContentLength );
1791 }
1792
1793 /**
1794 * xmlNanoHTTPRedir:
1795 * @ctx: the HTTP context
1796 *
1797 * Provides the specified redirection URL if available from the HTTP header.
1798 *
1799 * Return the specified redirection URL or NULL if not redirected.
1800 */
1801 const char *
xmlNanoHTTPRedir(void * ctx)1802 xmlNanoHTTPRedir( void * ctx ) {
1803 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr)ctx;
1804
1805 return ( ( ctxt == NULL ) ? NULL : ctxt->location );
1806 }
1807
1808 /**
1809 * xmlNanoHTTPEncoding:
1810 * @ctx: the HTTP context
1811 *
1812 * Provides the specified encoding if specified in the HTTP headers.
1813 *
1814 * Return the specified encoding or NULL if not available
1815 */
1816 const char *
xmlNanoHTTPEncoding(void * ctx)1817 xmlNanoHTTPEncoding( void * ctx ) {
1818 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr)ctx;
1819
1820 return ( ( ctxt == NULL ) ? NULL : ctxt->encoding );
1821 }
1822
1823 /**
1824 * xmlNanoHTTPMimeType:
1825 * @ctx: the HTTP context
1826 *
1827 * Provides the specified Mime-Type if specified in the HTTP headers.
1828 *
1829 * Return the specified Mime-Type or NULL if not available
1830 */
1831 const char *
xmlNanoHTTPMimeType(void * ctx)1832 xmlNanoHTTPMimeType( void * ctx ) {
1833 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr)ctx;
1834
1835 return ( ( ctxt == NULL ) ? NULL : ctxt->mimeType );
1836 }
1837
1838 /**
1839 * xmlNanoHTTPFetchContent:
1840 * @ctx: the HTTP context
1841 * @ptr: pointer to set to the content buffer.
1842 * @len: integer pointer to hold the length of the content
1843 *
1844 * Check if all the content was read
1845 *
1846 * Returns 0 if all the content was read and available, returns
1847 * -1 if received content length was less than specified or an error
1848 * occurred.
1849 */
1850 static int
xmlNanoHTTPFetchContent(void * ctx,char ** ptr,int * len)1851 xmlNanoHTTPFetchContent( void * ctx, char ** ptr, int * len ) {
1852 xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr)ctx;
1853
1854 int rc = 0;
1855 int cur_lgth;
1856 int rcvd_lgth;
1857 int dummy_int;
1858 char * dummy_ptr = NULL;
1859
1860 /* Dummy up return input parameters if not provided */
1861
1862 if ( len == NULL )
1863 len = &dummy_int;
1864
1865 if ( ptr == NULL )
1866 ptr = &dummy_ptr;
1867
1868 /* But can't work without the context pointer */
1869
1870 if ( ( ctxt == NULL ) || ( ctxt->content == NULL ) ) {
1871 *len = 0;
1872 *ptr = NULL;
1873 return ( -1 );
1874 }
1875
1876 rcvd_lgth = ctxt->inptr - ctxt->content;
1877
1878 while ( (cur_lgth = xmlNanoHTTPRecv( ctxt )) > 0 ) {
1879
1880 rcvd_lgth += cur_lgth;
1881 if ( (ctxt->ContentLength > 0) && (rcvd_lgth >= ctxt->ContentLength) )
1882 break;
1883 }
1884
1885 *ptr = ctxt->content;
1886 *len = rcvd_lgth;
1887
1888 if ( ( ctxt->ContentLength > 0 ) && ( rcvd_lgth < ctxt->ContentLength ) )
1889 rc = -1;
1890 else if ( rcvd_lgth == 0 )
1891 rc = -1;
1892
1893 return ( rc );
1894 }
1895
1896 #ifdef STANDALONE
main(int argc,char ** argv)1897 int main(int argc, char **argv) {
1898 char *contentType = NULL;
1899
1900 if (argv[1] != NULL) {
1901 if (argv[2] != NULL)
1902 xmlNanoHTTPFetch(argv[1], argv[2], &contentType);
1903 else
1904 xmlNanoHTTPFetch(argv[1], "-", &contentType);
1905 if (contentType != NULL) xmlFree(contentType);
1906 } else {
1907 xmlGenericError(xmlGenericErrorContext,
1908 "%s: minimal HTTP GET implementation\n", argv[0]);
1909 xmlGenericError(xmlGenericErrorContext,
1910 "\tusage %s [ URL [ filename ] ]\n", argv[0]);
1911 }
1912 xmlNanoHTTPCleanup();
1913 xmlMemoryDump();
1914 return(0);
1915 }
1916 #endif /* STANDALONE */
1917 #else /* !LIBXML_HTTP_ENABLED */
1918 #ifdef STANDALONE
1919 #include <stdio.h>
main(int argc,char ** argv)1920 int main(int argc, char **argv) {
1921 xmlGenericError(xmlGenericErrorContext,
1922 "%s : HTTP support not compiled in\n", argv[0]);
1923 return(0);
1924 }
1925 #endif /* STANDALONE */
1926 #endif /* LIBXML_HTTP_ENABLED */
1927 #define bottom_nanohttp
1928 #include "elfgcchack.h"
1929