nanohttp.c 43.6 KB
Newer Older
1 2 3 4 5 6 7 8 9 10
/*
 * nanohttp.c: minimalist HTTP GET implementation to fetch external subsets.
 *             focuses on size, streamability, reentrancy and portability
 *
 * This is clearly not a general purpose HTTP implementation
 * If you look for one, check:
 *         http://www.w3.org/Library/
 *
 * See Copyright for the status of this software.
 *
11
 * daniel@veillard.com
12 13
 */
 
14
#define NEED_SOCKETS
15
#define IN_LIBXML
16
#include "libxml.h"
17 18 19 20 21 22 23 24 25 26

#ifdef LIBXML_HTTP_ENABLED
#include <string.h>

#ifdef HAVE_STDLIB_H
#include <stdlib.h>
#endif
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
27 28 29
#ifdef HAVE_SYS_TYPES_H
#include <sys/types.h>
#endif
30 31 32 33 34 35 36 37 38 39 40 41
#ifdef HAVE_SYS_SOCKET_H
#include <sys/socket.h>
#endif
#ifdef HAVE_NETINET_IN_H
#include <netinet/in.h>
#endif
#ifdef HAVE_ARPA_INET_H
#include <arpa/inet.h>
#endif
#ifdef HAVE_NETDB_H
#include <netdb.h>
#endif
42
#ifdef HAVE_RESOLV_H
43 44 45
#ifdef HAVE_ARPA_NAMESER_H
#include <arpa/nameser.h>
#endif
46 47
#include <resolv.h>
#endif
48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65
#ifdef HAVE_FCNTL_H
#include <fcntl.h> 
#endif
#ifdef HAVE_ERRNO_H
#include <errno.h>
#endif
#ifdef HAVE_SYS_TIME_H
#include <sys/time.h>
#endif
#ifdef HAVE_SYS_SELECT_H
#include <sys/select.h>
#endif
#ifdef HAVE_STRINGS_H
#include <strings.h>
#endif
#ifdef SUPPORT_IP6
#include <resolv.h>
#endif
66 67 68 69
#ifdef HAVE_ZLIB_H
#include <zlib.h>
#endif

70 71 72

#ifdef VMS
#include <stropts>
73
#define XML_SOCKLEN_T unsigned int
74 75 76
#define SOCKET int
#endif

77
#if defined(__MINGW32__) || defined(_WIN32_WCE)
78 79 80
#define _WINSOCKAPI_
#include <wsockcompat.h>
#include <winsock2.h>
81 82
#undef XML_SOCKLEN_T
#define XML_SOCKLEN_T unsigned int
83 84 85
#endif


86
#include <libxml/globals.h>
87
#include <libxml/xmlerror.h>
88 89 90
#include <libxml/xmlmemory.h>
#include <libxml/parser.h> /* for xmlStr(n)casecmp() */
#include <libxml/nanohttp.h>
91
#include <libxml/globals.h>
92
#include <libxml/uri.h>
93 94 95 96 97

/**
 * A couple portability macros
 */
#ifndef _WINSOCKAPI_
98
#ifndef __BEOS__
99
#define closesocket(s) close(s)
100
#endif
101 102 103
#define SOCKET int
#endif

104 105 106 107 108 109
#ifdef __BEOS__
#ifndef PF_INET
#define PF_INET AF_INET
#endif
#endif

110 111
#ifndef XML_SOCKLEN_T
#define XML_SOCKLEN_T unsigned int
112 113 114 115
#endif
#ifndef SOCKET
#define SOCKET int
#endif
116

117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136
#ifdef STANDALONE
#define DEBUG_HTTP
#define xmlStrncasecmp(a, b, n) strncasecmp((char *)a, (char *)b, n)
#define xmlStrcasecmpi(a, b) strcasecmp((char *)a, (char *)b)
#endif

#define XML_NANO_HTTP_MAX_REDIR	10

#define XML_NANO_HTTP_CHUNK	4096

#define XML_NANO_HTTP_CLOSED	0
#define XML_NANO_HTTP_WRITE	1
#define XML_NANO_HTTP_READ	2
#define XML_NANO_HTTP_NONE	4

typedef struct xmlNanoHTTPCtxt {
    char *protocol;	/* the protocol name */
    char *hostname;	/* the host name */
    int port;		/* the port */
    char *path;		/* the path within the URL */
137
    char *query;	/* the query string */
138 139 140 141 142 143 144 145 146 147 148
    SOCKET fd;		/* the file descriptor for the socket */
    int state;		/* WRITE / READ / CLOSED */
    char *out;		/* buffer sent (zero terminated) */
    char *outptr;	/* index within the buffer sent */
    char *in;		/* the receiving buffer */
    char *content;	/* the start of the content */
    char *inptr;	/* the next byte to read from network */
    char *inrptr;	/* the next byte to give back to the client */
    int inlen;		/* len of the input buffer */
    int last;		/* return code for last operation */
    int returnValue;	/* the protocol return value */
149
    int ContentLength;  /* specified content length from HTTP header */
150 151 152
    char *contentType;	/* the MIME type for the input */
    char *location;	/* the new URL in case of redirect */
    char *authHeader;	/* contents of {WWW,Proxy}-Authenticate header */
153
    char *encoding;	/* encoding extracted from the contentType */
154
    char *mimeType;	/* Mime-Type extracted from the contentType */
155 156 157 158
#ifdef HAVE_ZLIB_H
    z_stream *strm;	/* Zlib stream object */
    int usesGzip;	/* "Content-Encoding: gzip" was detected */
#endif
159 160 161 162 163 164 165
} xmlNanoHTTPCtxt, *xmlNanoHTTPCtxtPtr;

static int initialized = 0;
static char *proxy = NULL;	 /* the proxy name if any */
static int proxyPort;	/* the proxy port if any */
static unsigned int timeout = 60;/* the select() timeout in seconds */

166
static int xmlNanoHTTPFetchContent( void * ctx, char ** ptr, int * len );
167

168 169 170 171 172 173 174 175 176 177 178 179
/**
 * xmlHTTPErrMemory:
 * @extra:  extra informations
 *
 * Handle an out of memory condition
 */
static void
xmlHTTPErrMemory(const char *extra)
{
    __xmlSimpleError(XML_FROM_HTTP, XML_ERR_NO_MEMORY, NULL, NULL, extra);
}

180 181 182
/**
 * A portability function
 */
183
static int socket_errno(void) {
184 185 186 187 188 189 190
#ifdef _WINSOCKAPI_
    return(WSAGetLastError());
#else
    return(errno);
#endif
}

191
#ifdef SUPPORT_IP6
192 193
static
int have_ipv6(void) {
194 195 196 197 198 199 200 201 202 203 204
    int s;

    s = socket (AF_INET6, SOCK_STREAM, 0);
    if (s != -1) {
	close (s);
	return (1);
    }
    return (0);
}
#endif

205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229
/**
 * xmlNanoHTTPInit:
 *
 * Initialize the HTTP protocol layer.
 * Currently it just checks for proxy informations
 */

void
xmlNanoHTTPInit(void) {
    const char *env;
#ifdef _WINSOCKAPI_
    WSADATA wsaData;    
#endif

    if (initialized)
	return;

#ifdef _WINSOCKAPI_
    if (WSAStartup(MAKEWORD(1, 1), &wsaData) != 0)
	return;
#endif

    if (proxy == NULL) {
	proxyPort = 80;
	env = getenv("no_proxy");
230
	if (env && ((env[0] == '*') && (env[1] == 0)))
231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247
	    goto done;
	env = getenv("http_proxy");
	if (env != NULL) {
	    xmlNanoHTTPScanProxy(env);
	    goto done;
	}
	env = getenv("HTTP_PROXY");
	if (env != NULL) {
	    xmlNanoHTTPScanProxy(env);
	    goto done;
	}
    }
done:
    initialized = 1;
}

/**
248
 * xmlNanoHTTPCleanup:
249 250 251 252 253 254
 *
 * Cleanup the HTTP protocol layer.
 */

void
xmlNanoHTTPCleanup(void) {
255
    if (proxy != NULL) {
256
	xmlFree(proxy);
257 258
	proxy = NULL;
    }
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277
#ifdef _WINSOCKAPI_
    if (initialized)
	WSACleanup();
#endif
    initialized = 0;
    return;
}

/**
 * xmlNanoHTTPScanURL:
 * @ctxt:  an HTTP context
 * @URL:  The URL used to initialize the context
 *
 * (Re)Initialize an HTTP context by parsing the URL and finding
 * the protocol host port and path it indicates.
 */

static void
xmlNanoHTTPScanURL(xmlNanoHTTPCtxtPtr ctxt, const char *URL) {
278 279 280 281
    xmlURIPtr uri;
    /*
     * Clear any existing data from the context
     */
282 283 284 285 286 287 288 289 290 291 292 293
    if (ctxt->protocol != NULL) { 
        xmlFree(ctxt->protocol);
	ctxt->protocol = NULL;
    }
    if (ctxt->hostname != NULL) { 
        xmlFree(ctxt->hostname);
	ctxt->hostname = NULL;
    }
    if (ctxt->path != NULL) { 
        xmlFree(ctxt->path);
	ctxt->path = NULL;
    }
294 295 296 297
    if (ctxt->query != NULL) { 
        xmlFree(ctxt->query);
	ctxt->query = NULL;
    }
298
    if (URL == NULL) return;
299

300
    uri = xmlParseURIRaw(URL, 1);
301 302
    if (uri == NULL)
	return;
303

304 305 306
    if ((uri->scheme == NULL) || (uri->server == NULL)) {
	xmlFreeURI(uri);
	return;
307
    }
308 309 310 311 312 313 314
    
    ctxt->protocol = xmlMemStrdup(uri->scheme);
    ctxt->hostname = xmlMemStrdup(uri->server);
    if (uri->path != NULL)
	ctxt->path = xmlMemStrdup(uri->path);
    else
	ctxt->path = xmlMemStrdup("/");
315 316
    if (uri->query != NULL)
	ctxt->query = xmlMemStrdup(uri->query);
317 318 319 320
    if (uri->port != 0)
	ctxt->port = uri->port;

    xmlFreeURI(uri);
321 322 323 324 325 326 327 328 329 330 331 332 333 334
}

/**
 * xmlNanoHTTPScanProxy:
 * @URL:  The proxy URL used to initialize the proxy context
 *
 * (Re)Initialize the HTTP Proxy context by parsing the URL and finding
 * the protocol host port it indicates.
 * Should be like http://myproxy/ or http://myproxy:3128/
 * A NULL URL cleans up proxy informations.
 */

void
xmlNanoHTTPScanProxy(const char *URL) {
335
    xmlURIPtr uri;
336 337 338 339 340

    if (proxy != NULL) { 
        xmlFree(proxy);
	proxy = NULL;
    }
341 342
    proxyPort = 0;

343 344 345 346 347 348 349 350 351
#ifdef DEBUG_HTTP
    if (URL == NULL)
	xmlGenericError(xmlGenericErrorContext,
		"Removing HTTP proxy info\n");
    else
	xmlGenericError(xmlGenericErrorContext,
		"Using HTTP proxy %s\n", URL);
#endif
    if (URL == NULL) return;
352

353
    uri = xmlParseURIRaw(URL, 1);
354 355 356 357 358 359
    if ((uri == NULL) || (uri->scheme == NULL) ||
	(strcmp(uri->scheme, "http")) || (uri->server == NULL)) {
	__xmlIOErr(XML_FROM_HTTP, XML_HTTP_URL_SYNTAX, "Syntax Error\n");
	if (uri != NULL)
	    xmlFreeURI(uri);
	return;
360
    }
361 362 363 364 365 366
    
    proxy = xmlMemStrdup(uri->server);
    if (uri->port != 0)
	proxyPort = uri->port;

    xmlFreeURI(uri);
367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382
}

/**
 * xmlNanoHTTPNewCtxt:
 * @URL:  The URL used to initialize the context
 *
 * Allocate and initialize a new HTTP context.
 *
 * Returns an HTTP context or NULL in case of error.
 */

static xmlNanoHTTPCtxtPtr
xmlNanoHTTPNewCtxt(const char *URL) {
    xmlNanoHTTPCtxtPtr ret;

    ret = (xmlNanoHTTPCtxtPtr) xmlMalloc(sizeof(xmlNanoHTTPCtxt));
383 384 385 386
    if (ret == NULL) {
        xmlHTTPErrMemory("allocating context");
        return(NULL);
    }
387 388 389 390 391

    memset(ret, 0, sizeof(xmlNanoHTTPCtxt));
    ret->port = 80;
    ret->returnValue = 0;
    ret->fd = -1;
392
    ret->ContentLength = -1;
393

394
    xmlNanoHTTPScanURL(ret, URL);
395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411

    return(ret);
}

/**
 * xmlNanoHTTPFreeCtxt:
 * @ctxt:  an HTTP context
 *
 * Frees the context after closing the connection.
 */

static void
xmlNanoHTTPFreeCtxt(xmlNanoHTTPCtxtPtr ctxt) {
    if (ctxt == NULL) return;
    if (ctxt->hostname != NULL) xmlFree(ctxt->hostname);
    if (ctxt->protocol != NULL) xmlFree(ctxt->protocol);
    if (ctxt->path != NULL) xmlFree(ctxt->path);
412
    if (ctxt->query != NULL) xmlFree(ctxt->query);
413 414 415
    if (ctxt->out != NULL) xmlFree(ctxt->out);
    if (ctxt->in != NULL) xmlFree(ctxt->in);
    if (ctxt->contentType != NULL) xmlFree(ctxt->contentType);
416
    if (ctxt->encoding != NULL) xmlFree(ctxt->encoding);
417
    if (ctxt->mimeType != NULL) xmlFree(ctxt->mimeType);
418 419
    if (ctxt->location != NULL) xmlFree(ctxt->location);
    if (ctxt->authHeader != NULL) xmlFree(ctxt->authHeader);
420 421 422 423 424 425 426
#ifdef HAVE_ZLIB_H
    if (ctxt->strm != NULL) {
	inflateEnd(ctxt->strm);
	xmlFree(ctxt->strm);
    }
#endif

427 428 429 430 431 432 433 434 435 436 437
    ctxt->state = XML_NANO_HTTP_NONE;
    if (ctxt->fd >= 0) closesocket(ctxt->fd);
    ctxt->fd = -1;
    xmlFree(ctxt);
}

/**
 * xmlNanoHTTPSend:
 * @ctxt:  an HTTP context
 *
 * Send the input needed to initiate the processing on the server side
438
 * Returns number of bytes sent or -1 on error.
439 440
 */

441 442 443 444 445 446 447 448 449
static int
xmlNanoHTTPSend(xmlNanoHTTPCtxtPtr ctxt, const char * xmt_ptr, int outlen) {

    int 	total_sent = 0;

    if ( (ctxt->state & XML_NANO_HTTP_WRITE) && (xmt_ptr != NULL ) ) {
        while (total_sent < outlen) {
            int nsent = send(ctxt->fd, xmt_ptr + total_sent,
                                      outlen - total_sent, 0);
450 451
            if (nsent>0)
                total_sent += nsent;
452
	    else if ( ( nsent == -1 ) && 
453
#if defined(EAGAIN) && EAGAIN != EWOULDBLOCK
454
	    	      ( socket_errno( ) != EAGAIN ) &&
455
#endif
456
		        ( socket_errno( ) != EWOULDBLOCK ) ) {
457
		__xmlIOErr(XML_FROM_HTTP, 0, "send failed\n");
458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475
		if ( total_sent == 0 )
		    total_sent = -1;
		break;
	    }
	    else {
	        /*
		**  No data sent
		**  Since non-blocking sockets are used, wait for 
		**  socket to be writable or default timeout prior
		**  to retrying.
		*/

		struct timeval	tv;
		fd_set		wfd;

		tv.tv_sec = timeout;
		tv.tv_usec = 0;
		FD_ZERO( &wfd );
476 477 478 479
#ifdef _MSC_VER
#pragma warning(push)
#pragma warning(disable: 4018)
#endif
480
		FD_SET( ctxt->fd, &wfd );
481 482 483
#ifdef _MSC_VER
#pragma warning(pop)
#endif
484 485
		(void)select( ctxt->fd + 1, NULL, &wfd, NULL, &tv );
	    }
486
	}
487
    }
488 489

    return total_sent;
490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509
}

/**
 * xmlNanoHTTPRecv:
 * @ctxt:  an HTTP context
 *
 * Read information coming from the HTTP connection.
 * This is a blocking call (but it blocks in select(), not read()).
 *
 * Returns the number of byte read or -1 in case of error.
 */

static int
xmlNanoHTTPRecv(xmlNanoHTTPCtxtPtr ctxt) {
    fd_set rfd;
    struct timeval tv;


    while (ctxt->state & XML_NANO_HTTP_READ) {
	if (ctxt->in == NULL) {
510
	    ctxt->in = (char *) xmlMallocAtomic(65000 * sizeof(char));
511
	    if (ctxt->in == NULL) {
512
		xmlHTTPErrMemory("allocating input");
513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
	        ctxt->last = -1;
		return(-1);
	    }
	    ctxt->inlen = 65000;
	    ctxt->inptr = ctxt->content = ctxt->inrptr = ctxt->in;
	}
	if (ctxt->inrptr > ctxt->in + XML_NANO_HTTP_CHUNK) {
	    int delta = ctxt->inrptr - ctxt->in;
	    int len = ctxt->inptr - ctxt->inrptr;
	    
	    memmove(ctxt->in, ctxt->inrptr, len);
	    ctxt->inrptr -= delta;
	    ctxt->content -= delta;
	    ctxt->inptr -= delta;
	}
        if ((ctxt->in + ctxt->inlen) < (ctxt->inptr + XML_NANO_HTTP_CHUNK)) {
	    int d_inptr = ctxt->inptr - ctxt->in;
	    int d_content = ctxt->content - ctxt->in;
	    int d_inrptr = ctxt->inrptr - ctxt->in;
532
	    char *	tmp_ptr = ctxt->in;
533 534

	    ctxt->inlen *= 2;
535
            ctxt->in = (char *) xmlRealloc(tmp_ptr, ctxt->inlen);
536
	    if (ctxt->in == NULL) {
537
		xmlHTTPErrMemory("allocating input buffer");
538
		xmlFree( tmp_ptr );
539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561
	        ctxt->last = -1;
		return(-1);
	    }
            ctxt->inptr = ctxt->in + d_inptr;
            ctxt->content = ctxt->in + d_content;
            ctxt->inrptr = ctxt->in + d_inrptr;
	}
	ctxt->last = recv(ctxt->fd, ctxt->inptr, XML_NANO_HTTP_CHUNK, 0);
	if (ctxt->last > 0) {
	    ctxt->inptr += ctxt->last;
	    return(ctxt->last);
	}
	if (ctxt->last == 0) {
	    return(0);
	}
	if (ctxt->last == -1) {
	    switch (socket_errno()) {
		case EINPROGRESS:
		case EWOULDBLOCK:
#if defined(EAGAIN) && EAGAIN != EWOULDBLOCK
		case EAGAIN:
#endif
		    break;
562 563 564 565 566

		case ECONNRESET:
		case ESHUTDOWN:
		    return ( 0 );

567
		default:
568
		    __xmlIOErr(XML_FROM_HTTP, 0, "recv failed\n");
569
		    return(-1);
570 571 572 573 574 575
	    }
	}

	tv.tv_sec = timeout;
	tv.tv_usec = 0;
	FD_ZERO(&rfd);
576 577 578 579
#ifdef _MSC_VER
#pragma warning(push)
#pragma warning(disable: 4018)
#endif
580
	FD_SET(ctxt->fd, &rfd);
581 582 583
#ifdef _MSC_VER
#pragma warning(pop)
#endif
584
	
585 586 587 588 589
	if ( (select(ctxt->fd+1, &rfd, NULL, NULL, &tv)<1)
#if defined(EINTR)
		&& (errno != EINTR)
#endif
	)
590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609
		return(0);
    }
    return(0);
}

/**
 * xmlNanoHTTPReadLine:
 * @ctxt:  an HTTP context
 *
 * Read one line in the HTTP server output, usually for extracting
 * the HTTP protocol informations from the answer header.
 *
 * Returns a newly allocated string with a copy of the line, or NULL
 *         which indicate the end of the input.
 */

static char *
xmlNanoHTTPReadLine(xmlNanoHTTPCtxtPtr ctxt) {
    char buf[4096];
    char *bp = buf;
610
    int	rc;
611 612 613
    
    while (bp - buf < 4095) {
	if (ctxt->inrptr == ctxt->inptr) {
614
	    if ( (rc = xmlNanoHTTPRecv(ctxt)) == 0) {
615 616 617 618 619 620
		if (bp == buf)
		    return(NULL);
		else
		    *bp = 0;
		return(xmlMemStrdup(buf));
	    }
621 622 623
	    else if ( rc == -1 ) {
	        return ( NULL );
	    }
624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645
	}
	*bp = *ctxt->inrptr++;
	if (*bp == '\n') {
	    *bp = 0;
	    return(xmlMemStrdup(buf));
	}
	if (*bp != '\r')
	    bp++;
    }
    buf[4095] = 0;
    return(xmlMemStrdup(buf));
}


/**
 * xmlNanoHTTPScanAnswer:
 * @ctxt:  an HTTP context
 * @line:  an HTTP header line
 *
 * Try to extract useful informations from the server answer.
 * We currently parse and process:
 *  - The HTTP revision/ return code
646
 *  - The Content-Type, Mime-Type and charset used
647
 *  - The Location for redirect processing.
648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689
 *
 * Returns -1 in case of failure, the file descriptor number otherwise
 */

static void
xmlNanoHTTPScanAnswer(xmlNanoHTTPCtxtPtr ctxt, const char *line) {
    const char *cur = line;

    if (line == NULL) return;

    if (!strncmp(line, "HTTP/", 5)) {
        int version = 0;
	int ret = 0;

	cur += 5;
	while ((*cur >= '0') && (*cur <= '9')) {
	    version *= 10;
	    version += *cur - '0';
	    cur++;
	}
	if (*cur == '.') {
	    cur++;
	    if ((*cur >= '0') && (*cur <= '9')) {
		version *= 10;
		version += *cur - '0';
		cur++;
	    }
	    while ((*cur >= '0') && (*cur <= '9'))
		cur++;
	} else
	    version *= 10;
	if ((*cur != ' ') && (*cur != '\t')) return;
	while ((*cur == ' ') || (*cur == '\t')) cur++;
	if ((*cur < '0') || (*cur > '9')) return;
	while ((*cur >= '0') && (*cur <= '9')) {
	    ret *= 10;
	    ret += *cur - '0';
	    cur++;
	}
	if ((*cur != 0) && (*cur != ' ') && (*cur != '\t')) return;
	ctxt->returnValue = ret;
    } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"Content-Type:", 13)) {
690
        const xmlChar *charset, *last, *mime;
691 692 693 694 695
        cur += 13;
	while ((*cur == ' ') || (*cur == '\t')) cur++;
	if (ctxt->contentType != NULL)
	    xmlFree(ctxt->contentType);
	ctxt->contentType = xmlMemStrdup(cur);
696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714
	mime = (const xmlChar *) cur;
	last = mime;
	while ((*last != 0) && (*last != ' ') && (*last != '\t') &&
	       (*last != ';') && (*last != ','))
	    last++;
	if (ctxt->mimeType != NULL)
	    xmlFree(ctxt->mimeType);
	ctxt->mimeType = (char *) xmlStrndup(mime, last - mime);
	charset = xmlStrstr(BAD_CAST ctxt->contentType, BAD_CAST "charset=");
	if (charset != NULL) {
	    charset += 8;
	    last = charset;
	    while ((*last != 0) && (*last != ' ') && (*last != '\t') &&
	           (*last != ';') && (*last != ','))
		last++;
	    if (ctxt->encoding != NULL)
	        xmlFree(ctxt->encoding);
	    ctxt->encoding = (char *) xmlStrndup(charset, last - charset);
	}
715
    } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"ContentType:", 12)) {
716
        const xmlChar *charset, *last, *mime;
717 718 719 720
        cur += 12;
	if (ctxt->contentType != NULL) return;
	while ((*cur == ' ') || (*cur == '\t')) cur++;
	ctxt->contentType = xmlMemStrdup(cur);
721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739
	mime = (const xmlChar *) cur;
	last = mime;
	while ((*last != 0) && (*last != ' ') && (*last != '\t') &&
	       (*last != ';') && (*last != ','))
	    last++;
	if (ctxt->mimeType != NULL)
	    xmlFree(ctxt->mimeType);
	ctxt->mimeType = (char *) xmlStrndup(mime, last - mime);
	charset = xmlStrstr(BAD_CAST ctxt->contentType, BAD_CAST "charset=");
	if (charset != NULL) {
	    charset += 8;
	    last = charset;
	    while ((*last != 0) && (*last != ' ') && (*last != '\t') &&
	           (*last != ';') && (*last != ','))
		last++;
	    if (ctxt->encoding != NULL)
	        xmlFree(ctxt->encoding);
	    ctxt->encoding = (char *) xmlStrndup(charset, last - charset);
	}
740 741 742 743 744
    } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"Location:", 9)) {
        cur += 9;
	while ((*cur == ' ') || (*cur == '\t')) cur++;
	if (ctxt->location != NULL)
	    xmlFree(ctxt->location);
745 746 747 748 749 750 751 752 753
	if (*cur == '/') {
	    xmlChar *tmp_http = xmlStrdup(BAD_CAST "http://");
	    xmlChar *tmp_loc = 
	        xmlStrcat(tmp_http, (const xmlChar *) ctxt->hostname);
	    ctxt->location = 
	        (char *) xmlStrcat (tmp_loc, (const xmlChar *) cur);
	} else {
	    ctxt->location = xmlMemStrdup(cur);
	}
754 755 756 757 758 759 760 761 762 763 764 765
    } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"WWW-Authenticate:", 17)) {
        cur += 17;
	while ((*cur == ' ') || (*cur == '\t')) cur++;
	if (ctxt->authHeader != NULL)
	    xmlFree(ctxt->authHeader);
	ctxt->authHeader = xmlMemStrdup(cur);
    } else if (!xmlStrncasecmp(BAD_CAST line, BAD_CAST"Proxy-Authenticate:", 19)) {
        cur += 19;
	while ((*cur == ' ') || (*cur == '\t')) cur++;
	if (ctxt->authHeader != NULL)
	    xmlFree(ctxt->authHeader);
	ctxt->authHeader = xmlMemStrdup(cur);
766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785
#ifdef HAVE_ZLIB_H
    } else if ( !xmlStrncasecmp( BAD_CAST line, BAD_CAST"Content-Encoding:", 17) ) {
	cur += 17;
	while ((*cur == ' ') || (*cur == '\t')) cur++;
	if ( !xmlStrncasecmp( BAD_CAST cur, BAD_CAST"gzip", 4) ) {
	    ctxt->usesGzip = 1;

	    ctxt->strm = xmlMalloc(sizeof(z_stream));

	    if (ctxt->strm != NULL) {
		ctxt->strm->zalloc = Z_NULL;
		ctxt->strm->zfree = Z_NULL;
		ctxt->strm->opaque = Z_NULL;
		ctxt->strm->avail_in = 0;
		ctxt->strm->next_in = Z_NULL;

		inflateInit2( ctxt->strm, 31 );
	    }
	}
#endif
786 787 788
    } else if ( !xmlStrncasecmp( BAD_CAST line, BAD_CAST"Content-Length:", 15) ) {
	cur += 15;
	ctxt->ContentLength = strtol( cur, NULL, 10 );
789 790 791 792 793
    }
}

/**
 * xmlNanoHTTPConnectAttempt:
794
 * @addr:  a socket address structure
795 796 797 798 799 800 801 802 803
 *
 * Attempt a connection to the given IP:port endpoint. It forces
 * non-blocking semantic on the socket, and allow 60 seconds for
 * the host to answer.
 *
 * Returns -1 in case of failure, the file descriptor number otherwise
 */

static int
804
xmlNanoHTTPConnectAttempt(struct sockaddr *addr)
805 806
{
    fd_set wfd;
807 808 809
#ifdef _WINSOCKAPI_
    fd_set xfd;
#endif
810 811
    struct timeval tv;
    int status;
812 813
    int addrlen;
    SOCKET s;
814
    
815 816 817 818 819 820 821 822 823 824 825
#ifdef SUPPORT_IP6
    if (addr->sa_family == AF_INET6) {
	s = socket (PF_INET6, SOCK_STREAM, IPPROTO_TCP);
	addrlen = sizeof (struct sockaddr_in6);
    }
    else
#endif
    {
	s = socket (PF_INET, SOCK_STREAM, IPPROTO_TCP);
	addrlen = sizeof (struct sockaddr_in);
    }
826 827 828 829
    if (s==-1) {
#ifdef DEBUG_HTTP
	perror("socket");
#endif
830
	__xmlIOErr(XML_FROM_HTTP, 0, "socket failed\n");
831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846
	return(-1);
    }
    
#ifdef _WINSOCKAPI_
    {
	u_long one = 1;

	status = ioctlsocket(s, FIONBIO, &one) == SOCKET_ERROR ? -1 : 0;
    }
#else /* _WINSOCKAPI_ */
#if defined(VMS)
    {
	int enable = 1;
	status = ioctl(s, FIONBIO, &enable);
    }
#else /* VMS */
847 848 849 850 851 852
#if defined(__BEOS__)
	{
		bool noblock = true;
		status = setsockopt(s, SOL_SOCKET, SO_NONBLOCK, &noblock, sizeof(noblock));
	}
#else /* __BEOS__ */
853 854 855 856 857 858 859 860 861 862 863 864 865 866
    if ((status = fcntl(s, F_GETFL, 0)) != -1) {
#ifdef O_NONBLOCK
	status |= O_NONBLOCK;
#else /* O_NONBLOCK */
#ifdef F_NDELAY
	status |= F_NDELAY;
#endif /* F_NDELAY */
#endif /* !O_NONBLOCK */
	status = fcntl(s, F_SETFL, status);
    }
    if (status < 0) {
#ifdef DEBUG_HTTP
	perror("nonblocking");
#endif
867
	__xmlIOErr(XML_FROM_HTTP, 0, "error setting non-blocking IO\n");
868 869 870
	closesocket(s);
	return(-1);
    }
871
#endif /* !__BEOS__ */
872 873 874
#endif /* !VMS */
#endif /* !_WINSOCKAPI_ */

875
    if (connect (s, addr, addrlen) == -1) {
876 877 878 879 880
	switch (socket_errno()) {
	    case EINPROGRESS:
	    case EWOULDBLOCK:
		break;
	    default:
881
		__xmlIOErr(XML_FROM_HTTP, 0, "error connecting to HTTP server");
882 883 884 885 886 887 888
		closesocket(s);
		return(-1);
	}
    }	
    
    tv.tv_sec = timeout;
    tv.tv_usec = 0;
889 890 891 892 893

#ifdef _MSC_VER
#pragma warning(push)
#pragma warning(disable: 4018)
#endif
894 895
    FD_ZERO(&wfd);
    FD_SET(s, &wfd);
896 897 898 899

#ifdef _WINSOCKAPI_    
    FD_ZERO(&xfd);
    FD_SET(s, &xfd);
900
    
901 902
    switch(select(s+1, NULL, &wfd, &xfd, &tv))
#else
903
    switch(select(s+1, NULL, &wfd, NULL, &tv))
904 905 906
#endif
#ifdef _MSC_VER
#pragma warning(pop)
907
#endif
908 909 910
    {
	case 0:
	    /* Time out */
911
	    __xmlIOErr(XML_FROM_HTTP, 0, "Connect attempt timed out");
912 913 914 915
	    closesocket(s);
	    return(-1);
	case -1:
	    /* Ermm.. ?? */
916
	    __xmlIOErr(XML_FROM_HTTP, 0, "Connect failed");
917 918 919 920
	    closesocket(s);
	    return(-1);
    }

921 922 923 924 925
    if ( FD_ISSET(s, &wfd)
#ifdef _WINSOCKAPI_
                           || FD_ISSET(s, &xfd)
#endif
                                                ) {
926
	XML_SOCKLEN_T len;
927
	len = sizeof(status);
928
#ifdef SO_ERROR
929 930
	if (getsockopt(s, SOL_SOCKET, SO_ERROR, (char*)&status, &len) < 0 ) {
	    /* Solaris error code */
931
	    __xmlIOErr(XML_FROM_HTTP, 0, "getsockopt failed\n");
932 933
	    return (-1);
	}
934
#endif
935
	if ( status ) {
936
	    __xmlIOErr(XML_FROM_HTTP, 0, "Error connecting to remote host");
937 938 939 940 941 942
	    closesocket(s);
	    errno = status;
	    return (-1);
	}
    } else {
	/* pbm */
943
	__xmlIOErr(XML_FROM_HTTP, 0, "select failed\n");
944
	closesocket(s);
945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965
	return (-1);
    }
    
    return(s);
}
 
/**
 * xmlNanoHTTPConnectHost:
 * @host:  the host name
 * @port:  the port number
 *
 * Attempt a connection to the given host:port endpoint. It tries
 * the multiple IP provided by the DNS if available.
 *
 * Returns -1 in case of failure, the file descriptor number otherwise
 */

static int
xmlNanoHTTPConnectHost(const char *host, int port)
{
    struct hostent *h;
966
    struct sockaddr *addr = NULL;
967
    struct in_addr ia;
968
    struct sockaddr_in sockin;
969

970 971
#ifdef SUPPORT_IP6
    struct in6_addr ia6;
972
    struct sockaddr_in6 sockin6;
973 974 975
#endif
    int i;
    int s;
976

977 978 979
    memset (&sockin, 0, sizeof(sockin));
#ifdef SUPPORT_IP6
    memset (&sockin6, 0, sizeof(sockin6));
980 981 982
#endif

#if !defined(HAVE_GETADDRINFO) && defined(SUPPORT_IP6) && defined(RES_USE_INET6)
983
    if (have_ipv6 ())
984
    {
985 986 987 988
	if (!(_res.options & RES_INIT))
	    res_init();
	_res.options |= RES_USE_INET6;
    }
989 990 991 992 993 994
#endif

#if defined(HAVE_GETADDRINFO) && defined(SUPPORT_IP6) && !defined(_WIN32)
    if (have_ipv6 ())
#endif
#if defined(HAVE_GETADDRINFO) && (defined(SUPPORT_IP6) || defined(_WIN32))
995
    {
996 997 998 999 1000 1001 1002 1003 1004
	int status;
	struct addrinfo hints, *res, *result;

	result = NULL;
	memset (&hints, 0,sizeof(hints));
	hints.ai_socktype = SOCK_STREAM;

	status = getaddrinfo (host, NULL, &hints, &result);
	if (status) {
1005
	    __xmlIOErr(XML_FROM_HTTP, 0, "getaddrinfo failed\n");
1006 1007 1008 1009
	    return (-1);
	}

	for (res = result; res; res = res->ai_next) {
1010 1011 1012 1013 1014
	    if (res->ai_family == AF_INET) {
		if (res->ai_addrlen > sizeof(sockin)) {
		    __xmlIOErr(XML_FROM_HTTP, 0, "address size mismatch\n");
		    freeaddrinfo (result);
		    return (-1);
1015
		}
1016 1017 1018 1019 1020 1021 1022
		memcpy (&sockin, res->ai_addr, res->ai_addrlen);
		sockin.sin_port = htons (port);
		addr = (struct sockaddr *)&sockin;
#ifdef SUPPORT_IP6
	    } else if (have_ipv6 () && (res->ai_family == AF_INET6)) {
		if (res->ai_addrlen > sizeof(sockin6)) {
		    __xmlIOErr(XML_FROM_HTTP, 0, "address size mismatch\n");
1023
		    freeaddrinfo (result);
1024
		    return (-1);
1025
		}
1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036
		memcpy (&sockin6, res->ai_addr, res->ai_addrlen);
		sockin6.sin6_port = htons (port);
		addr = (struct sockaddr *)&sockin6;
#endif
	    } else
		continue;              /* for */

	    s = xmlNanoHTTPConnectAttempt (addr);
	    if (s != -1) {
		freeaddrinfo (result);
		return (s);
1037 1038
	    }
	}
1039

1040 1041
	if (result)
	    freeaddrinfo (result);
1042
    }
1043
#endif
1044 1045
#if defined(HAVE_GETADDRINFO) && defined(SUPPORT_IP6) && !defined(_WIN32)
    else
1046
#endif
1047 1048
#if !defined(HAVE_GETADDRINFO) || !defined(_WIN32)
    {
1049 1050
	h = gethostbyname (host);
	if (h == NULL) {
1051 1052 1053 1054 1055 1056 1057

/*
 * Okay, I got fed up by the non-portability of this error message
 * extraction code. it work on Linux, if it work on your platform
 * and one want to enable it, send me the defined(foobar) needed
 */
#if defined(HAVE_NETDB_H) && defined(HOST_NOT_FOUND) && defined(linux)
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083
	    const char *h_err_txt = "";

	    switch (h_errno) {
		case HOST_NOT_FOUND:
		    h_err_txt = "Authoritive host not found";
		    break;

		case TRY_AGAIN:
		    h_err_txt =
			"Non-authoritive host not found or server failure.";
		    break;

		case NO_RECOVERY:
		    h_err_txt =
			"Non-recoverable errors:  FORMERR, REFUSED, or NOTIMP.";
		    break;

		case NO_ADDRESS:
		    h_err_txt =
			"Valid name, no data record of requested type.";
		    break;

		default:
		    h_err_txt = "No error text defined.";
		    break;
	    }
1084
	    __xmlIOErr(XML_FROM_HTTP, 0, h_err_txt);
1085
#else
1086
	    __xmlIOErr(XML_FROM_HTTP, 0, "Failed to resolve host");
1087
#endif
1088 1089
	    return (-1);
	}
1090

1091 1092 1093
	for (i = 0; h->h_addr_list[i]; i++) {
	    if (h->h_addrtype == AF_INET) {
		/* A records (IPv4) */
1094 1095 1096 1097
		if ((unsigned int) h->h_length > sizeof(ia)) {
		    __xmlIOErr(XML_FROM_HTTP, 0, "address size mismatch\n");
		    return (-1);
		}
1098 1099 1100
		memcpy (&ia, h->h_addr_list[i], h->h_length);
		sockin.sin_family = h->h_addrtype;
		sockin.sin_addr = ia;
1101
		sockin.sin_port = (u_short)htons ((unsigned short)port);
1102
		addr = (struct sockaddr *) &sockin;
1103
#ifdef SUPPORT_IP6
1104 1105
	    } else if (have_ipv6 () && (h->h_addrtype == AF_INET6)) {
		/* AAAA records (IPv6) */
1106 1107 1108 1109
		if ((unsigned int) h->h_length > sizeof(ia6)) {
		    __xmlIOErr(XML_FROM_HTTP, 0, "address size mismatch\n");
		    return (-1);
		}
1110 1111 1112 1113 1114 1115 1116 1117
		memcpy (&ia6, h->h_addr_list[i], h->h_length);
		sockin6.sin6_family = h->h_addrtype;
		sockin6.sin6_addr = ia6;
		sockin6.sin6_port = htons (port);
		addr = (struct sockaddr *) &sockin6;
#endif
	    } else
		break;              /* for */
1118

1119 1120 1121 1122 1123
	    s = xmlNanoHTTPConnectAttempt (addr);
	    if (s != -1)
		return (s);
	}
    }
1124 1125
#endif

1126 1127
#ifdef DEBUG_HTTP
    xmlGenericError(xmlGenericErrorContext,
1128 1129
                    "xmlNanoHTTPConnectHost:  unable to connect to '%s'.\n",
                    host);
1130
#endif
1131
    return (-1);
1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150
}


/**
 * xmlNanoHTTPOpen:
 * @URL:  The URL to load
 * @contentType:  if available the Content-Type information will be
 *                returned at that location
 *
 * This function try to open a connection to the indicated resource
 * via HTTP GET.
 *
 * Returns NULL in case of failure, otherwise a request handler.
 *     The contentType, if provided must be freed by the caller
 */

void*
xmlNanoHTTPOpen(const char *URL, char **contentType) {
    if (contentType != NULL) *contentType = NULL;
1151
    return(xmlNanoHTTPMethod(URL, NULL, NULL, contentType, NULL, 0));
1152 1153 1154 1155 1156 1157 1158
}

/**
 * xmlNanoHTTPOpenRedir:
 * @URL:  The URL to load
 * @contentType:  if available the Content-Type information will be
 *                returned at that location
1159
 * @redir: if available the redirected URL will be returned
1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171
 *
 * This function try to open a connection to the indicated resource
 * via HTTP GET.
 *
 * Returns NULL in case of failure, otherwise a request handler.
 *     The contentType, if provided must be freed by the caller
 */

void*
xmlNanoHTTPOpenRedir(const char *URL, char **contentType, char **redir) {
    if (contentType != NULL) *contentType = NULL;
    if (redir != NULL) *redir = NULL;
1172
    return(xmlNanoHTTPMethodRedir(URL, NULL, NULL, contentType, redir, NULL,0));
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189
}

/**
 * xmlNanoHTTPRead:
 * @ctx:  the HTTP context
 * @dest:  a buffer
 * @len:  the buffer length
 *
 * This function tries to read @len bytes from the existing HTTP connection
 * and saves them in @dest. This is a blocking call.
 *
 * Returns the number of byte read. 0 is an indication of an end of connection.
 *         -1 indicates a parameter error.
 */
int
xmlNanoHTTPRead(void *ctx, void *dest, int len) {
    xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr) ctx;
1190 1191 1192 1193 1194
#ifdef HAVE_ZLIB_H
    int bytes_read = 0;
    int orig_avail_in;
    int z_ret;
#endif
1195 1196 1197 1198 1199

    if (ctx == NULL) return(-1);
    if (dest == NULL) return(-1);
    if (len <= 0) return(0);

1200 1201 1202 1203 1204 1205
#ifdef HAVE_ZLIB_H
    if (ctxt->usesGzip == 1) {
        if (ctxt->strm == NULL) return(0);
 
        ctxt->strm->next_out = dest;
        ctxt->strm->avail_out = len;
1206
	ctxt->strm->avail_in = ctxt->inptr - ctxt->inrptr;
1207

1208 1209
        while (ctxt->strm->avail_out > 0 &&
	       (ctxt->strm->avail_in > 0 || xmlNanoHTTPRecv(ctxt) > 0)) {
1210 1211
            orig_avail_in = ctxt->strm->avail_in =
			    ctxt->inptr - ctxt->inrptr - bytes_read;
1212 1213 1214 1215 1216 1217
            ctxt->strm->next_in = BAD_CAST (ctxt->inrptr + bytes_read);

            z_ret = inflate(ctxt->strm, Z_NO_FLUSH);
            bytes_read += orig_avail_in - ctxt->strm->avail_in;

            if (z_ret != Z_OK) break;
1218
	}
1219 1220 1221 1222 1223 1224

        ctxt->inrptr += bytes_read;
        return(len - ctxt->strm->avail_out);
    }
#endif

1225
    while (ctxt->inptr - ctxt->inrptr < len) {
1226
        if (xmlNanoHTTPRecv(ctxt) <= 0) break;
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251
    }
    if (ctxt->inptr - ctxt->inrptr < len)
        len = ctxt->inptr - ctxt->inrptr;
    memcpy(dest, ctxt->inrptr, len);
    ctxt->inrptr += len;
    return(len);
}

/**
 * xmlNanoHTTPClose:
 * @ctx:  the HTTP context
 *
 * This function closes an HTTP context, it ends up the connection and
 * free all data related to it.
 */
void
xmlNanoHTTPClose(void *ctx) {
    xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr) ctx;

    if (ctx == NULL) return;

    xmlNanoHTTPFreeCtxt(ctxt);
}

/**
1252
 * xmlNanoHTTPMethodRedir:
1253 1254 1255 1256
 * @URL:  The URL to load
 * @method:  the HTTP method to use
 * @input:  the input string if any
 * @contentType:  the Content-Type information IN and OUT
1257
 * @redir:  the redirected URL OUT
1258
 * @headers:  the extra headers
1259
 * @ilen:  input length
1260 1261 1262 1263 1264 1265
 *
 * This function try to open a connection to the indicated resource
 * via HTTP using the given @method, adding the given extra headers
 * and the input buffer for the request content.
 *
 * Returns NULL in case of failure, otherwise a request handler.
1266
 *     The contentType, or redir, if provided must be freed by the caller
1267 1268 1269
 */

void*
1270
xmlNanoHTTPMethodRedir(const char *URL, const char *method, const char *input,
1271 1272
                  char **contentType, char **redir,
		  const char *headers, int ilen ) {
1273 1274
    xmlNanoHTTPCtxtPtr ctxt;
    char *bp, *p;
1275
    int blen, ret;
1276 1277 1278
    int head;
    int nbRedirects = 0;
    char *redirURL = NULL;
William M. Brack's avatar
William M. Brack committed
1279 1280 1281
#ifdef DEBUG_HTTP
    int xmt_bytes;
#endif
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291
    
    if (URL == NULL) return(NULL);
    if (method == NULL) method = "GET";
    xmlNanoHTTPInit();

retry:
    if (redirURL == NULL)
	ctxt = xmlNanoHTTPNewCtxt(URL);
    else {
	ctxt = xmlNanoHTTPNewCtxt(redirURL);
1292
	ctxt->location = xmlMemStrdup(redirURL);
1293 1294
    }

1295 1296 1297 1298
    if ( ctxt == NULL ) {
	return ( NULL );
    }

1299
    if ((ctxt->protocol == NULL) || (strcmp(ctxt->protocol, "http"))) {
1300
	__xmlIOErr(XML_FROM_HTTP, XML_HTTP_URL_SYNTAX, "Not a valid HTTP URI");
1301 1302 1303 1304 1305
        xmlNanoHTTPFreeCtxt(ctxt);
	if (redirURL != NULL) xmlFree(redirURL);
        return(NULL);
    }
    if (ctxt->hostname == NULL) {
1306 1307
	__xmlIOErr(XML_FROM_HTTP, XML_HTTP_UNKNOWN_HOST,
	           "Failed to identify host in URI");
1308
        xmlNanoHTTPFreeCtxt(ctxt);
1309
	if (redirURL != NULL) xmlFree(redirURL);
1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321
        return(NULL);
    }
    if (proxy) {
	blen = strlen(ctxt->hostname) * 2 + 16;
	ret = xmlNanoHTTPConnectHost(proxy, proxyPort);
    }
    else {
	blen = strlen(ctxt->hostname);
	ret = xmlNanoHTTPConnectHost(ctxt->hostname, ctxt->port);
    }
    if (ret < 0) {
        xmlNanoHTTPFreeCtxt(ctxt);
1322
	if (redirURL != NULL) xmlFree(redirURL);
1323 1324 1325 1326
        return(NULL);
    }
    ctxt->fd = ret;

1327
    if (input == NULL)
1328
	ilen = 0;
1329 1330 1331
    else
	blen += 36;

1332
    if (headers != NULL)
1333
	blen += strlen(headers) + 2;
1334 1335
    if (contentType && *contentType)
	blen += strlen(*contentType) + 16;
1336 1337
    if (ctxt->query != NULL)
	blen += strlen(ctxt->query) + 1;
1338
    blen += strlen(method) + strlen(ctxt->path) + 24;
1339 1340 1341
#ifdef HAVE_ZLIB_H
    blen += 23;
#endif
1342
    bp = (char*)xmlMallocAtomic(blen);
1343 1344
    if ( bp == NULL ) {
        xmlNanoHTTPFreeCtxt( ctxt );
1345
	xmlHTTPErrMemory("allocating header buffer");
1346 1347 1348 1349 1350
	return ( NULL );
    }

    p = bp;

1351 1352
    if (proxy) {
	if (ctxt->port != 80) {
1353 1354
	    p += snprintf( p, blen - (p - bp), "%s http://%s:%d%s", 
			method, ctxt->hostname,
1355
		 	ctxt->port, ctxt->path );
1356
	}
1357 1358
	else 
	    p += snprintf( p, blen - (p - bp), "%s http://%s%s", method,
1359
	    		ctxt->hostname, ctxt->path);
1360 1361
    }
    else
1362
	p += snprintf( p, blen - (p - bp), "%s %s", method, ctxt->path);
1363

1364 1365 1366
    if (ctxt->query != NULL)
	p += snprintf( p, blen - (p - bp), "?%s", ctxt->query);

1367 1368
    if (ctxt->port == 80) {
        p += snprintf( p, blen - (p - bp), " HTTP/1.0\r\nHost: %s\r\n", 
1369
		    ctxt->hostname);
1370 1371 1372 1373
    } else {
        p += snprintf( p, blen - (p - bp), " HTTP/1.0\r\nHost: %s:%d\r\n",
		    ctxt->hostname, ctxt->port);
    }
1374

1375 1376 1377 1378
#ifdef HAVE_ZLIB_H
    p += snprintf(p, blen - (p - bp), "Accept-Encoding: gzip\r\n");
#endif

1379
    if (contentType != NULL && *contentType) 
1380
	p += snprintf(p, blen - (p - bp), "Content-Type: %s\r\n", *contentType);
1381 1382

    if (headers != NULL)
1383
	p += snprintf( p, blen - (p - bp), "%s", headers );
1384

1385
    if (input != NULL)
1386
	snprintf(p, blen - (p - bp), "Content-Length: %d\r\n\r\n", ilen );
1387
    else
1388
	snprintf(p, blen - (p - bp), "\r\n");
1389

1390 1391 1392 1393 1394 1395 1396 1397 1398
#ifdef DEBUG_HTTP
    xmlGenericError(xmlGenericErrorContext,
	    "-> %s%s", proxy? "(Proxy) " : "", bp);
    if ((blen -= strlen(bp)+1) < 0)
	xmlGenericError(xmlGenericErrorContext,
		"ERROR: overflowed buffer by %d bytes\n", -blen);
#endif
    ctxt->outptr = ctxt->out = bp;
    ctxt->state = XML_NANO_HTTP_WRITE;
1399 1400
    blen = strlen( ctxt->out );
#ifdef DEBUG_HTTP
William M. Brack's avatar
William M. Brack committed
1401
    xmt_bytes = xmlNanoHTTPSend(ctxt, ctxt->out, blen );
1402 1403 1404 1405 1406 1407
    if ( xmt_bytes != blen )
        xmlGenericError( xmlGenericErrorContext,
			"xmlNanoHTTPMethodRedir:  Only %d of %d %s %s\n",
			xmt_bytes, blen,
			"bytes of HTTP headers sent to host",
			ctxt->hostname );
William M. Brack's avatar
William M. Brack committed
1408 1409
#else
    xmlNanoHTTPSend(ctxt, ctxt->out, blen );
1410 1411 1412
#endif

    if ( input != NULL ) {
William M. Brack's avatar
William M. Brack committed
1413
#ifdef DEBUG_HTTP
1414 1415 1416 1417 1418 1419 1420 1421
        xmt_bytes = xmlNanoHTTPSend( ctxt, input, ilen );

	if ( xmt_bytes != ilen )
	    xmlGenericError( xmlGenericErrorContext,
	    		"xmlNanoHTTPMethodRedir:  Only %d of %d %s %s\n",
			xmt_bytes, ilen,
			"bytes of HTTP content sent to host",
			ctxt->hostname );
William M. Brack's avatar
William M. Brack committed
1422 1423
#else
	xmlNanoHTTPSend( ctxt, input, ilen );
1424 1425 1426
#endif
    }

1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450
    ctxt->state = XML_NANO_HTTP_READ;
    head = 1;

    while ((p = xmlNanoHTTPReadLine(ctxt)) != NULL) {
        if (head && (*p == 0)) {
	    head = 0;
	    ctxt->content = ctxt->inrptr;
	    xmlFree(p);
	    break;
	}
	xmlNanoHTTPScanAnswer(ctxt, p);

#ifdef DEBUG_HTTP
	xmlGenericError(xmlGenericErrorContext, "<- %s\n", p);
#endif
        xmlFree(p);
    }

    if ((ctxt->location != NULL) && (ctxt->returnValue >= 300) &&
        (ctxt->returnValue < 400)) {
#ifdef DEBUG_HTTP
	xmlGenericError(xmlGenericErrorContext,
		"\nRedirect to: %s\n", ctxt->location);
#endif
1451
	while ( xmlNanoHTTPRecv(ctxt) > 0 ) ;
1452 1453
        if (nbRedirects < XML_NANO_HTTP_MAX_REDIR) {
	    nbRedirects++;
1454 1455
	    if (redirURL != NULL)
		xmlFree(redirURL);
1456 1457 1458 1459 1460
	    redirURL = xmlMemStrdup(ctxt->location);
	    xmlNanoHTTPFreeCtxt(ctxt);
	    goto retry;
	}
	xmlNanoHTTPFreeCtxt(ctxt);
1461
	if (redirURL != NULL) xmlFree(redirURL);
1462 1463
#ifdef DEBUG_HTTP
	xmlGenericError(xmlGenericErrorContext,
1464
		"xmlNanoHTTPMethodRedir: Too many redirects, aborting ...\n");
1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
#endif
	return(NULL);
    }

    if (contentType != NULL) {
	if (ctxt->contentType != NULL)
	    *contentType = xmlMemStrdup(ctxt->contentType);
	else
	    *contentType = NULL;
    }

1476 1477 1478 1479 1480 1481 1482 1483 1484
    if ((redir != NULL) && (redirURL != NULL)) {
	*redir = redirURL;
    } else {
	if (redirURL != NULL)
	    xmlFree(redirURL);
	if (redir != NULL)
	    *redir = NULL;
    }

1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498
#ifdef DEBUG_HTTP
    if (ctxt->contentType != NULL)
	xmlGenericError(xmlGenericErrorContext,
		"\nCode %d, content-type '%s'\n\n",
	       ctxt->returnValue, ctxt->contentType);
    else
	xmlGenericError(xmlGenericErrorContext,
		"\nCode %d, no content-type\n\n",
	       ctxt->returnValue);
#endif

    return((void *) ctxt);
}

1499 1500 1501 1502 1503 1504 1505
/**
 * xmlNanoHTTPMethod:
 * @URL:  The URL to load
 * @method:  the HTTP method to use
 * @input:  the input string if any
 * @contentType:  the Content-Type information IN and OUT
 * @headers:  the extra headers
1506
 * @ilen:  input length
1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517
 *
 * This function try to open a connection to the indicated resource
 * via HTTP using the given @method, adding the given extra headers
 * and the input buffer for the request content.
 *
 * Returns NULL in case of failure, otherwise a request handler.
 *     The contentType, if provided must be freed by the caller
 */

void*
xmlNanoHTTPMethod(const char *URL, const char *method, const char *input,
1518
                  char **contentType, const char *headers, int ilen) {
1519
    return(xmlNanoHTTPMethodRedir(URL, method, input, contentType,
1520
		                  NULL, headers, ilen));
1521 1522
}

1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537
/**
 * xmlNanoHTTPFetch:
 * @URL:  The URL to load
 * @filename:  the filename where the content should be saved
 * @contentType:  if available the Content-Type information will be
 *                returned at that location
 *
 * This function try to fetch the indicated resource via HTTP GET
 * and save it's content in the file.
 *
 * Returns -1 in case of failure, 0 incase of success. The contentType,
 *     if provided must be freed by the caller
 */
int
xmlNanoHTTPFetch(const char *URL, const char *filename, char **contentType) {
1538 1539
    void *ctxt = NULL;
    char *buf = NULL;
1540 1541 1542
    int fd;
    int len;
    
1543
    if (filename == NULL) return(-1);
1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560
    ctxt = xmlNanoHTTPOpen(URL, contentType);
    if (ctxt == NULL) return(-1);

    if (!strcmp(filename, "-")) 
        fd = 0;
    else {
        fd = open(filename, O_CREAT | O_WRONLY, 00644);
	if (fd < 0) {
	    xmlNanoHTTPClose(ctxt);
	    if ((contentType != NULL) && (*contentType != NULL)) {
	        xmlFree(*contentType);
		*contentType = NULL;
	    }
	    return(-1);
	}
    }

1561 1562
    xmlNanoHTTPFetchContent( ctxt, &buf, &len );
    if ( len > 0 ) {
1563 1564 1565 1566 1567 1568 1569 1570
	write(fd, buf, len);
    }

    xmlNanoHTTPClose(ctxt);
    close(fd);
    return(0);
}

1571
#ifdef LIBXML_OUTPUT_ENABLED
1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583
/**
 * xmlNanoHTTPSave:
 * @ctxt:  the HTTP context
 * @filename:  the filename where the content should be saved
 *
 * This function saves the output of the HTTP transaction to a file
 * It closes and free the context at the end
 *
 * Returns -1 in case of failure, 0 incase of success.
 */
int
xmlNanoHTTPSave(void *ctxt, const char *filename) {
1584
    char *buf = NULL;
1585 1586 1587
    int fd;
    int len;
    
1588
    if ((ctxt == NULL) || (filename == NULL)) return(-1);
1589 1590 1591 1592

    if (!strcmp(filename, "-")) 
        fd = 0;
    else {
1593
        fd = open(filename, O_CREAT | O_WRONLY, 0666);
1594 1595 1596 1597 1598 1599
	if (fd < 0) {
	    xmlNanoHTTPClose(ctxt);
	    return(-1);
	}
    }

1600 1601
    xmlNanoHTTPFetchContent( ctxt, &buf, &len );
    if ( len > 0 ) {
1602 1603 1604 1605
	write(fd, buf, len);
    }

    xmlNanoHTTPClose(ctxt);
1606
    close(fd);
1607 1608
    return(0);
}
1609
#endif /* LIBXML_OUTPUT_ENABLED */
1610 1611 1612 1613 1614

/**
 * xmlNanoHTTPReturnCode:
 * @ctx:  the HTTP context
 *
1615 1616
 * Get the latest HTTP return code received
 *
1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631
 * Returns the HTTP return code for the request.
 */
int
xmlNanoHTTPReturnCode(void *ctx) {
    xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr) ctx;

    if (ctxt == NULL) return(-1);

    return(ctxt->returnValue);
}

/**
 * xmlNanoHTTPAuthHeader:
 * @ctx:  the HTTP context
 *
1632 1633
 * Get the authentication header of an HTTP context
 *
1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645
 * Returns the stashed value of the WWW-Authenticate or Proxy-Authenticate
 * header.
 */
const char *
xmlNanoHTTPAuthHeader(void *ctx) {
    xmlNanoHTTPCtxtPtr ctxt = (xmlNanoHTTPCtxtPtr) ctx;

    if (ctxt == NULL) return(NULL);

    return(ctxt->authHeader);
}

1646
/**
1647
 * xmlNanoHTTPContentLength:
1648 1649
 * @ctx:  the HTTP context
 *
1650 1651
 * Provides the specified content length from the HTTP header.
 *
1652 1653 1654 1655 1656 1657
 * Return the specified content length from the HTTP header.  Note that
 * a value of -1 indicates that the content length element was not included in
 * the response header.
 */
int
xmlNanoHTTPContentLength( void *