http.c revision 63069 
1/*-
2 * Copyright (c) 2000 Dag-Erling Co�dan Sm�rgrav
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer
10 *    in this position and unchanged.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 *    notice, this list of conditions and the following disclaimer in the
13 *    documentation and/or other materials provided with the distribution.
14 * 3. The name of the author may not be used to endorse or promote products
15 *    derived from this software without specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
18 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
19 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
20 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
21 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
22 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
23 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
24 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
26 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27 *
28 *      $FreeBSD: head/lib/libfetch/http.c 63069 2000-07-13 09:13:58Z des $
29 */
30
31#include <sys/param.h>
32#include <sys/socket.h>
33
34#include <ctype.h>
35#include <err.h>
36#include <errno.h>
37#include <locale.h>
38#include <netdb.h>
39#include <stdarg.h>
40#include <stdio.h>
41#include <stdlib.h>
42#include <string.h>
43#include <time.h>
44#include <unistd.h>
45
46#include "fetch.h"
47#include "common.h"
48#include "httperr.h"
49
50extern char *__progname; /* XXX not portable */
51
52/* Maximum number of redirects to follow */
53#define MAX_REDIRECT 5
54
55/* Symbolic names for reply codes we care about */
56#define HTTP_OK			200
57#define HTTP_PARTIAL		206
58#define HTTP_MOVED_PERM		301
59#define HTTP_MOVED_TEMP		302
60#define HTTP_SEE_OTHER		303
61#define HTTP_NEED_AUTH		401
62#define HTTP_NEED_PROXY_AUTH	403
63#define HTTP_PROTOCOL_ERROR	999
64
65#define HTTP_REDIRECT(xyz) ((xyz) == HTTP_MOVED_PERM \
66                            || (xyz) == HTTP_MOVED_TEMP \
67                            || (xyz) == HTTP_SEE_OTHER)
68
69
70
71/*****************************************************************************
72 * I/O functions for decoding chunked streams
73 */
74
75struct cookie
76{
77    int		 fd;
78    char	*buf;
79    size_t	 b_size;
80    size_t	 b_len;
81    int		 b_pos;
82    int		 eof;
83    int		 error;
84    long	 chunksize;
85#ifdef DEBUG
86    long	 total;
87#endif
88};
89
90/*
91 * Get next chunk header
92 */
93static int
94_http_new_chunk(struct cookie *c)
95{
96    char *p;
97
98    if (_fetch_getln(c->fd, &c->buf, &c->b_size, &c->b_len) == -1)
99	return -1;
100
101    if (c->b_len < 2 || !ishexnumber(*c->buf))
102	return -1;
103
104    for (p = c->buf; !isspace(*p) && *p != ';' && p < c->buf + c->b_len; ++p)
105	if (!ishexnumber(*p))
106	    return -1;
107	else if (isdigit(*p))
108	    c->chunksize = c->chunksize * 16 + *p - '0';
109	else
110	    c->chunksize = c->chunksize * 16 + 10 + tolower(*p) - 'a';
111
112#ifdef DEBUG
113    c->total += c->chunksize;
114    if (c->chunksize == 0)
115	fprintf(stderr, "\033[1m_http_fillbuf(): "
116		"end of last chunk\033[m\n");
117    else
118	fprintf(stderr, "\033[1m_http_fillbuf(): "
119		"new chunk: %ld (%ld)\033[m\n", c->chunksize, c->total);
120#endif
121
122    return c->chunksize;
123}
124
125/*
126 * Fill the input buffer, do chunk decoding on the fly
127 */
128static int
129_http_fillbuf(struct cookie *c)
130{
131    if (c->error)
132	return -1;
133    if (c->eof)
134	return 0;
135
136    if (c->chunksize == 0) {
137	switch (_http_new_chunk(c)) {
138	case -1:
139	    c->error = 1;
140	    return -1;
141	case 0:
142	    c->eof = 1;
143	    return 0;
144	}
145    }
146
147    if (c->b_size < c->chunksize) {
148	char *tmp;
149
150	if ((tmp = realloc(c->buf, c->chunksize)) == NULL)
151	    return -1;
152	c->buf = tmp;
153	c->b_size = c->chunksize;
154    }
155
156    if ((c->b_len = read(c->fd, c->buf, c->chunksize)) == -1)
157	return -1;
158    c->chunksize -= c->b_len;
159
160    if (c->chunksize == 0) {
161	char endl[2];
162	read(c->fd, endl, 2);
163    }
164
165    c->b_pos = 0;
166
167    return c->b_len;
168}
169
170/*
171 * Read function
172 */
173static int
174_http_readfn(void *v, char *buf, int len)
175{
176    struct cookie *c = (struct cookie *)v;
177    int l, pos;
178
179    if (c->error)
180	return -1;
181    if (c->eof)
182	return 0;
183
184    for (pos = 0; len > 0; pos += l, len -= l) {
185	/* empty buffer */
186	if (!c->buf || c->b_pos == c->b_len)
187	    if (_http_fillbuf(c) < 1)
188		break;
189	l = c->b_len - c->b_pos;
190	if (len < l)
191	    l = len;
192	bcopy(c->buf + c->b_pos, buf + pos, l);
193	c->b_pos += l;
194    }
195
196    if (!pos && c->error)
197	return -1;
198    return pos;
199}
200
201/*
202 * Write function
203 */
204static int
205_http_writefn(void *v, const char *buf, int len)
206{
207    struct cookie *c = (struct cookie *)v;
208
209    return write(c->fd, buf, len);
210}
211
212/*
213 * Close function
214 */
215static int
216_http_closefn(void *v)
217{
218    struct cookie *c = (struct cookie *)v;
219    int r;
220
221    r = close(c->fd);
222    if (c->buf)
223	free(c->buf);
224    free(c);
225    return r;
226}
227
228/*
229 * Wrap a file descriptor up
230 */
231static FILE *
232_http_funopen(int fd)
233{
234    struct cookie *c;
235    FILE *f;
236
237    if ((c = calloc(1, sizeof *c)) == NULL) {
238	_fetch_syserr();
239	return NULL;
240    }
241    c->fd = fd;
242    if (!(f = funopen(c, _http_readfn, _http_writefn, NULL, _http_closefn))) {
243	_fetch_syserr();
244	free(c);
245	return NULL;
246    }
247    return f;
248}
249
250
251/*****************************************************************************
252 * Helper functions for talking to the server and parsing its replies
253 */
254
255/* Header types */
256typedef enum {
257    hdr_syserror = -2,
258    hdr_error = -1,
259    hdr_end = 0,
260    hdr_unknown = 1,
261    hdr_content_length,
262    hdr_content_range,
263    hdr_last_modified,
264    hdr_location,
265    hdr_transfer_encoding
266} hdr;
267
268/* Names of interesting headers */
269static struct {
270    hdr		 num;
271    char	*name;
272} hdr_names[] = {
273    { hdr_content_length,	"Content-Length" },
274    { hdr_content_range,	"Content-Range" },
275    { hdr_last_modified,	"Last-Modified" },
276    { hdr_location,		"Location" },
277    { hdr_transfer_encoding,	"Transfer-Encoding" },
278    { hdr_unknown,		NULL },
279};
280
281static char	*reply_buf;
282static size_t	 reply_size;
283static size_t	 reply_length;
284
285/*
286 * Send a formatted line; optionally echo to terminal
287 */
288static int
289_http_cmd(int fd, char *fmt, ...)
290{
291    va_list ap;
292    size_t len;
293    char *msg;
294    int r;
295
296    va_start(ap, fmt);
297    len = vasprintf(&msg, fmt, ap);
298    va_end(ap);
299
300    if (msg == NULL) {
301	errno = ENOMEM;
302	_fetch_syserr();
303	return -1;
304    }
305
306    r = _fetch_putln(fd, msg, len);
307    free(msg);
308
309    if (r == -1) {
310	_fetch_syserr();
311	return -1;
312    }
313
314    return 0;
315}
316
317/*
318 * Get and parse status line
319 */
320static int
321_http_get_reply(int fd)
322{
323    if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
324	return -1;
325    /*
326     * A valid status line looks like "HTTP/m.n xyz reason" where m
327     * and n are the major and minor protocol version numbers and xyz
328     * is the reply code.
329     * We grok HTTP 1.0 and 1.1, so m must be 1 and n must be 0 or 1.
330     * We don't care about the reason phrase.
331     */
332    if (strncmp(reply_buf, "HTTP/1.", 7) != 0
333	|| (reply_buf[7] != '0' && reply_buf[7] != '1') || reply_buf[8] != ' '
334	|| !isdigit(reply_buf[9])
335	|| !isdigit(reply_buf[10])
336	|| !isdigit(reply_buf[11]))
337	return HTTP_PROTOCOL_ERROR;
338
339    return ((reply_buf[9] - '0') * 100
340	    + (reply_buf[10] - '0') * 10
341	    + (reply_buf[11] - '0'));
342}
343
344/*
345 * Check a header; if the type matches the given string, return a
346 * pointer to the beginning of the value.
347 */
348static char *
349_http_match(char *str, char *hdr)
350{
351    while (*str && *hdr && tolower(*str++) == tolower(*hdr++))
352	/* nothing */;
353    if (*str || *hdr != ':')
354	return NULL;
355    while (*hdr && isspace(*++hdr))
356	/* nothing */;
357    return hdr;
358}
359
360/*
361 * Get the next header and return the appropriate symbolic code.
362 */
363static hdr
364_http_next_header(int fd, char **p)
365{
366    int i;
367
368    if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
369	return hdr_syserror;
370    while (reply_length && isspace(reply_buf[reply_length-1]))
371	reply_length--;
372    reply_buf[reply_length] = 0;
373    if (reply_length == 0)
374	return hdr_end;
375    /*
376     * We could check for malformed headers but we don't really care.
377     * A valid header starts with a token immediately followed by a
378     * colon; a token is any sequence of non-control, non-whitespace
379     * characters except "()<>@,;:\\\"{}".
380     */
381    for (i = 0; hdr_names[i].num != hdr_unknown; i++)
382	if ((*p = _http_match(hdr_names[i].name, reply_buf)) != NULL)
383	    return hdr_names[i].num;
384    return hdr_unknown;
385}
386
387/*
388 * Parse a last-modified header
389 */
390static time_t
391_http_parse_mtime(char *p)
392{
393    char locale[64];
394    struct tm tm;
395
396    strncpy(locale, setlocale(LC_TIME, NULL), sizeof locale);
397    setlocale(LC_TIME, "C");
398    strptime(p, "%a, %d %b %Y %H:%M:%S GMT", &tm);
399    /* XXX should add support for date-2 and date-3 */
400    setlocale(LC_TIME, locale);
401    DEBUG(fprintf(stderr, "last modified: [\033[1m%04d-%02d-%02d "
402		  "%02d:%02d:%02d\033[m]\n",
403		  tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday,
404		  tm.tm_hour, tm.tm_min, tm.tm_sec));
405    return timegm(&tm);
406}
407
408/*
409 * Parse a content-length header
410 */
411static off_t
412_http_parse_length(char *p)
413{
414    off_t len;
415
416    for (len = 0; *p && isdigit(*p); ++p)
417	len = len * 10 + (*p - '0');
418    DEBUG(fprintf(stderr, "content length: [\033[1m%lld\033[m]\n", len));
419    return len;
420}
421
422/*
423 * Parse a content-range header
424 */
425static off_t
426_http_parse_range(char *p)
427{
428    off_t off;
429
430    if (strncasecmp(p, "bytes ", 6) != 0)
431	return -1;
432    for (p += 6, off = 0; *p && isdigit(*p); ++p)
433	off = off * 10 + *p - '0';
434    if (*p != '-')
435	return -1;
436    DEBUG(fprintf(stderr, "content range: [\033[1m%lld-\033[m]\n", off));
437    return off;
438}
439
440
441/*****************************************************************************
442 * Helper functions for authorization
443 */
444
445/*
446 * Base64 encoding
447 */
448static char *
449_http_base64(char *src)
450{
451    static const char base64[] =
452	"ABCDEFGHIJKLMNOPQRSTUVWXYZ"
453	"abcdefghijklmnopqrstuvwxyz"
454	"0123456789+/";
455    char *str, *dst;
456    size_t l;
457    int t, r;
458
459    l = strlen(src);
460    if ((str = malloc(((l + 2) / 3) * 4)) == NULL)
461	return NULL;
462    dst = str;
463    r = 0;
464
465    while (l >= 3) {
466	t = (src[0] << 16) | (src[1] << 8) | src[2];
467	dst[0] = base64[(t >> 18) & 0x3f];
468	dst[1] = base64[(t >> 12) & 0x3f];
469	dst[2] = base64[(t >> 6) & 0x3f];
470	dst[3] = base64[(t >> 0) & 0x3f];
471	src += 3; l -= 3;
472	dst += 4; r += 4;
473    }
474
475    switch (l) {
476    case 2:
477	t = (src[0] << 16) | (src[1] << 8);
478	dst[0] = base64[(t >> 18) & 0x3f];
479	dst[1] = base64[(t >> 12) & 0x3f];
480	dst[2] = base64[(t >> 6) & 0x3f];
481	dst[3] = '=';
482	dst += 4;
483	r += 4;
484	break;
485    case 1:
486	t = src[0] << 16;
487	dst[0] = base64[(t >> 18) & 0x3f];
488	dst[1] = base64[(t >> 12) & 0x3f];
489	dst[2] = dst[3] = '=';
490	dst += 4;
491	r += 4;
492	break;
493    case 0:
494	break;
495    }
496
497    *dst = 0;
498    return str;
499}
500
501/*
502 * Encode username and password
503 */
504static int
505_http_basic_auth(int fd, char *hdr, char *usr, char *pwd)
506{
507    char *upw, *auth;
508    int r;
509
510    if (asprintf(&upw, "%s:%s", usr, pwd) == -1)
511	return -1;
512    auth = _http_base64(upw);
513    free(upw);
514    if (auth == NULL)
515	return -1;
516    r = _http_cmd(fd, "%s: Basic %s", hdr, auth);
517    free(auth);
518    return r;
519}
520
521/*
522 * Send an authorization header
523 */
524static int
525_http_authorize(int fd, char *hdr, char *p)
526{
527    /* basic authorization */
528    if (strncasecmp(p, "basic:", 6) == 0) {
529	char *user, *pwd, *str;
530	int r;
531
532	/* skip realm */
533	for (p += 6; *p && *p != ':'; ++p)
534	    /* nothing */ ;
535	if (!*p || strchr(++p, ':') == NULL)
536	    return -1;
537	if ((str = strdup(p)) == NULL)
538	    return -1; /* XXX */
539	user = str;
540	pwd = strchr(str, ':');
541	*pwd++ = '\0';
542	r = _http_basic_auth(fd, hdr, user, pwd);
543	free(str);
544	return r;
545    }
546    return -1;
547}
548
549
550/*****************************************************************************
551 * Helper functions for connecting to a server or proxy
552 */
553
554/*
555 * Connect to the specified HTTP proxy server.
556 */
557static int
558_http_proxy_connect(char *proxy, int af, int verbose)
559{
560    char *hostname, *p;
561    int fd, port;
562
563    /* get hostname */
564    hostname = NULL;
565#ifdef INET6
566    /* host part can be an IPv6 address enclosed in square brackets */
567    if (*proxy == '[') {
568	if ((p = strchr(proxy, ']')) == NULL) {
569	    /* no terminating bracket */
570	    /* XXX should set an error code */
571	    goto ouch;
572	}
573	if (p[1] != '\0' && p[1] != ':') {
574	    /* garbage after address */
575	    /* XXX should set an error code */
576	    goto ouch;
577	}
578	if ((hostname = malloc(p - proxy)) == NULL) {
579	    errno = ENOMEM;
580	    _fetch_syserr();
581	    goto ouch;
582	}
583	strncpy(hostname, proxy + 1, p - proxy - 1);
584	hostname[p - proxy - 1] = '\0';
585	++p;
586    } else {
587#endif /* INET6 */
588	if ((p = strchr(proxy, ':')) == NULL)
589	    p = strchr(proxy, '\0');
590	if ((hostname = malloc(p - proxy + 1)) == NULL) {
591	    errno = ENOMEM;
592	    _fetch_syserr();
593	    goto ouch;
594	}
595	strncpy(hostname, proxy, p - proxy);
596	hostname[p - proxy] = '\0';
597#ifdef INET6
598    }
599#endif /* INET6 */
600    DEBUG(fprintf(stderr, "proxy name: [%s]\n", hostname));
601
602    /* get port number */
603    port = 0;
604    if (*p == ':') {
605	++p;
606	if (strspn(p, "0123456789") != strlen(p) || strlen(p) > 5) {
607	    /* port number is non-numeric or too long */
608	    /* XXX should set an error code */
609	    goto ouch;
610	}
611	port = atoi(p);
612	if (port < 1 || port > 65535) {
613	    /* port number is out of range */
614	    /* XXX should set an error code */
615	    goto ouch;
616	}
617    }
618
619    if (!port) {
620#if 0
621	/*
622	 * commented out, since there is currently no service name
623	 * for HTTP proxies
624	 */
625	struct servent *se;
626
627	if ((se = getservbyname("xxxx", "tcp")) != NULL)
628	    port = ntohs(se->s_port);
629	else
630#endif
631	    port = 3128;
632    }
633    DEBUG(fprintf(stderr, "proxy port: %d\n", port));
634
635    /* connect */
636    if ((fd = _fetch_connect(hostname, port, af, verbose)) == -1)
637	_fetch_syserr();
638    return fd;
639
640 ouch:
641    if (hostname)
642	free(hostname);
643    return -1;
644}
645
646/*
647 * Connect to the correct HTTP server or proxy.
648 */
649static int
650_http_connect(struct url *URL, int *proxy, char *flags)
651{
652    int direct, verbose;
653    int af, fd;
654    char *p;
655
656#ifdef INET6
657    af = AF_UNSPEC;
658#else
659    af = AF_INET;
660#endif
661
662    direct = (flags && strchr(flags, 'd'));
663    verbose = (flags && strchr(flags, 'v'));
664    if (flags && strchr(flags, '4'))
665	af = AF_INET;
666    else if (flags && strchr(flags, '6'))
667	af = AF_INET6;
668
669    /* check port */
670    if (!URL->port) {
671	struct servent *se;
672
673	/* Scheme can be ftp if we're using a proxy */
674	if (strcasecmp(URL->scheme, "ftp") == 0)
675	    if ((se = getservbyname("ftp", "tcp")) != NULL)
676		URL->port = ntohs(se->s_port);
677	    else
678		URL->port = 21;
679	else
680	    if ((se = getservbyname("http", "tcp")) != NULL)
681		URL->port = ntohs(se->s_port);
682	    else
683		URL->port = 80;
684    }
685
686    if (!direct && (p = getenv("HTTP_PROXY")) != NULL) {
687	/* attempt to connect to proxy server */
688	if ((fd = _http_proxy_connect(p, af, verbose)) == -1)
689	    return -1;
690	*proxy = 1;
691    } else {
692	/* if no proxy is configured, try direct */
693	if (strcasecmp(URL->scheme, "ftp") == 0) {
694	    /* can't talk http to an ftp server */
695	    /* XXX should set an error code */
696	    return -1;
697	}
698	if ((fd = _fetch_connect(URL->host, URL->port, af, verbose)) == -1)
699	    /* _fetch_connect() has already set an error code */
700	    return -1;
701	*proxy = 0;
702    }
703
704    return fd;
705}
706
707
708/*****************************************************************************
709 * Core
710 */
711
712/*
713 * Send a request and process the reply
714 */
715static FILE *
716_http_request(struct url *URL, char *op, struct url_stat *us, char *flags)
717{
718    struct url *url, *new;
719    int chunked, need_auth, noredirect, proxy, verbose;
720    int code, fd, i, n;
721    off_t offset;
722    char *p;
723    FILE *f;
724    hdr h;
725    char *host;
726#ifdef INET6
727    char hbuf[MAXHOSTNAMELEN + 1];
728#endif
729
730    noredirect = (flags && strchr(flags, 'A'));
731    verbose = (flags && strchr(flags, 'v'));
732
733    n = noredirect ? 1 : MAX_REDIRECT;
734
735    /* just to appease compiler warnings */
736    code = HTTP_PROTOCOL_ERROR;
737    chunked = 0;
738    offset = 0;
739    fd = -1;
740
741    for (url = URL, i = 0; i < n; ++i) {
742	new = NULL;
743	us->size = -1;
744	us->atime = us->mtime = 0;
745	chunked = 0;
746	need_auth = 0;
747	offset = 0;
748	fd = -1;
749    retry:
750	/* connect to server or proxy */
751	if ((fd = _http_connect(url, &proxy, flags)) == -1)
752	    goto ouch;
753
754	host = url->host;
755#ifdef INET6
756	if (strchr(url->host, ':')) {
757	    snprintf(hbuf, sizeof(hbuf), "[%s]", url->host);
758	    host = hbuf;
759	}
760#endif
761
762	/* send request */
763	if (verbose)
764	    _fetch_info("requesting %s://%s:%d%s",
765			url->scheme, host, url->port, url->doc);
766	if (proxy) {
767	    _http_cmd(fd, "%s %s://%s:%d%s HTTP/1.1",
768		      op, url->scheme, host, url->port, url->doc);
769	} else {
770	    _http_cmd(fd, "%s %s HTTP/1.1",
771		      op, url->doc);
772	}
773
774	/* proxy authorization */
775	if (proxy && (p = getenv("HTTP_PROXY_AUTH")) != NULL)
776	    _http_authorize(fd, "Proxy-Authorization", p);
777
778	/* server authorization */
779	if (need_auth) {
780	    if (*url->user || *url->pwd)
781		_http_basic_auth(fd, "Authorization",
782				 url->user ? url->user : "",
783				 url->pwd ? url->pwd : "");
784	    else if ((p = getenv("HTTP_AUTH")) != NULL)
785		_http_authorize(fd, "Authorization", p);
786	    else {
787		_http_seterr(HTTP_NEED_AUTH);
788		goto ouch;
789	    }
790	}
791
792	/* other headers */
793	_http_cmd(fd, "Host: %s:%d", host, url->port);
794	_http_cmd(fd, "User-Agent: %s " _LIBFETCH_VER, __progname);
795	if (URL->offset)
796	    _http_cmd(fd, "Range: bytes=%lld-", url->offset);
797	_http_cmd(fd, "Connection: close");
798	_http_cmd(fd, "");
799
800	/* get reply */
801	switch ((code = _http_get_reply(fd))) {
802	case HTTP_OK:
803	case HTTP_PARTIAL:
804	    /* fine */
805	    break;
806	case HTTP_MOVED_PERM:
807	case HTTP_MOVED_TEMP:
808	    /*
809	     * Not so fine, but we still have to read the headers to
810	     * get the new location.
811	     */
812	    break;
813	case HTTP_NEED_AUTH:
814	    if (need_auth) {
815		/*
816		 * We already sent out authorization code, so there's
817		 * nothing more we can do.
818		 */
819		_http_seterr(code);
820		goto ouch;
821	    }
822	    /* try again, but send the password this time */
823	    if (verbose)
824		_fetch_info("server requires authorization");
825	    need_auth = 1;
826	    close(fd);
827	    goto retry;
828	case HTTP_NEED_PROXY_AUTH:
829	    /*
830	     * If we're talking to a proxy, we already sent our proxy
831	     * authorization code, so there's nothing more we can do.
832	     */
833	    _http_seterr(code);
834	    goto ouch;
835	case HTTP_PROTOCOL_ERROR:
836	    /* fall through */
837	case -1:
838	    _fetch_syserr();
839	    goto ouch;
840	default:
841	    _http_seterr(code);
842	    goto ouch;
843	}
844
845	/* get headers */
846	do {
847	    switch ((h = _http_next_header(fd, &p))) {
848	    case hdr_syserror:
849		_fetch_syserr();
850		goto ouch;
851	    case hdr_error:
852		_http_seterr(HTTP_PROTOCOL_ERROR);
853		goto ouch;
854	    case hdr_content_length:
855		us->size = _http_parse_length(p);
856		break;
857	    case hdr_content_range:
858		offset = _http_parse_range(p);
859		break;
860	    case hdr_last_modified:
861		us->atime = us->mtime = _http_parse_mtime(p);
862		break;
863	    case hdr_location:
864		if (!HTTP_REDIRECT(code))
865		    break;
866		if (new)
867		    free(new);
868		if (verbose)
869		    _fetch_info("%d redirect to %s", code, p);
870		if (*p == '/')
871		    /* absolute path */
872		    new = fetchMakeURL(url->scheme, url->host, url->port, p,
873				       url->user, url->pwd);
874		else
875		    new = fetchParseURL(p);
876		if (new == NULL) {
877		    /* XXX should set an error code */
878		    DEBUG(fprintf(stderr, "failed to parse new URL\n"));
879		    goto ouch;
880		}
881		if (!*new->user && !*new->pwd) {
882		    strcpy(new->user, url->user);
883		    strcpy(new->pwd, url->pwd);
884		}
885		new->offset = url->offset;
886		new->length = url->length;
887		break;
888	    case hdr_transfer_encoding:
889		/* XXX weak test*/
890		chunked = (strcasecmp(p, "chunked") == 0);
891		break;
892	    case hdr_end:
893		/* fall through */
894	    case hdr_unknown:
895		/* ignore */
896		break;
897	    }
898	} while (h > hdr_end);
899
900	/* we either have a hit, or a redirect with no Location: header */
901	if (code == HTTP_OK || code == HTTP_PARTIAL || !new)
902	    break;
903
904	/* we have a redirect */
905	close(fd);
906	if (url != URL)
907	    fetchFreeURL(url);
908	url = new;
909    }
910
911    /* no success */
912    if (fd == -1) {
913	_http_seterr(code);
914	goto ouch;
915    }
916
917    /* wrap it up in a FILE */
918    if ((f = chunked ? _http_funopen(fd) : fdopen(fd, "r")) == NULL) {
919	_fetch_syserr();
920	goto ouch;
921    }
922
923    while (offset++ < url->offset)
924	if (fgetc(f) == EOF) {
925	    _fetch_syserr();
926	    fclose(f);
927	    f = NULL;
928	}
929
930    if (url != URL)
931	fetchFreeURL(url);
932
933    return f;
934
935 ouch:
936    if (url != URL)
937	fetchFreeURL(url);
938    if (fd != -1)
939	close(fd);
940    return NULL;
941}
942
943
944/*****************************************************************************
945 * Entry points
946 */
947
948/*
949 * Retrieve a file by HTTP
950 */
951FILE *
952fetchGetHTTP(struct url *URL, char *flags)
953{
954    struct url_stat us;
955
956    return _http_request(URL, "GET", &us, flags);
957}
958
959FILE *
960fetchPutHTTP(struct url *URL, char *flags)
961{
962    warnx("fetchPutHTTP(): not implemented");
963    return NULL;
964}
965
966/*
967 * Get an HTTP document's metadata
968 */
969int
970fetchStatHTTP(struct url *URL, struct url_stat *us, char *flags)
971{
972    FILE *f;
973
974    if ((f = _http_request(URL, "HEAD", us, flags)) == NULL)
975	return -1;
976    fclose(f);
977    return 0;
978}
979
980/*
981 * List a directory
982 */
983struct url_ent *
984fetchListHTTP(struct url *url, char *flags)
985{
986    warnx("fetchListHTTP(): not implemented");
987    return NULL;
988}
989