http.c revision 66325 
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 66325 2000-09-24 12:22:12Z des $
29 */
30
31/*
32 * The following copyright applies to the base64 code:
33 *
34 *-
35 * Copyright 1997 Massachusetts Institute of Technology
36 *
37 * Permission to use, copy, modify, and distribute this software and
38 * its documentation for any purpose and without fee is hereby
39 * granted, provided that both the above copyright notice and this
40 * permission notice appear in all copies, that both the above
41 * copyright notice and this permission notice appear in all
42 * supporting documentation, and that the name of M.I.T. not be used
43 * in advertising or publicity pertaining to distribution of the
44 * software without specific, written prior permission.  M.I.T. makes
45 * no representations about the suitability of this software for any
46 * purpose.  It is provided "as is" without express or implied
47 * warranty.
48 *
49 * THIS SOFTWARE IS PROVIDED BY M.I.T. ``AS IS''.  M.I.T. DISCLAIMS
50 * ALL EXPRESS OR IMPLIED WARRANTIES WITH REGARD TO THIS SOFTWARE,
51 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
52 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT
53 * SHALL M.I.T. BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
54 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
55 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
56 * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
57 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
58 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
59 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
60 * SUCH DAMAGE.
61 */
62
63#include <sys/param.h>
64#include <sys/socket.h>
65
66#include <ctype.h>
67#include <err.h>
68#include <errno.h>
69#include <locale.h>
70#include <netdb.h>
71#include <stdarg.h>
72#include <stdio.h>
73#include <stdlib.h>
74#include <string.h>
75#include <time.h>
76#include <unistd.h>
77
78#include "fetch.h"
79#include "common.h"
80#include "httperr.h"
81
82extern char *__progname; /* XXX not portable */
83
84/* Maximum number of redirects to follow */
85#define MAX_REDIRECT 5
86
87/* Symbolic names for reply codes we care about */
88#define HTTP_OK			200
89#define HTTP_PARTIAL		206
90#define HTTP_MOVED_PERM		301
91#define HTTP_MOVED_TEMP		302
92#define HTTP_SEE_OTHER		303
93#define HTTP_NEED_AUTH		401
94#define HTTP_NEED_PROXY_AUTH	403
95#define HTTP_PROTOCOL_ERROR	999
96
97#define HTTP_REDIRECT(xyz) ((xyz) == HTTP_MOVED_PERM \
98                            || (xyz) == HTTP_MOVED_TEMP \
99                            || (xyz) == HTTP_SEE_OTHER)
100
101
102
103/*****************************************************************************
104 * I/O functions for decoding chunked streams
105 */
106
107struct cookie
108{
109    int		 fd;
110    char	*buf;
111    size_t	 b_size;
112    size_t	 b_len;
113    int		 b_pos;
114    int		 eof;
115    int		 error;
116    long	 chunksize;
117#ifndef NDEBUG
118    long	 total;
119#endif
120};
121
122/*
123 * Get next chunk header
124 */
125static int
126_http_new_chunk(struct cookie *c)
127{
128    char *p;
129
130    if (_fetch_getln(c->fd, &c->buf, &c->b_size, &c->b_len) == -1)
131	return -1;
132
133    if (c->b_len < 2 || !ishexnumber(*c->buf))
134	return -1;
135
136    for (p = c->buf; !isspace(*p) && *p != ';' && p < c->buf + c->b_len; ++p)
137	if (!ishexnumber(*p))
138	    return -1;
139	else if (isdigit(*p))
140	    c->chunksize = c->chunksize * 16 + *p - '0';
141	else
142	    c->chunksize = c->chunksize * 16 + 10 + tolower(*p) - 'a';
143
144#ifndef NDEBUG
145    c->total += c->chunksize;
146    if (c->chunksize == 0)
147	fprintf(stderr, "\033[1m_http_fillbuf(): "
148		"end of last chunk\033[m\n");
149    else
150	fprintf(stderr, "\033[1m_http_fillbuf(): "
151		"new chunk: %ld (%ld)\033[m\n", c->chunksize, c->total);
152#endif
153
154    return c->chunksize;
155}
156
157/*
158 * Fill the input buffer, do chunk decoding on the fly
159 */
160static int
161_http_fillbuf(struct cookie *c)
162{
163    if (c->error)
164	return -1;
165    if (c->eof)
166	return 0;
167
168    if (c->chunksize == 0) {
169	switch (_http_new_chunk(c)) {
170	case -1:
171	    c->error = 1;
172	    return -1;
173	case 0:
174	    c->eof = 1;
175	    return 0;
176	}
177    }
178
179    if (c->b_size < c->chunksize) {
180	char *tmp;
181
182	if ((tmp = realloc(c->buf, c->chunksize)) == NULL)
183	    return -1;
184	c->buf = tmp;
185	c->b_size = c->chunksize;
186    }
187
188    if ((c->b_len = read(c->fd, c->buf, c->chunksize)) == -1)
189	return -1;
190    c->chunksize -= c->b_len;
191
192    if (c->chunksize == 0) {
193	char endl[2];
194	read(c->fd, endl, 2);
195    }
196
197    c->b_pos = 0;
198
199    return c->b_len;
200}
201
202/*
203 * Read function
204 */
205static int
206_http_readfn(void *v, char *buf, int len)
207{
208    struct cookie *c = (struct cookie *)v;
209    int l, pos;
210
211    if (c->error)
212	return -1;
213    if (c->eof)
214	return 0;
215
216    for (pos = 0; len > 0; pos += l, len -= l) {
217	/* empty buffer */
218	if (!c->buf || c->b_pos == c->b_len)
219	    if (_http_fillbuf(c) < 1)
220		break;
221	l = c->b_len - c->b_pos;
222	if (len < l)
223	    l = len;
224	bcopy(c->buf + c->b_pos, buf + pos, l);
225	c->b_pos += l;
226    }
227
228    if (!pos && c->error)
229	return -1;
230    return pos;
231}
232
233/*
234 * Write function
235 */
236static int
237_http_writefn(void *v, const char *buf, int len)
238{
239    struct cookie *c = (struct cookie *)v;
240
241    return write(c->fd, buf, len);
242}
243
244/*
245 * Close function
246 */
247static int
248_http_closefn(void *v)
249{
250    struct cookie *c = (struct cookie *)v;
251    int r;
252
253    r = close(c->fd);
254    if (c->buf)
255	free(c->buf);
256    free(c);
257    return r;
258}
259
260/*
261 * Wrap a file descriptor up
262 */
263static FILE *
264_http_funopen(int fd)
265{
266    struct cookie *c;
267    FILE *f;
268
269    if ((c = calloc(1, sizeof *c)) == NULL) {
270	_fetch_syserr();
271	return NULL;
272    }
273    c->fd = fd;
274    if (!(f = funopen(c, _http_readfn, _http_writefn, NULL, _http_closefn))) {
275	_fetch_syserr();
276	free(c);
277	return NULL;
278    }
279    return f;
280}
281
282
283/*****************************************************************************
284 * Helper functions for talking to the server and parsing its replies
285 */
286
287/* Header types */
288typedef enum {
289    hdr_syserror = -2,
290    hdr_error = -1,
291    hdr_end = 0,
292    hdr_unknown = 1,
293    hdr_content_length,
294    hdr_content_range,
295    hdr_last_modified,
296    hdr_location,
297    hdr_transfer_encoding
298} hdr;
299
300/* Names of interesting headers */
301static struct {
302    hdr		 num;
303    char	*name;
304} hdr_names[] = {
305    { hdr_content_length,	"Content-Length" },
306    { hdr_content_range,	"Content-Range" },
307    { hdr_last_modified,	"Last-Modified" },
308    { hdr_location,		"Location" },
309    { hdr_transfer_encoding,	"Transfer-Encoding" },
310    { hdr_unknown,		NULL },
311};
312
313static char	*reply_buf;
314static size_t	 reply_size;
315static size_t	 reply_length;
316
317/*
318 * Send a formatted line; optionally echo to terminal
319 */
320static int
321_http_cmd(int fd, char *fmt, ...)
322{
323    va_list ap;
324    size_t len;
325    char *msg;
326    int r;
327
328    va_start(ap, fmt);
329    len = vasprintf(&msg, fmt, ap);
330    va_end(ap);
331
332    if (msg == NULL) {
333	errno = ENOMEM;
334	_fetch_syserr();
335	return -1;
336    }
337
338    r = _fetch_putln(fd, msg, len);
339    free(msg);
340
341    if (r == -1) {
342	_fetch_syserr();
343	return -1;
344    }
345
346    return 0;
347}
348
349/*
350 * Get and parse status line
351 */
352static int
353_http_get_reply(int fd)
354{
355    char *p;
356
357    if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
358	return -1;
359    /*
360     * A valid status line looks like "HTTP/m.n xyz reason" where m
361     * and n are the major and minor protocol version numbers and xyz
362     * is the reply code.
363     * Unfortunately, there are servers out there (NCSA 1.5.1, to name
364     * just one) that do not send a version number, so we can't rely
365     * on finding one, but if we do, insist on it being 1.0 or 1.1.
366     * We don't care about the reason phrase.
367     */
368    if (strncmp(reply_buf, "HTTP", 4) != 0)
369	return HTTP_PROTOCOL_ERROR;
370    p = reply_buf + 4;
371    if (*p == '/') {
372	if (p[1] != '1' || p[2] != '.' || (p[3] != '0' && p[3] != '1'))
373	    return HTTP_PROTOCOL_ERROR;
374	p += 4;
375    }
376    if (*p != ' '
377	|| !isdigit(p[1])
378	|| !isdigit(p[2])
379	|| !isdigit(p[3]))
380	return HTTP_PROTOCOL_ERROR;
381
382    return ((p[1] - '0') * 100 + (p[2] - '0') * 10 + (p[3] - '0'));
383}
384
385/*
386 * Check a header; if the type matches the given string, return a
387 * pointer to the beginning of the value.
388 */
389static char *
390_http_match(char *str, char *hdr)
391{
392    while (*str && *hdr && tolower(*str++) == tolower(*hdr++))
393	/* nothing */;
394    if (*str || *hdr != ':')
395	return NULL;
396    while (*hdr && isspace(*++hdr))
397	/* nothing */;
398    return hdr;
399}
400
401/*
402 * Get the next header and return the appropriate symbolic code.
403 */
404static hdr
405_http_next_header(int fd, char **p)
406{
407    int i;
408
409    if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
410	return hdr_syserror;
411    while (reply_length && isspace(reply_buf[reply_length-1]))
412	reply_length--;
413    reply_buf[reply_length] = 0;
414    if (reply_length == 0)
415	return hdr_end;
416    /*
417     * We could check for malformed headers but we don't really care.
418     * A valid header starts with a token immediately followed by a
419     * colon; a token is any sequence of non-control, non-whitespace
420     * characters except "()<>@,;:\\\"{}".
421     */
422    for (i = 0; hdr_names[i].num != hdr_unknown; i++)
423	if ((*p = _http_match(hdr_names[i].name, reply_buf)) != NULL)
424	    return hdr_names[i].num;
425    return hdr_unknown;
426}
427
428/*
429 * Parse a last-modified header
430 */
431static int
432_http_parse_mtime(char *p, time_t *mtime)
433{
434    char locale[64], *r;
435    struct tm tm;
436
437    strncpy(locale, setlocale(LC_TIME, NULL), sizeof locale);
438    setlocale(LC_TIME, "C");
439    r = strptime(p, "%a, %d %b %Y %H:%M:%S GMT", &tm);
440    /* XXX should add support for date-2 and date-3 */
441    setlocale(LC_TIME, locale);
442    if (r == NULL)
443	return -1;
444    DEBUG(fprintf(stderr, "last modified: [\033[1m%04d-%02d-%02d "
445		  "%02d:%02d:%02d\033[m]\n",
446		  tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday,
447		  tm.tm_hour, tm.tm_min, tm.tm_sec));
448    *mtime = timegm(&tm);
449    return 0;
450}
451
452/*
453 * Parse a content-length header
454 */
455static int
456_http_parse_length(char *p, off_t *length)
457{
458    off_t len;
459
460    for (len = 0; *p && isdigit(*p); ++p)
461	len = len * 10 + (*p - '0');
462    DEBUG(fprintf(stderr, "content length: [\033[1m%lld\033[m]\n", len));
463    *length = len;
464    return 0;
465}
466
467/*
468 * Parse a content-range header
469 */
470static int
471_http_parse_range(char *p, off_t *offset, off_t *length, off_t *size)
472{
473    int first, last, len;
474
475    if (strncasecmp(p, "bytes ", 6) != 0)
476	return -1;
477    for (first = 0, p += 6; *p && isdigit(*p); ++p)
478	first = first * 10 + *p - '0';
479    if (*p != '-')
480	return -1;
481    for (last = 0, ++p; *p && isdigit(*p); ++p)
482	last = last * 10 + *p - '0';
483    if (first > last || *p != '/')
484	return -1;
485    for (len = 0, ++p; *p && isdigit(*p); ++p)
486	len = len * 10 + *p - '0';
487    if (len < last - first + 1)
488	return -1;
489    DEBUG(fprintf(stderr, "content range: [\033[1m%d-%d/%d\033[m]\n",
490		  first, last, len));
491    *offset = first;
492    *length = last - first + 1;
493    *size = len;
494    return 0;
495}
496
497
498/*****************************************************************************
499 * Helper functions for authorization
500 */
501
502/*
503 * Base64 encoding
504 */
505static char *
506_http_base64(char *src)
507{
508    static const char base64[] =
509	"ABCDEFGHIJKLMNOPQRSTUVWXYZ"
510	"abcdefghijklmnopqrstuvwxyz"
511	"0123456789+/";
512    char *str, *dst;
513    size_t l;
514    int t, r;
515
516    l = strlen(src);
517    if ((str = malloc(((l + 2) / 3) * 4)) == NULL)
518	return NULL;
519    dst = str;
520    r = 0;
521
522    while (l >= 3) {
523	t = (src[0] << 16) | (src[1] << 8) | src[2];
524	dst[0] = base64[(t >> 18) & 0x3f];
525	dst[1] = base64[(t >> 12) & 0x3f];
526	dst[2] = base64[(t >> 6) & 0x3f];
527	dst[3] = base64[(t >> 0) & 0x3f];
528	src += 3; l -= 3;
529	dst += 4; r += 4;
530    }
531
532    switch (l) {
533    case 2:
534	t = (src[0] << 16) | (src[1] << 8);
535	dst[0] = base64[(t >> 18) & 0x3f];
536	dst[1] = base64[(t >> 12) & 0x3f];
537	dst[2] = base64[(t >> 6) & 0x3f];
538	dst[3] = '=';
539	dst += 4;
540	r += 4;
541	break;
542    case 1:
543	t = src[0] << 16;
544	dst[0] = base64[(t >> 18) & 0x3f];
545	dst[1] = base64[(t >> 12) & 0x3f];
546	dst[2] = dst[3] = '=';
547	dst += 4;
548	r += 4;
549	break;
550    case 0:
551	break;
552    }
553
554    *dst = 0;
555    return str;
556}
557
558/*
559 * Encode username and password
560 */
561static int
562_http_basic_auth(int fd, char *hdr, char *usr, char *pwd)
563{
564    char *upw, *auth;
565    int r;
566
567    if (asprintf(&upw, "%s:%s", usr, pwd) == -1)
568	return -1;
569    auth = _http_base64(upw);
570    free(upw);
571    if (auth == NULL)
572	return -1;
573    r = _http_cmd(fd, "%s: Basic %s", hdr, auth);
574    free(auth);
575    return r;
576}
577
578/*
579 * Send an authorization header
580 */
581static int
582_http_authorize(int fd, char *hdr, char *p)
583{
584    /* basic authorization */
585    if (strncasecmp(p, "basic:", 6) == 0) {
586	char *user, *pwd, *str;
587	int r;
588
589	/* skip realm */
590	for (p += 6; *p && *p != ':'; ++p)
591	    /* nothing */ ;
592	if (!*p || strchr(++p, ':') == NULL)
593	    return -1;
594	if ((str = strdup(p)) == NULL)
595	    return -1; /* XXX */
596	user = str;
597	pwd = strchr(str, ':');
598	*pwd++ = '\0';
599	r = _http_basic_auth(fd, hdr, user, pwd);
600	free(str);
601	return r;
602    }
603    return -1;
604}
605
606
607/*****************************************************************************
608 * Helper functions for connecting to a server or proxy
609 */
610
611/*
612 * Return the default port for this scheme
613 */
614static int
615_http_default_port(char *scheme)
616{
617    struct servent *se;
618
619    if ((se = getservbyname(scheme, "tcp")) != NULL)
620	return ntohs(se->s_port);
621    if (strcasecmp(scheme, "ftp") == 0)
622	return FTP_DEFAULT_PORT;
623    if (strcasecmp(scheme, "http") == 0)
624	return HTTP_DEFAULT_PORT;
625    return 0;
626}
627
628/*
629 * Connect to the specified HTTP proxy server.
630 */
631static int
632_http_proxy_connect(char *proxy, int af, int verbose)
633{
634    char *hostname, *p;
635    int fd, port;
636
637    /* get hostname */
638    hostname = NULL;
639#ifdef INET6
640    /* host part can be an IPv6 address enclosed in square brackets */
641    if (*proxy == '[') {
642	if ((p = strchr(proxy, ']')) == NULL) {
643	    /* no terminating bracket */
644	    /* XXX should set an error code */
645	    goto ouch;
646	}
647	if (p[1] != '\0' && p[1] != ':') {
648	    /* garbage after address */
649	    /* XXX should set an error code */
650	    goto ouch;
651	}
652	if ((hostname = malloc(p - proxy)) == NULL) {
653	    errno = ENOMEM;
654	    _fetch_syserr();
655	    goto ouch;
656	}
657	strncpy(hostname, proxy + 1, p - proxy - 1);
658	hostname[p - proxy - 1] = '\0';
659	++p;
660    } else {
661#endif /* INET6 */
662	if ((p = strchr(proxy, ':')) == NULL)
663	    p = strchr(proxy, '\0');
664	if ((hostname = malloc(p - proxy + 1)) == NULL) {
665	    errno = ENOMEM;
666	    _fetch_syserr();
667	    goto ouch;
668	}
669	strncpy(hostname, proxy, p - proxy);
670	hostname[p - proxy] = '\0';
671#ifdef INET6
672    }
673#endif /* INET6 */
674    DEBUG(fprintf(stderr, "proxy name: [%s]\n", hostname));
675
676    /* get port number */
677    port = 0;
678    if (*p == ':') {
679	++p;
680	if (strspn(p, "0123456789") != strlen(p) || strlen(p) > 5) {
681	    /* port number is non-numeric or too long */
682	    /* XXX should set an error code */
683	    goto ouch;
684	}
685	port = atoi(p);
686	if (port < 1 || port > 65535) {
687	    /* port number is out of range */
688	    /* XXX should set an error code */
689	    goto ouch;
690	}
691    }
692
693    if (!port) {
694#if 0
695	/*
696	 * commented out, since there is currently no service name
697	 * for HTTP proxies
698	 */
699	struct servent *se;
700
701	if ((se = getservbyname("xxxx", "tcp")) != NULL)
702	    port = ntohs(se->s_port);
703	else
704#endif
705	    port = 3128;
706    }
707    DEBUG(fprintf(stderr, "proxy port: %d\n", port));
708
709    /* connect */
710    if ((fd = _fetch_connect(hostname, port, af, verbose)) == -1)
711	_fetch_syserr();
712    return fd;
713
714 ouch:
715    if (hostname)
716	free(hostname);
717    return -1;
718}
719
720/*
721 * Connect to the correct HTTP server or proxy.
722 */
723static int
724_http_connect(struct url *URL, int *proxy, char *flags)
725{
726    int direct, verbose;
727    int af, fd;
728    char *p;
729
730#ifdef INET6
731    af = AF_UNSPEC;
732#else
733    af = AF_INET;
734#endif
735
736    direct = (flags && strchr(flags, 'd'));
737    verbose = (flags && strchr(flags, 'v'));
738    if (flags && strchr(flags, '4'))
739	af = AF_INET;
740    else if (flags && strchr(flags, '6'))
741	af = AF_INET6;
742
743    /* check port */
744    if (!URL->port)
745	URL->port = _http_default_port(URL->scheme);
746
747    if (!direct && (p = getenv("HTTP_PROXY")) != NULL && *p != '\0') {
748	/* attempt to connect to proxy server */
749	if ((fd = _http_proxy_connect(p, af, verbose)) == -1)
750	    return -1;
751	*proxy = 1;
752    } else {
753	/* if no proxy is configured, try direct */
754	if (strcasecmp(URL->scheme, "ftp") == 0) {
755	    /* can't talk http to an ftp server */
756	    /* XXX should set an error code */
757	    return -1;
758	}
759	if ((fd = _fetch_connect(URL->host, URL->port, af, verbose)) == -1)
760	    /* _fetch_connect() has already set an error code */
761	    return -1;
762	*proxy = 0;
763    }
764
765    return fd;
766}
767
768
769/*****************************************************************************
770 * Core
771 */
772
773/*
774 * Send a request and process the reply
775 */
776static FILE *
777_http_request(struct url *URL, char *op, struct url_stat *us, char *flags)
778{
779    struct url *url, *new;
780    int chunked, need_auth, noredirect, proxy, verbose;
781    int code, fd, i, n;
782    off_t offset, clength, length, size;
783    time_t mtime;
784    char *p;
785    FILE *f;
786    hdr h;
787    char *host;
788#ifdef INET6
789    char hbuf[MAXHOSTNAMELEN + 1];
790#endif
791
792    noredirect = (flags && strchr(flags, 'A'));
793    verbose = (flags && strchr(flags, 'v'));
794
795    /* try the provided URL first */
796    url = URL;
797
798    /* if the A flag is set, we only get one try */
799    n = noredirect ? 1 : MAX_REDIRECT;
800    i = 0;
801
802    do {
803	new = NULL;
804	chunked = 0;
805	need_auth = 0;
806	offset = 0;
807	clength = -1;
808	length = -1;
809	size = -1;
810	mtime = 0;
811    retry:
812	/* connect to server or proxy */
813	if ((fd = _http_connect(url, &proxy, flags)) == -1)
814	    goto ouch;
815
816	host = url->host;
817#ifdef INET6
818	if (strchr(url->host, ':')) {
819	    snprintf(hbuf, sizeof(hbuf), "[%s]", url->host);
820	    host = hbuf;
821	}
822#endif
823
824	/* send request */
825	if (verbose)
826	    _fetch_info("requesting %s://%s:%d%s",
827			url->scheme, host, url->port, url->doc);
828	if (proxy) {
829	    _http_cmd(fd, "%s %s://%s:%d%s HTTP/1.1",
830		      op, url->scheme, host, url->port, url->doc);
831	} else {
832	    _http_cmd(fd, "%s %s HTTP/1.1",
833		      op, url->doc);
834	}
835
836	/* proxy authorization */
837	if (proxy && (p = getenv("HTTP_PROXY_AUTH")) != NULL && *p != '\0')
838	    _http_authorize(fd, "Proxy-Authorization", p);
839
840	/* server authorization */
841	if (need_auth) {
842	    if (*url->user || *url->pwd)
843		_http_basic_auth(fd, "Authorization",
844				 url->user ? url->user : "",
845				 url->pwd ? url->pwd : "");
846	    else if ((p = getenv("HTTP_AUTH")) != NULL && *p != '\0')
847		_http_authorize(fd, "Authorization", p);
848	    else {
849		_http_seterr(HTTP_NEED_AUTH);
850		goto ouch;
851	    }
852	}
853
854	/* other headers */
855	if (url->port == _http_default_port(url->scheme))
856	    _http_cmd(fd, "Host: %s", host);
857	else
858	    _http_cmd(fd, "Host: %s:%d", host, url->port);
859	_http_cmd(fd, "User-Agent: %s " _LIBFETCH_VER, __progname);
860	if (url->offset)
861	    _http_cmd(fd, "Range: bytes=%lld-", url->offset);
862	_http_cmd(fd, "Connection: close");
863	_http_cmd(fd, "");
864
865	/* get reply */
866	switch ((code = _http_get_reply(fd))) {
867	case HTTP_OK:
868	case HTTP_PARTIAL:
869	    /* fine */
870	    break;
871	case HTTP_MOVED_PERM:
872	case HTTP_MOVED_TEMP:
873	    /*
874	     * Not so fine, but we still have to read the headers to
875	     * get the new location.
876	     */
877	    break;
878	case HTTP_NEED_AUTH:
879	    if (need_auth) {
880		/*
881		 * We already sent out authorization code, so there's
882		 * nothing more we can do.
883		 */
884		_http_seterr(code);
885		goto ouch;
886	    }
887	    /* try again, but send the password this time */
888	    if (verbose)
889		_fetch_info("server requires authorization");
890	    need_auth = 1;
891	    close(fd);
892	    goto retry;
893	case HTTP_NEED_PROXY_AUTH:
894	    /*
895	     * If we're talking to a proxy, we already sent our proxy
896	     * authorization code, so there's nothing more we can do.
897	     */
898	    _http_seterr(code);
899	    goto ouch;
900	case HTTP_PROTOCOL_ERROR:
901	    /* fall through */
902	case -1:
903	    _fetch_syserr();
904	    goto ouch;
905	default:
906	    _http_seterr(code);
907	    goto ouch;
908	}
909
910	/* get headers */
911	do {
912	    switch ((h = _http_next_header(fd, &p))) {
913	    case hdr_syserror:
914		_fetch_syserr();
915		goto ouch;
916	    case hdr_error:
917		_http_seterr(HTTP_PROTOCOL_ERROR);
918		goto ouch;
919	    case hdr_content_length:
920		_http_parse_length(p, &clength);
921		break;
922	    case hdr_content_range:
923		_http_parse_range(p, &offset, &length, &size);
924		break;
925	    case hdr_last_modified:
926		_http_parse_mtime(p, &mtime);
927		break;
928	    case hdr_location:
929		if (!HTTP_REDIRECT(code))
930		    break;
931		if (new)
932		    free(new);
933		if (verbose)
934		    _fetch_info("%d redirect to %s", code, p);
935		if (*p == '/')
936		    /* absolute path */
937		    new = fetchMakeURL(url->scheme, url->host, url->port, p,
938				       url->user, url->pwd);
939		else
940		    new = fetchParseURL(p);
941		if (new == NULL) {
942		    /* XXX should set an error code */
943		    DEBUG(fprintf(stderr, "failed to parse new URL\n"));
944		    goto ouch;
945		}
946		if (!*new->user && !*new->pwd) {
947		    strcpy(new->user, url->user);
948		    strcpy(new->pwd, url->pwd);
949		}
950		new->offset = url->offset;
951		new->length = url->length;
952		break;
953	    case hdr_transfer_encoding:
954		/* XXX weak test*/
955		chunked = (strcasecmp(p, "chunked") == 0);
956		break;
957	    case hdr_end:
958		/* fall through */
959	    case hdr_unknown:
960		/* ignore */
961		break;
962	    }
963	} while (h > hdr_end);
964
965	/* we either have a hit, or a redirect with no Location: header */
966	if (code == HTTP_OK || code == HTTP_PARTIAL || !new)
967	    break;
968
969	/* we have a redirect */
970	close(fd);
971	fd = -1;
972	if (url != URL)
973	    fetchFreeURL(url);
974	url = new;
975    } while (++i < n);
976
977    /* no success */
978    if (fd == -1) {
979	_http_seterr(code);
980	goto ouch;
981    }
982
983    DEBUG(fprintf(stderr, "offset %lld, length %lld, size %lld, clength %lld\n",
984		  offset, length, size, clength));
985
986    /* check for inconsistencies */
987    if (clength != -1 && length != -1 && clength != length) {
988	_http_seterr(HTTP_PROTOCOL_ERROR);
989	goto ouch;
990    }
991    if (clength == -1)
992	clength = length;
993    if (clength != -1)
994	length = offset + clength;
995    if (length != -1 && size != -1 && length != size) {
996	_http_seterr(HTTP_PROTOCOL_ERROR);
997	goto ouch;
998    }
999    if (size == -1)
1000	size = length;
1001
1002    /* fill in stats */
1003    if (us) {
1004	us->size = size;
1005	us->atime = us->mtime = mtime;
1006    }
1007
1008    /* too far? */
1009    if (offset > URL->offset) {
1010	_http_seterr(HTTP_PROTOCOL_ERROR);
1011	goto ouch;
1012    }
1013
1014    /* report back real offset and size */
1015    URL->offset = offset;
1016    URL->length = clength;
1017
1018    /* wrap it up in a FILE */
1019    if ((f = chunked ? _http_funopen(fd) : fdopen(fd, "r")) == NULL) {
1020	_fetch_syserr();
1021	goto ouch;
1022    }
1023
1024    if (url != URL)
1025	fetchFreeURL(url);
1026
1027    return f;
1028
1029 ouch:
1030    if (url != URL)
1031	fetchFreeURL(url);
1032    if (fd != -1)
1033	close(fd);
1034    return NULL;
1035}
1036
1037
1038/*****************************************************************************
1039 * Entry points
1040 */
1041
1042/*
1043 * Retrieve and stat a file by HTTP
1044 */
1045FILE *
1046fetchXGetHTTP(struct url *URL, struct url_stat *us, char *flags)
1047{
1048    return _http_request(URL, "GET", us, flags);
1049}
1050
1051/*
1052 * Retrieve a file by HTTP
1053 */
1054FILE *
1055fetchGetHTTP(struct url *URL, char *flags)
1056{
1057    return fetchXGetHTTP(URL, NULL, flags);
1058}
1059
1060/*
1061 * Store a file by HTTP
1062 */
1063FILE *
1064fetchPutHTTP(struct url *URL, char *flags)
1065{
1066    warnx("fetchPutHTTP(): not implemented");
1067    return NULL;
1068}
1069
1070/*
1071 * Get an HTTP document's metadata
1072 */
1073int
1074fetchStatHTTP(struct url *URL, struct url_stat *us, char *flags)
1075{
1076    FILE *f;
1077
1078    if ((f = _http_request(URL, "HEAD", us, flags)) == NULL)
1079	return -1;
1080    fclose(f);
1081    return 0;
1082}
1083
1084/*
1085 * List a directory
1086 */
1087struct url_ent *
1088fetchListHTTP(struct url *url, char *flags)
1089{
1090    warnx("fetchListHTTP(): not implemented");
1091    return NULL;
1092}
1093