http.c revision 90267 
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
29#include <sys/cdefs.h>
30__FBSDID("$FreeBSD: head/lib/libfetch/http.c 90267 2002-02-05 22:13:51Z des $");
31
32/*
33 * The following copyright applies to the base64 code:
34 *
35 *-
36 * Copyright 1997 Massachusetts Institute of Technology
37 *
38 * Permission to use, copy, modify, and distribute this software and
39 * its documentation for any purpose and without fee is hereby
40 * granted, provided that both the above copyright notice and this
41 * permission notice appear in all copies, that both the above
42 * copyright notice and this permission notice appear in all
43 * supporting documentation, and that the name of M.I.T. not be used
44 * in advertising or publicity pertaining to distribution of the
45 * software without specific, written prior permission.  M.I.T. makes
46 * no representations about the suitability of this software for any
47 * purpose.  It is provided "as is" without express or implied
48 * warranty.
49 *
50 * THIS SOFTWARE IS PROVIDED BY M.I.T. ``AS IS''.  M.I.T. DISCLAIMS
51 * ALL EXPRESS OR IMPLIED WARRANTIES WITH REGARD TO THIS SOFTWARE,
52 * INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
53 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. IN NO EVENT
54 * SHALL M.I.T. BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
55 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
56 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
57 * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
58 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
59 * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
60 * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
61 * SUCH DAMAGE.
62 */
63
64#include <sys/param.h>
65#include <sys/socket.h>
66
67#include <ctype.h>
68#include <err.h>
69#include <errno.h>
70#include <locale.h>
71#include <netdb.h>
72#include <stdarg.h>
73#include <stdio.h>
74#include <stdlib.h>
75#include <string.h>
76#include <time.h>
77#include <unistd.h>
78
79#include "fetch.h"
80#include "common.h"
81#include "httperr.h"
82
83extern char	*__progname; /* XXX not portable */
84
85/* Maximum number of redirects to follow */
86#define MAX_REDIRECT 5
87
88/* Symbolic names for reply codes we care about */
89#define HTTP_OK			200
90#define HTTP_PARTIAL		206
91#define HTTP_MOVED_PERM		301
92#define HTTP_MOVED_TEMP		302
93#define HTTP_SEE_OTHER		303
94#define HTTP_NEED_AUTH		401
95#define HTTP_NEED_PROXY_AUTH	407
96#define HTTP_PROTOCOL_ERROR	999
97
98#define HTTP_REDIRECT(xyz) ((xyz) == HTTP_MOVED_PERM \
99			    || (xyz) == HTTP_MOVED_TEMP \
100			    || (xyz) == HTTP_SEE_OTHER)
101
102#define HTTP_ERROR(xyz) ((xyz) > 400 && (xyz) < 599)
103
104
105/*****************************************************************************
106 * I/O functions for decoding chunked streams
107 */
108
109struct cookie
110{
111	int		 fd;
112	char		*buf;
113	size_t		 b_size;
114	ssize_t		 b_len;
115	int		 b_pos;
116	int		 eof;
117	int		 error;
118	size_t		 chunksize;
119#ifndef NDEBUG
120	size_t		 total;
121#endif
122};
123
124/*
125 * Get next chunk header
126 */
127static int
128_http_new_chunk(struct cookie *c)
129{
130	char *p;
131
132	if (_fetch_getln(c->fd, &c->buf, &c->b_size, &c->b_len) == -1)
133		return (-1);
134
135	if (c->b_len < 2 || !ishexnumber(*c->buf))
136		return (-1);
137
138	for (p = c->buf; !isspace(*p) && p < c->buf + c->b_len; ++p) {
139		if (*p == ';')
140			break;
141		if (!ishexnumber(*p))
142			return (-1);
143		if (isdigit(*p)) {
144			c->chunksize = c->chunksize * 16 +
145			    *p - '0';
146		} else {
147			c->chunksize = c->chunksize * 16 +
148			    10 + tolower(*p) - 'a';
149		}
150	}
151
152#ifndef NDEBUG
153	if (fetchDebug) {
154		c->total += c->chunksize;
155		if (c->chunksize == 0)
156			fprintf(stderr, "_http_fillbuf(): "
157			    "end of last chunk\n");
158		else
159			fprintf(stderr, "_http_fillbuf(): "
160			    "new chunk: %lu (%lu)\n",
161			    (unsigned long)c->chunksize, (unsigned long)c->total);
162	}
163#endif
164
165	return (c->chunksize);
166}
167
168/*
169 * Fill the input buffer, do chunk decoding on the fly
170 */
171static int
172_http_fillbuf(struct cookie *c)
173{
174	if (c->error)
175		return (-1);
176	if (c->eof)
177		return (0);
178
179	if (c->chunksize == 0) {
180		switch (_http_new_chunk(c)) {
181		case -1:
182			c->error = 1;
183			return (-1);
184		case 0:
185			c->eof = 1;
186			return (0);
187		}
188	}
189
190	if (c->b_size < c->chunksize) {
191		char *tmp;
192
193		if ((tmp = realloc(c->buf, c->chunksize)) == NULL)
194			return (-1);
195		c->buf = tmp;
196		c->b_size = c->chunksize;
197	}
198
199	if ((c->b_len = read(c->fd, c->buf, c->chunksize)) == -1)
200		return (-1);
201	c->chunksize -= c->b_len;
202
203	if (c->chunksize == 0) {
204		char endl;
205		if (read(c->fd, &endl, 1) == -1 ||
206		    read(c->fd, &endl, 1) == -1)
207			return (-1);
208	}
209
210	c->b_pos = 0;
211
212	return (c->b_len);
213}
214
215/*
216 * Read function
217 */
218static int
219_http_readfn(void *v, char *buf, int len)
220{
221	struct cookie *c = (struct cookie *)v;
222	int l, pos;
223
224	if (c->error)
225		return (-1);
226	if (c->eof)
227		return (0);
228
229	for (pos = 0; len > 0; pos += l, len -= l) {
230		/* empty buffer */
231		if (!c->buf || c->b_pos == c->b_len)
232			if (_http_fillbuf(c) < 1)
233				break;
234		l = c->b_len - c->b_pos;
235		if (len < l)
236			l = len;
237		bcopy(c->buf + c->b_pos, buf + pos, l);
238		c->b_pos += l;
239	}
240
241	if (!pos && c->error)
242		return (-1);
243	return (pos);
244}
245
246/*
247 * Write function
248 */
249static int
250_http_writefn(void *v, const char *buf, int len)
251{
252	struct cookie *c = (struct cookie *)v;
253
254	return (write(c->fd, buf, len));
255}
256
257/*
258 * Close function
259 */
260static int
261_http_closefn(void *v)
262{
263	struct cookie *c = (struct cookie *)v;
264	int r;
265
266	r = close(c->fd);
267	if (c->buf)
268		free(c->buf);
269	free(c);
270	return (r);
271}
272
273/*
274 * Wrap a file descriptor up
275 */
276static FILE *
277_http_funopen(int fd)
278{
279	struct cookie *c;
280	FILE *f;
281
282	if ((c = calloc(1, sizeof *c)) == NULL) {
283		_fetch_syserr();
284		return (NULL);
285	}
286	c->fd = fd;
287	f = funopen(c, _http_readfn, _http_writefn, NULL, _http_closefn);
288	if (f == NULL) {
289		_fetch_syserr();
290		free(c);
291		return (NULL);
292	}
293	return (f);
294}
295
296
297/*****************************************************************************
298 * Helper functions for talking to the server and parsing its replies
299 */
300
301/* Header types */
302typedef enum {
303	hdr_syserror = -2,
304	hdr_error = -1,
305	hdr_end = 0,
306	hdr_unknown = 1,
307	hdr_content_length,
308	hdr_content_range,
309	hdr_last_modified,
310	hdr_location,
311	hdr_transfer_encoding,
312	hdr_www_authenticate
313} hdr_t;
314
315/* Names of interesting headers */
316static struct {
317	hdr_t		 num;
318	const char	*name;
319} hdr_names[] = {
320	{ hdr_content_length,		"Content-Length" },
321	{ hdr_content_range,		"Content-Range" },
322	{ hdr_last_modified,		"Last-Modified" },
323	{ hdr_location,			"Location" },
324	{ hdr_transfer_encoding,	"Transfer-Encoding" },
325	{ hdr_www_authenticate,		"WWW-Authenticate" },
326	{ hdr_unknown,			NULL },
327};
328
329static char		*reply_buf;
330static size_t		 reply_size;
331static size_t		 reply_length;
332
333/*
334 * Send a formatted line; optionally echo to terminal
335 */
336static int
337_http_cmd(int fd, const char *fmt, ...)
338{
339	va_list ap;
340	size_t len;
341	char *msg;
342	int r;
343
344	va_start(ap, fmt);
345	len = vasprintf(&msg, fmt, ap);
346	va_end(ap);
347
348	if (msg == NULL) {
349		errno = ENOMEM;
350		_fetch_syserr();
351		return (-1);
352	}
353
354	r = _fetch_putln(fd, msg, len);
355	free(msg);
356
357	if (r == -1) {
358		_fetch_syserr();
359		return (-1);
360	}
361
362	return (0);
363}
364
365/*
366 * Get and parse status line
367 */
368static int
369_http_get_reply(int fd)
370{
371	char *p;
372
373	if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
374		return (-1);
375	/*
376	 * A valid status line looks like "HTTP/m.n xyz reason" where m
377	 * and n are the major and minor protocol version numbers and xyz
378	 * is the reply code.
379	 * Unfortunately, there are servers out there (NCSA 1.5.1, to name
380	 * just one) that do not send a version number, so we can't rely
381	 * on finding one, but if we do, insist on it being 1.0 or 1.1.
382	 * We don't care about the reason phrase.
383	 */
384	if (strncmp(reply_buf, "HTTP", 4) != 0)
385		return (HTTP_PROTOCOL_ERROR);
386	p = reply_buf + 4;
387	if (*p == '/') {
388		if (p[1] != '1' || p[2] != '.' || (p[3] != '0' && p[3] != '1'))
389			return (HTTP_PROTOCOL_ERROR);
390		p += 4;
391	}
392	if (*p != ' ' || !isdigit(p[1]) || !isdigit(p[2]) || !isdigit(p[3]))
393		return (HTTP_PROTOCOL_ERROR);
394
395	return ((p[1] - '0') * 100 + (p[2] - '0') * 10 + (p[3] - '0'));
396}
397
398/*
399 * Check a header; if the type matches the given string, return a pointer
400 * to the beginning of the value.
401 */
402static const char *
403_http_match(const char *str, const char *hdr)
404{
405	while (*str && *hdr && tolower(*str++) == tolower(*hdr++))
406		/* nothing */;
407	if (*str || *hdr != ':')
408		return (NULL);
409	while (*hdr && isspace(*++hdr))
410		/* nothing */;
411	return (hdr);
412}
413
414/*
415 * Get the next header and return the appropriate symbolic code.
416 */
417static hdr_t
418_http_next_header(int fd, const char **p)
419{
420	int i;
421
422	if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
423		return (hdr_syserror);
424	while (reply_length && isspace(reply_buf[reply_length-1]))
425		reply_length--;
426	reply_buf[reply_length] = 0;
427	if (reply_length == 0)
428	return (hdr_end);
429	/*
430	 * We could check for malformed headers but we don't really care.
431	 * A valid header starts with a token immediately followed by a
432	 * colon; a token is any sequence of non-control, non-whitespace
433	 * characters except "()<>@,;:\\\"{}".
434	 */
435	for (i = 0; hdr_names[i].num != hdr_unknown; i++)
436		if ((*p = _http_match(hdr_names[i].name, reply_buf)) != NULL)
437			return (hdr_names[i].num);
438	return (hdr_unknown);
439}
440
441/*
442 * Parse a last-modified header
443 */
444static int
445_http_parse_mtime(const char *p, time_t *mtime)
446{
447	char locale[64], *r;
448	struct tm tm;
449
450	strncpy(locale, setlocale(LC_TIME, NULL), sizeof locale);
451	setlocale(LC_TIME, "C");
452	r = strptime(p, "%a, %d %b %Y %H:%M:%S GMT", &tm);
453	/* XXX should add support for date-2 and date-3 */
454	setlocale(LC_TIME, locale);
455	if (r == NULL)
456		return (-1);
457	DEBUG(fprintf(stderr, "last modified: [%04d-%02d-%02d "
458		  "%02d:%02d:%02d]\n",
459		  tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday,
460		  tm.tm_hour, tm.tm_min, tm.tm_sec));
461	*mtime = timegm(&tm);
462	return (0);
463}
464
465/*
466 * Parse a content-length header
467 */
468static int
469_http_parse_length(const char *p, off_t *length)
470{
471	off_t len;
472
473	for (len = 0; *p && isdigit(*p); ++p)
474		len = len * 10 + (*p - '0');
475	if (*p)
476		return (-1);
477	DEBUG(fprintf(stderr, "content length: [%lld]\n",
478	    (long long)len));
479	*length = len;
480	return (0);
481}
482
483/*
484 * Parse a content-range header
485 */
486static int
487_http_parse_range(const char *p, off_t *offset, off_t *length, off_t *size)
488{
489	off_t first, last, len;
490
491	if (strncasecmp(p, "bytes ", 6) != 0)
492		return (-1);
493	for (first = 0, p += 6; *p && isdigit(*p); ++p)
494		first = first * 10 + *p - '0';
495	if (*p != '-')
496		return (-1);
497	for (last = 0, ++p; *p && isdigit(*p); ++p)
498		last = last * 10 + *p - '0';
499	if (first > last || *p != '/')
500		return (-1);
501	for (len = 0, ++p; *p && isdigit(*p); ++p)
502		len = len * 10 + *p - '0';
503	if (*p || len < last - first + 1)
504		return (-1);
505	DEBUG(fprintf(stderr, "content range: [%lld-%lld/%lld]\n",
506	    (long long)first, (long long)last, (long long)len));
507	*offset = first;
508	*length = last - first + 1;
509	*size = len;
510	return (0);
511}
512
513
514/*****************************************************************************
515 * Helper functions for authorization
516 */
517
518/*
519 * Base64 encoding
520 */
521static char *
522_http_base64(const char *src)
523{
524	static const char base64[] =
525	    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
526	    "abcdefghijklmnopqrstuvwxyz"
527	    "0123456789+/";
528	char *str, *dst;
529	size_t l;
530	int t, r;
531
532	l = strlen(src);
533	if ((str = malloc(((l + 2) / 3) * 4)) == NULL)
534		return (NULL);
535	dst = str;
536	r = 0;
537
538	while (l >= 3) {
539		t = (src[0] << 16) | (src[1] << 8) | src[2];
540		dst[0] = base64[(t >> 18) & 0x3f];
541		dst[1] = base64[(t >> 12) & 0x3f];
542		dst[2] = base64[(t >> 6) & 0x3f];
543		dst[3] = base64[(t >> 0) & 0x3f];
544		src += 3; l -= 3;
545		dst += 4; r += 4;
546	}
547
548	switch (l) {
549	case 2:
550		t = (src[0] << 16) | (src[1] << 8);
551		dst[0] = base64[(t >> 18) & 0x3f];
552		dst[1] = base64[(t >> 12) & 0x3f];
553		dst[2] = base64[(t >> 6) & 0x3f];
554		dst[3] = '=';
555		dst += 4;
556		r += 4;
557		break;
558	case 1:
559		t = src[0] << 16;
560		dst[0] = base64[(t >> 18) & 0x3f];
561		dst[1] = base64[(t >> 12) & 0x3f];
562		dst[2] = dst[3] = '=';
563		dst += 4;
564		r += 4;
565		break;
566	case 0:
567		break;
568	}
569
570	*dst = 0;
571	return (str);
572}
573
574/*
575 * Encode username and password
576 */
577static int
578_http_basic_auth(int fd, const char *hdr, const char *usr, const char *pwd)
579{
580	char *upw, *auth;
581	int r;
582
583	DEBUG(fprintf(stderr, "usr: [%s]\n", usr));
584	DEBUG(fprintf(stderr, "pwd: [%s]\n", pwd));
585	if (asprintf(&upw, "%s:%s", usr, pwd) == -1)
586		return (-1);
587	auth = _http_base64(upw);
588	free(upw);
589	if (auth == NULL)
590		return (-1);
591	r = _http_cmd(fd, "%s: Basic %s", hdr, auth);
592	free(auth);
593	return (r);
594}
595
596/*
597 * Send an authorization header
598 */
599static int
600_http_authorize(int fd, const char *hdr, const char *p)
601{
602	/* basic authorization */
603	if (strncasecmp(p, "basic:", 6) == 0) {
604		char *user, *pwd, *str;
605		int r;
606
607		/* skip realm */
608		for (p += 6; *p && *p != ':'; ++p)
609			/* nothing */ ;
610		if (!*p || strchr(++p, ':') == NULL)
611			return (-1);
612		if ((str = strdup(p)) == NULL)
613			return (-1); /* XXX */
614		user = str;
615		pwd = strchr(str, ':');
616		*pwd++ = '\0';
617		r = _http_basic_auth(fd, hdr, user, pwd);
618		free(str);
619		return (r);
620	}
621	return (-1);
622}
623
624
625/*****************************************************************************
626 * Helper functions for connecting to a server or proxy
627 */
628
629/*
630 * Connect to the correct HTTP server or proxy.
631 */
632static int
633_http_connect(struct url *URL, struct url *purl, const char *flags)
634{
635	int verbose;
636	int af, fd;
637
638#ifdef INET6
639	af = AF_UNSPEC;
640#else
641	af = AF_INET;
642#endif
643
644	verbose = CHECK_FLAG('v');
645	if (CHECK_FLAG('4'))
646		af = AF_INET;
647#ifdef INET6
648	else if (CHECK_FLAG('6'))
649		af = AF_INET6;
650#endif
651
652	if (purl) {
653		URL = purl;
654	} else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) {
655		/* can't talk http to an ftp server */
656		/* XXX should set an error code */
657		return (-1);
658	}
659
660	if ((fd = _fetch_connect(URL->host, URL->port, af, verbose)) == -1)
661		/* _fetch_connect() has already set an error code */
662		return (-1);
663	return (fd);
664}
665
666static struct url *
667_http_get_proxy(void)
668{
669	struct url *purl;
670	char *p;
671
672	if (((p = getenv("HTTP_PROXY")) || (p = getenv("http_proxy"))) &&
673	    (purl = fetchParseURL(p))) {
674		if (!*purl->scheme)
675			strcpy(purl->scheme, SCHEME_HTTP);
676		if (!purl->port)
677			purl->port = _fetch_default_proxy_port(purl->scheme);
678		if (strcasecmp(purl->scheme, SCHEME_HTTP) == 0)
679			return (purl);
680		fetchFreeURL(purl);
681	}
682	return (NULL);
683}
684
685static void
686_http_print_html(FILE *out, FILE *in)
687{
688	size_t len;
689	char *line, *p, *q;
690	int comment, tag;
691
692	comment = tag = 0;
693	while ((line = fgetln(in, &len)) != NULL) {
694		while (len && isspace(line[len - 1]))
695			--len;
696		for (p = q = line; q < line + len; ++q) {
697			if (comment && *q == '-') {
698				if (q + 2 < line + len &&
699				    strcmp(q, "-->") == 0) {
700					tag = comment = 0;
701					q += 2;
702				}
703			} else if (tag && !comment && *q == '>') {
704				p = q + 1;
705				tag = 0;
706			} else if (!tag && *q == '<') {
707				if (q > p)
708					fwrite(p, q - p, 1, out);
709				tag = 1;
710				if (q + 3 < line + len &&
711				    strcmp(q, "<!--") == 0) {
712					comment = 1;
713					q += 3;
714				}
715			}
716		}
717		if (!tag && q > p)
718			fwrite(p, q - p, 1, out);
719		fputc('\n', out);
720	}
721}
722
723
724/*****************************************************************************
725 * Core
726 */
727
728/*
729 * Send a request and process the reply
730 */
731FILE *
732_http_request(struct url *URL, const char *op, struct url_stat *us,
733    struct url *purl, const char *flags)
734{
735	struct url *url, *new;
736	int chunked, direct, need_auth, noredirect, verbose;
737	int code, fd, i, n;
738	off_t offset, clength, length, size;
739	time_t mtime;
740	const char *p;
741	FILE *f;
742	hdr_t h;
743	char *host;
744#ifdef INET6
745	char hbuf[MAXHOSTNAMELEN + 1];
746#endif
747
748	direct = CHECK_FLAG('d');
749	noredirect = CHECK_FLAG('A');
750	verbose = CHECK_FLAG('v');
751
752	if (direct && purl) {
753		fetchFreeURL(purl);
754		purl = NULL;
755	}
756
757	/* try the provided URL first */
758	url = URL;
759
760	/* if the A flag is set, we only get one try */
761	n = noredirect ? 1 : MAX_REDIRECT;
762	i = 0;
763
764	need_auth = 0;
765	do {
766		new = NULL;
767		chunked = 0;
768		offset = 0;
769		clength = -1;
770		length = -1;
771		size = -1;
772		mtime = 0;
773
774		/* check port */
775		if (!url->port)
776			url->port = _fetch_default_port(url->scheme);
777
778		/* were we redirected to an FTP URL? */
779		if (purl == NULL && strcmp(url->scheme, SCHEME_FTP) == 0) {
780			if (strcmp(op, "GET") == 0)
781				return (_ftp_request(url, "RETR", us, purl, flags));
782			else if (strcmp(op, "HEAD") == 0)
783				return (_ftp_request(url, "STAT", us, purl, flags));
784		}
785
786		/* connect to server or proxy */
787		if ((fd = _http_connect(url, purl, flags)) == -1)
788			goto ouch;
789
790		host = url->host;
791#ifdef INET6
792		if (strchr(url->host, ':')) {
793			snprintf(hbuf, sizeof(hbuf), "[%s]", url->host);
794			host = hbuf;
795		}
796#endif
797
798		/* send request */
799		if (verbose)
800			_fetch_info("requesting %s://%s:%d%s",
801			    url->scheme, host, url->port, url->doc);
802		if (purl) {
803			_http_cmd(fd, "%s %s://%s:%d%s HTTP/1.1",
804			    op, url->scheme, host, url->port, url->doc);
805		} else {
806			_http_cmd(fd, "%s %s HTTP/1.1",
807			    op, url->doc);
808		}
809
810		/* virtual host */
811		if (url->port == _fetch_default_port(url->scheme))
812			_http_cmd(fd, "Host: %s", host);
813		else
814			_http_cmd(fd, "Host: %s:%d", host, url->port);
815
816		/* proxy authorization */
817		if (purl) {
818			if (*purl->user || *purl->pwd)
819				_http_basic_auth(fd, "Proxy-Authorization",
820				    purl->user, purl->pwd);
821			else if ((p = getenv("HTTP_PROXY_AUTH")) != NULL && *p != '\0')
822				_http_authorize(fd, "Proxy-Authorization", p);
823		}
824
825		/* server authorization */
826		if (need_auth || *url->user || *url->pwd) {
827			if (*url->user || *url->pwd)
828				_http_basic_auth(fd, "Authorization", url->user, url->pwd);
829			else if ((p = getenv("HTTP_AUTH")) != NULL && *p != '\0')
830				_http_authorize(fd, "Authorization", p);
831			else if (fetchAuthMethod && fetchAuthMethod(url) == 0) {
832				_http_basic_auth(fd, "Authorization", url->user, url->pwd);
833			} else {
834				_http_seterr(HTTP_NEED_AUTH);
835				goto ouch;
836			}
837		}
838
839		/* other headers */
840		if ((p = getenv("HTTP_USER_AGENT")) != NULL && *p != '\0')
841			_http_cmd(fd, "User-Agent: %s", p);
842		else
843			_http_cmd(fd, "User-Agent: %s " _LIBFETCH_VER, __progname);
844		if (url->offset)
845			_http_cmd(fd, "Range: bytes=%lld-", (long long)url->offset);
846		_http_cmd(fd, "Connection: close");
847		_http_cmd(fd, "");
848
849		/* get reply */
850		switch ((code = _http_get_reply(fd))) {
851		case HTTP_OK:
852		case HTTP_PARTIAL:
853			/* fine */
854			break;
855		case HTTP_MOVED_PERM:
856		case HTTP_MOVED_TEMP:
857		case HTTP_SEE_OTHER:
858			/*
859			 * Not so fine, but we still have to read the headers to
860			 * get the new location.
861			 */
862			break;
863		case HTTP_NEED_AUTH:
864			if (need_auth) {
865				/*
866				 * We already sent out authorization code, so there's
867				 * nothing more we can do.
868				 */
869				_http_seterr(code);
870				goto ouch;
871			}
872			/* try again, but send the password this time */
873			if (verbose)
874				_fetch_info("server requires authorization");
875			break;
876		case HTTP_NEED_PROXY_AUTH:
877			/*
878			 * If we're talking to a proxy, we already sent our proxy
879			 * authorization code, so there's nothing more we can do.
880			 */
881			_http_seterr(code);
882			goto ouch;
883		case HTTP_PROTOCOL_ERROR:
884			/* fall through */
885		case -1:
886			_fetch_syserr();
887			goto ouch;
888		default:
889			_http_seterr(code);
890			if (!verbose)
891				goto ouch;
892			/* fall through so we can get the full error message */
893		}
894
895		/* get headers */
896		do {
897			switch ((h = _http_next_header(fd, &p))) {
898			case hdr_syserror:
899				_fetch_syserr();
900				goto ouch;
901			case hdr_error:
902				_http_seterr(HTTP_PROTOCOL_ERROR);
903				goto ouch;
904			case hdr_content_length:
905				_http_parse_length(p, &clength);
906				break;
907			case hdr_content_range:
908				_http_parse_range(p, &offset, &length, &size);
909				break;
910			case hdr_last_modified:
911				_http_parse_mtime(p, &mtime);
912				break;
913			case hdr_location:
914				if (!HTTP_REDIRECT(code))
915					break;
916				if (new)
917					free(new);
918				if (verbose)
919					_fetch_info("%d redirect to %s", code, p);
920				if (*p == '/')
921					/* absolute path */
922					new = fetchMakeURL(url->scheme, url->host, url->port, p,
923					    url->user, url->pwd);
924				else
925					new = fetchParseURL(p);
926				if (new == NULL) {
927					/* XXX should set an error code */
928					DEBUG(fprintf(stderr, "failed to parse new URL\n"));
929					goto ouch;
930				}
931				if (!*new->user && !*new->pwd) {
932					strcpy(new->user, url->user);
933					strcpy(new->pwd, url->pwd);
934				}
935				new->offset = url->offset;
936				new->length = url->length;
937				break;
938			case hdr_transfer_encoding:
939				/* XXX weak test*/
940				chunked = (strcasecmp(p, "chunked") == 0);
941				break;
942			case hdr_www_authenticate:
943				if (code != HTTP_NEED_AUTH)
944					break;
945				/* if we were smarter, we'd check the method and realm */
946				break;
947			case hdr_end:
948				/* fall through */
949			case hdr_unknown:
950				/* ignore */
951				break;
952			}
953		} while (h > hdr_end);
954
955		/* we have a hit or an error */
956		if (code == HTTP_OK || code == HTTP_PARTIAL || HTTP_ERROR(code))
957			break;
958
959		/* we need to provide authentication */
960		if (code == HTTP_NEED_AUTH) {
961			need_auth = 1;
962			close(fd);
963			fd = -1;
964			continue;
965		}
966
967		/* all other cases: we got a redirect */
968		need_auth = 0;
969		close(fd);
970		fd = -1;
971		if (!new) {
972			DEBUG(fprintf(stderr, "redirect with no new location\n"));
973			break;
974		}
975		if (url != URL)
976			fetchFreeURL(url);
977		url = new;
978	} while (++i < n);
979
980	/* we failed, or ran out of retries */
981	if (fd == -1) {
982		_http_seterr(code);
983		goto ouch;
984	}
985
986	DEBUG(fprintf(stderr, "offset %lld, length %lld,"
987		  " size %lld, clength %lld\n",
988		  (long long)offset, (long long)length,
989		  (long long)size, (long long)clength));
990
991	/* check for inconsistencies */
992	if (clength != -1 && length != -1 && clength != length) {
993		_http_seterr(HTTP_PROTOCOL_ERROR);
994		goto ouch;
995	}
996	if (clength == -1)
997		clength = length;
998	if (clength != -1)
999		length = offset + clength;
1000	if (length != -1 && size != -1 && length != size) {
1001		_http_seterr(HTTP_PROTOCOL_ERROR);
1002		goto ouch;
1003	}
1004	if (size == -1)
1005		size = length;
1006
1007	/* fill in stats */
1008	if (us) {
1009		us->size = size;
1010		us->atime = us->mtime = mtime;
1011	}
1012
1013	/* too far? */
1014	if (offset > URL->offset) {
1015		_http_seterr(HTTP_PROTOCOL_ERROR);
1016		goto ouch;
1017	}
1018
1019	/* report back real offset and size */
1020	URL->offset = offset;
1021	URL->length = clength;
1022
1023	/* wrap it up in a FILE */
1024	if ((f = chunked ? _http_funopen(fd) : fdopen(fd, "r")) == NULL) {
1025		_fetch_syserr();
1026		goto ouch;
1027	}
1028
1029	if (url != URL)
1030		fetchFreeURL(url);
1031	if (purl)
1032		fetchFreeURL(purl);
1033
1034	if (HTTP_ERROR(code)) {
1035		_http_print_html(stderr, f);
1036		fclose(f);
1037		f = NULL;
1038	}
1039
1040	return (f);
1041
1042ouch:
1043	if (url != URL)
1044		fetchFreeURL(url);
1045	if (purl)
1046		fetchFreeURL(purl);
1047	if (fd != -1)
1048		close(fd);
1049	return (NULL);
1050}
1051
1052
1053/*****************************************************************************
1054 * Entry points
1055 */
1056
1057/*
1058 * Retrieve and stat a file by HTTP
1059 */
1060FILE *
1061fetchXGetHTTP(struct url *URL, struct url_stat *us, const char *flags)
1062{
1063	return (_http_request(URL, "GET", us, _http_get_proxy(), flags));
1064}
1065
1066/*
1067 * Retrieve a file by HTTP
1068 */
1069FILE *
1070fetchGetHTTP(struct url *URL, const char *flags)
1071{
1072	return (fetchXGetHTTP(URL, NULL, flags));
1073}
1074
1075/*
1076 * Store a file by HTTP
1077 */
1078FILE *
1079fetchPutHTTP(struct url *URL __unused, const char *flags __unused)
1080{
1081	warnx("fetchPutHTTP(): not implemented");
1082	return (NULL);
1083}
1084
1085/*
1086 * Get an HTTP document's metadata
1087 */
1088int
1089fetchStatHTTP(struct url *URL, struct url_stat *us, const char *flags)
1090{
1091	FILE *f;
1092
1093	if ((f = _http_request(URL, "HEAD", us, _http_get_proxy(), flags)) == NULL)
1094		return (-1);
1095	fclose(f);
1096	return (0);
1097}
1098
1099/*
1100 * List a directory
1101 */
1102struct url_ent *
1103fetchListHTTP(struct url *url __unused, const char *flags __unused)
1104{
1105	warnx("fetchListHTTP(): not implemented");
1106	return (NULL);
1107}
1108