http.c revision 93399 
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 93399 2002-03-29 22:43:43Z markm $");
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
83/* Maximum number of redirects to follow */
84#define MAX_REDIRECT 5
85
86/* Symbolic names for reply codes we care about */
87#define HTTP_OK			200
88#define HTTP_PARTIAL		206
89#define HTTP_MOVED_PERM		301
90#define HTTP_MOVED_TEMP		302
91#define HTTP_SEE_OTHER		303
92#define HTTP_NEED_AUTH		401
93#define HTTP_NEED_PROXY_AUTH	407
94#define HTTP_PROTOCOL_ERROR	999
95
96#define HTTP_REDIRECT(xyz) ((xyz) == HTTP_MOVED_PERM \
97			    || (xyz) == HTTP_MOVED_TEMP \
98			    || (xyz) == HTTP_SEE_OTHER)
99
100#define HTTP_ERROR(xyz) ((xyz) > 400 && (xyz) < 599)
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	ssize_t		 b_len;
113	int		 b_pos;
114	int		 eof;
115	int		 error;
116	size_t		 chunksize;
117#ifndef NDEBUG
118	size_t		 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 < c->buf + c->b_len; ++p) {
137		if (*p == ';')
138			break;
139		if (!ishexnumber(*p))
140			return (-1);
141		if (isdigit(*p)) {
142			c->chunksize = c->chunksize * 16 +
143			    *p - '0';
144		} else {
145			c->chunksize = c->chunksize * 16 +
146			    10 + tolower(*p) - 'a';
147		}
148	}
149
150#ifndef NDEBUG
151	if (fetchDebug) {
152		c->total += c->chunksize;
153		if (c->chunksize == 0)
154			fprintf(stderr, "_http_fillbuf(): "
155			    "end of last chunk\n");
156		else
157			fprintf(stderr, "_http_fillbuf(): "
158			    "new chunk: %lu (%lu)\n",
159			    (unsigned long)c->chunksize, (unsigned long)c->total);
160	}
161#endif
162
163	return (c->chunksize);
164}
165
166/*
167 * Fill the input buffer, do chunk decoding on the fly
168 */
169static int
170_http_fillbuf(struct cookie *c)
171{
172	if (c->error)
173		return (-1);
174	if (c->eof)
175		return (0);
176
177	if (c->chunksize == 0) {
178		switch (_http_new_chunk(c)) {
179		case -1:
180			c->error = 1;
181			return (-1);
182		case 0:
183			c->eof = 1;
184			return (0);
185		}
186	}
187
188	if (c->b_size < c->chunksize) {
189		char *tmp;
190
191		if ((tmp = realloc(c->buf, c->chunksize)) == NULL)
192			return (-1);
193		c->buf = tmp;
194		c->b_size = c->chunksize;
195	}
196
197	if ((c->b_len = read(c->fd, c->buf, c->chunksize)) == -1)
198		return (-1);
199	c->chunksize -= c->b_len;
200
201	if (c->chunksize == 0) {
202		char endl;
203		if (read(c->fd, &endl, 1) == -1 ||
204		    read(c->fd, &endl, 1) == -1)
205			return (-1);
206	}
207
208	c->b_pos = 0;
209
210	return (c->b_len);
211}
212
213/*
214 * Read function
215 */
216static int
217_http_readfn(void *v, char *buf, int len)
218{
219	struct cookie *c = (struct cookie *)v;
220	int l, pos;
221
222	if (c->error)
223		return (-1);
224	if (c->eof)
225		return (0);
226
227	for (pos = 0; len > 0; pos += l, len -= l) {
228		/* empty buffer */
229		if (!c->buf || c->b_pos == c->b_len)
230			if (_http_fillbuf(c) < 1)
231				break;
232		l = c->b_len - c->b_pos;
233		if (len < l)
234			l = len;
235		bcopy(c->buf + c->b_pos, buf + pos, l);
236		c->b_pos += l;
237	}
238
239	if (!pos && c->error)
240		return (-1);
241	return (pos);
242}
243
244/*
245 * Write function
246 */
247static int
248_http_writefn(void *v, const char *buf, int len)
249{
250	struct cookie *c = (struct cookie *)v;
251
252	return (write(c->fd, buf, len));
253}
254
255/*
256 * Close function
257 */
258static int
259_http_closefn(void *v)
260{
261	struct cookie *c = (struct cookie *)v;
262	int r;
263
264	r = close(c->fd);
265	if (c->buf)
266		free(c->buf);
267	free(c);
268	return (r);
269}
270
271/*
272 * Wrap a file descriptor up
273 */
274static FILE *
275_http_funopen(int fd)
276{
277	struct cookie *c;
278	FILE *f;
279
280	if ((c = calloc(1, sizeof *c)) == NULL) {
281		_fetch_syserr();
282		return (NULL);
283	}
284	c->fd = fd;
285	f = funopen(c, _http_readfn, _http_writefn, NULL, _http_closefn);
286	if (f == NULL) {
287		_fetch_syserr();
288		free(c);
289		return (NULL);
290	}
291	return (f);
292}
293
294
295/*****************************************************************************
296 * Helper functions for talking to the server and parsing its replies
297 */
298
299/* Header types */
300typedef enum {
301	hdr_syserror = -2,
302	hdr_error = -1,
303	hdr_end = 0,
304	hdr_unknown = 1,
305	hdr_content_length,
306	hdr_content_range,
307	hdr_last_modified,
308	hdr_location,
309	hdr_transfer_encoding,
310	hdr_www_authenticate
311} hdr_t;
312
313/* Names of interesting headers */
314static struct {
315	hdr_t		 num;
316	const char	*name;
317} hdr_names[] = {
318	{ hdr_content_length,		"Content-Length" },
319	{ hdr_content_range,		"Content-Range" },
320	{ hdr_last_modified,		"Last-Modified" },
321	{ hdr_location,			"Location" },
322	{ hdr_transfer_encoding,	"Transfer-Encoding" },
323	{ hdr_www_authenticate,		"WWW-Authenticate" },
324	{ hdr_unknown,			NULL },
325};
326
327static char		*reply_buf;
328static size_t		 reply_size;
329static size_t		 reply_length;
330
331/*
332 * Send a formatted line; optionally echo to terminal
333 */
334static int
335_http_cmd(int fd, const char *fmt, ...)
336{
337	va_list ap;
338	size_t len;
339	char *msg;
340	int r;
341
342	va_start(ap, fmt);
343	len = vasprintf(&msg, fmt, ap);
344	va_end(ap);
345
346	if (msg == NULL) {
347		errno = ENOMEM;
348		_fetch_syserr();
349		return (-1);
350	}
351
352	r = _fetch_putln(fd, msg, len);
353	free(msg);
354
355	if (r == -1) {
356		_fetch_syserr();
357		return (-1);
358	}
359
360	return (0);
361}
362
363/*
364 * Get and parse status line
365 */
366static int
367_http_get_reply(int fd)
368{
369	char *p;
370
371	if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
372		return (-1);
373	/*
374	 * A valid status line looks like "HTTP/m.n xyz reason" where m
375	 * and n are the major and minor protocol version numbers and xyz
376	 * is the reply code.
377	 * Unfortunately, there are servers out there (NCSA 1.5.1, to name
378	 * just one) that do not send a version number, so we can't rely
379	 * on finding one, but if we do, insist on it being 1.0 or 1.1.
380	 * We don't care about the reason phrase.
381	 */
382	if (strncmp(reply_buf, "HTTP", 4) != 0)
383		return (HTTP_PROTOCOL_ERROR);
384	p = reply_buf + 4;
385	if (*p == '/') {
386		if (p[1] != '1' || p[2] != '.' || (p[3] != '0' && p[3] != '1'))
387			return (HTTP_PROTOCOL_ERROR);
388		p += 4;
389	}
390	if (*p != ' ' || !isdigit(p[1]) || !isdigit(p[2]) || !isdigit(p[3]))
391		return (HTTP_PROTOCOL_ERROR);
392
393	return ((p[1] - '0') * 100 + (p[2] - '0') * 10 + (p[3] - '0'));
394}
395
396/*
397 * Check a header; if the type matches the given string, return a pointer
398 * to the beginning of the value.
399 */
400static const char *
401_http_match(const char *str, const char *hdr)
402{
403	while (*str && *hdr && tolower(*str++) == tolower(*hdr++))
404		/* nothing */;
405	if (*str || *hdr != ':')
406		return (NULL);
407	while (*hdr && isspace(*++hdr))
408		/* nothing */;
409	return (hdr);
410}
411
412/*
413 * Get the next header and return the appropriate symbolic code.
414 */
415static hdr_t
416_http_next_header(int fd, const char **p)
417{
418	int i;
419
420	if (_fetch_getln(fd, &reply_buf, &reply_size, &reply_length) == -1)
421		return (hdr_syserror);
422	while (reply_length && isspace(reply_buf[reply_length-1]))
423		reply_length--;
424	reply_buf[reply_length] = 0;
425	if (reply_length == 0)
426	return (hdr_end);
427	/*
428	 * We could check for malformed headers but we don't really care.
429	 * A valid header starts with a token immediately followed by a
430	 * colon; a token is any sequence of non-control, non-whitespace
431	 * characters except "()<>@,;:\\\"{}".
432	 */
433	for (i = 0; hdr_names[i].num != hdr_unknown; i++)
434		if ((*p = _http_match(hdr_names[i].name, reply_buf)) != NULL)
435			return (hdr_names[i].num);
436	return (hdr_unknown);
437}
438
439/*
440 * Parse a last-modified header
441 */
442static int
443_http_parse_mtime(const char *p, time_t *mtime)
444{
445	char locale[64], *r;
446	struct tm tm;
447
448	strncpy(locale, setlocale(LC_TIME, NULL), sizeof locale);
449	setlocale(LC_TIME, "C");
450	r = strptime(p, "%a, %d %b %Y %H:%M:%S GMT", &tm);
451	/* XXX should add support for date-2 and date-3 */
452	setlocale(LC_TIME, locale);
453	if (r == NULL)
454		return (-1);
455	DEBUG(fprintf(stderr, "last modified: [%04d-%02d-%02d "
456		  "%02d:%02d:%02d]\n",
457		  tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday,
458		  tm.tm_hour, tm.tm_min, tm.tm_sec));
459	*mtime = timegm(&tm);
460	return (0);
461}
462
463/*
464 * Parse a content-length header
465 */
466static int
467_http_parse_length(const char *p, off_t *length)
468{
469	off_t len;
470
471	for (len = 0; *p && isdigit(*p); ++p)
472		len = len * 10 + (*p - '0');
473	if (*p)
474		return (-1);
475	DEBUG(fprintf(stderr, "content length: [%lld]\n",
476	    (long long)len));
477	*length = len;
478	return (0);
479}
480
481/*
482 * Parse a content-range header
483 */
484static int
485_http_parse_range(const char *p, off_t *offset, off_t *length, off_t *size)
486{
487	off_t first, last, len;
488
489	if (strncasecmp(p, "bytes ", 6) != 0)
490		return (-1);
491	for (first = 0, p += 6; *p && isdigit(*p); ++p)
492		first = first * 10 + *p - '0';
493	if (*p != '-')
494		return (-1);
495	for (last = 0, ++p; *p && isdigit(*p); ++p)
496		last = last * 10 + *p - '0';
497	if (first > last || *p != '/')
498		return (-1);
499	for (len = 0, ++p; *p && isdigit(*p); ++p)
500		len = len * 10 + *p - '0';
501	if (*p || len < last - first + 1)
502		return (-1);
503	DEBUG(fprintf(stderr, "content range: [%lld-%lld/%lld]\n",
504	    (long long)first, (long long)last, (long long)len));
505	*offset = first;
506	*length = last - first + 1;
507	*size = len;
508	return (0);
509}
510
511
512/*****************************************************************************
513 * Helper functions for authorization
514 */
515
516/*
517 * Base64 encoding
518 */
519static char *
520_http_base64(const char *src)
521{
522	static const char base64[] =
523	    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
524	    "abcdefghijklmnopqrstuvwxyz"
525	    "0123456789+/";
526	char *str, *dst;
527	size_t l;
528	int t, r;
529
530	l = strlen(src);
531	if ((str = malloc(((l + 2) / 3) * 4)) == NULL)
532		return (NULL);
533	dst = str;
534	r = 0;
535
536	while (l >= 3) {
537		t = (src[0] << 16) | (src[1] << 8) | src[2];
538		dst[0] = base64[(t >> 18) & 0x3f];
539		dst[1] = base64[(t >> 12) & 0x3f];
540		dst[2] = base64[(t >> 6) & 0x3f];
541		dst[3] = base64[(t >> 0) & 0x3f];
542		src += 3; l -= 3;
543		dst += 4; r += 4;
544	}
545
546	switch (l) {
547	case 2:
548		t = (src[0] << 16) | (src[1] << 8);
549		dst[0] = base64[(t >> 18) & 0x3f];
550		dst[1] = base64[(t >> 12) & 0x3f];
551		dst[2] = base64[(t >> 6) & 0x3f];
552		dst[3] = '=';
553		dst += 4;
554		r += 4;
555		break;
556	case 1:
557		t = src[0] << 16;
558		dst[0] = base64[(t >> 18) & 0x3f];
559		dst[1] = base64[(t >> 12) & 0x3f];
560		dst[2] = dst[3] = '=';
561		dst += 4;
562		r += 4;
563		break;
564	case 0:
565		break;
566	}
567
568	*dst = 0;
569	return (str);
570}
571
572/*
573 * Encode username and password
574 */
575static int
576_http_basic_auth(int fd, const char *hdr, const char *usr, const char *pwd)
577{
578	char *upw, *auth;
579	int r;
580
581	DEBUG(fprintf(stderr, "usr: [%s]\n", usr));
582	DEBUG(fprintf(stderr, "pwd: [%s]\n", pwd));
583	if (asprintf(&upw, "%s:%s", usr, pwd) == -1)
584		return (-1);
585	auth = _http_base64(upw);
586	free(upw);
587	if (auth == NULL)
588		return (-1);
589	r = _http_cmd(fd, "%s: Basic %s", hdr, auth);
590	free(auth);
591	return (r);
592}
593
594/*
595 * Send an authorization header
596 */
597static int
598_http_authorize(int fd, const char *hdr, const char *p)
599{
600	/* basic authorization */
601	if (strncasecmp(p, "basic:", 6) == 0) {
602		char *user, *pwd, *str;
603		int r;
604
605		/* skip realm */
606		for (p += 6; *p && *p != ':'; ++p)
607			/* nothing */ ;
608		if (!*p || strchr(++p, ':') == NULL)
609			return (-1);
610		if ((str = strdup(p)) == NULL)
611			return (-1); /* XXX */
612		user = str;
613		pwd = strchr(str, ':');
614		*pwd++ = '\0';
615		r = _http_basic_auth(fd, hdr, user, pwd);
616		free(str);
617		return (r);
618	}
619	return (-1);
620}
621
622
623/*****************************************************************************
624 * Helper functions for connecting to a server or proxy
625 */
626
627/*
628 * Connect to the correct HTTP server or proxy.
629 */
630static int
631_http_connect(struct url *URL, struct url *purl, const char *flags)
632{
633	int verbose;
634	int af, fd;
635
636#ifdef INET6
637	af = AF_UNSPEC;
638#else
639	af = AF_INET;
640#endif
641
642	verbose = CHECK_FLAG('v');
643	if (CHECK_FLAG('4'))
644		af = AF_INET;
645#ifdef INET6
646	else if (CHECK_FLAG('6'))
647		af = AF_INET6;
648#endif
649
650	if (purl) {
651		URL = purl;
652	} else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) {
653		/* can't talk http to an ftp server */
654		/* XXX should set an error code */
655		return (-1);
656	}
657
658	if ((fd = _fetch_connect(URL->host, URL->port, af, verbose)) == -1)
659		/* _fetch_connect() has already set an error code */
660		return (-1);
661	return (fd);
662}
663
664static struct url *
665_http_get_proxy(void)
666{
667	struct url *purl;
668	char *p;
669
670	if (((p = getenv("HTTP_PROXY")) || (p = getenv("http_proxy"))) &&
671	    (purl = fetchParseURL(p))) {
672		if (!*purl->scheme)
673			strcpy(purl->scheme, SCHEME_HTTP);
674		if (!purl->port)
675			purl->port = _fetch_default_proxy_port(purl->scheme);
676		if (strcasecmp(purl->scheme, SCHEME_HTTP) == 0)
677			return (purl);
678		fetchFreeURL(purl);
679	}
680	return (NULL);
681}
682
683static void
684_http_print_html(FILE *out, FILE *in)
685{
686	size_t len;
687	char *line, *p, *q;
688	int comment, tag;
689
690	comment = tag = 0;
691	while ((line = fgetln(in, &len)) != NULL) {
692		while (len && isspace(line[len - 1]))
693			--len;
694		for (p = q = line; q < line + len; ++q) {
695			if (comment && *q == '-') {
696				if (q + 2 < line + len &&
697				    strcmp(q, "-->") == 0) {
698					tag = comment = 0;
699					q += 2;
700				}
701			} else if (tag && !comment && *q == '>') {
702				p = q + 1;
703				tag = 0;
704			} else if (!tag && *q == '<') {
705				if (q > p)
706					fwrite(p, q - p, 1, out);
707				tag = 1;
708				if (q + 3 < line + len &&
709				    strcmp(q, "<!--") == 0) {
710					comment = 1;
711					q += 3;
712				}
713			}
714		}
715		if (!tag && q > p)
716			fwrite(p, q - p, 1, out);
717		fputc('\n', out);
718	}
719}
720
721
722/*****************************************************************************
723 * Core
724 */
725
726/*
727 * Send a request and process the reply
728 */
729FILE *
730_http_request(struct url *URL, const char *op, struct url_stat *us,
731    struct url *purl, const char *flags)
732{
733	struct url *url, *new;
734	int chunked, direct, need_auth, noredirect, verbose;
735	int code, fd, i, n;
736	off_t offset, clength, length, size;
737	time_t mtime;
738	const char *p;
739	FILE *f;
740	hdr_t h;
741	char *host;
742#ifdef INET6
743	char hbuf[MAXHOSTNAMELEN + 1];
744#endif
745
746	direct = CHECK_FLAG('d');
747	noredirect = CHECK_FLAG('A');
748	verbose = CHECK_FLAG('v');
749
750	if (direct && purl) {
751		fetchFreeURL(purl);
752		purl = NULL;
753	}
754
755	/* try the provided URL first */
756	url = URL;
757
758	/* if the A flag is set, we only get one try */
759	n = noredirect ? 1 : MAX_REDIRECT;
760	i = 0;
761
762	need_auth = 0;
763	do {
764		new = NULL;
765		chunked = 0;
766		offset = 0;
767		clength = -1;
768		length = -1;
769		size = -1;
770		mtime = 0;
771
772		/* check port */
773		if (!url->port)
774			url->port = _fetch_default_port(url->scheme);
775
776		/* were we redirected to an FTP URL? */
777		if (purl == NULL && strcmp(url->scheme, SCHEME_FTP) == 0) {
778			if (strcmp(op, "GET") == 0)
779				return (_ftp_request(url, "RETR", us, purl, flags));
780			else if (strcmp(op, "HEAD") == 0)
781				return (_ftp_request(url, "STAT", us, purl, flags));
782		}
783
784		/* connect to server or proxy */
785		if ((fd = _http_connect(url, purl, flags)) == -1)
786			goto ouch;
787
788		host = url->host;
789#ifdef INET6
790		if (strchr(url->host, ':')) {
791			snprintf(hbuf, sizeof(hbuf), "[%s]", url->host);
792			host = hbuf;
793		}
794#endif
795
796		/* send request */
797		if (verbose)
798			_fetch_info("requesting %s://%s:%d%s",
799			    url->scheme, host, url->port, url->doc);
800		if (purl) {
801			_http_cmd(fd, "%s %s://%s:%d%s HTTP/1.1",
802			    op, url->scheme, host, url->port, url->doc);
803		} else {
804			_http_cmd(fd, "%s %s HTTP/1.1",
805			    op, url->doc);
806		}
807
808		/* virtual host */
809		if (url->port == _fetch_default_port(url->scheme))
810			_http_cmd(fd, "Host: %s", host);
811		else
812			_http_cmd(fd, "Host: %s:%d", host, url->port);
813
814		/* proxy authorization */
815		if (purl) {
816			if (*purl->user || *purl->pwd)
817				_http_basic_auth(fd, "Proxy-Authorization",
818				    purl->user, purl->pwd);
819			else if ((p = getenv("HTTP_PROXY_AUTH")) != NULL && *p != '\0')
820				_http_authorize(fd, "Proxy-Authorization", p);
821		}
822
823		/* server authorization */
824		if (need_auth || *url->user || *url->pwd) {
825			if (*url->user || *url->pwd)
826				_http_basic_auth(fd, "Authorization", url->user, url->pwd);
827			else if ((p = getenv("HTTP_AUTH")) != NULL && *p != '\0')
828				_http_authorize(fd, "Authorization", p);
829			else if (fetchAuthMethod && fetchAuthMethod(url) == 0) {
830				_http_basic_auth(fd, "Authorization", url->user, url->pwd);
831			} else {
832				_http_seterr(HTTP_NEED_AUTH);
833				goto ouch;
834			}
835		}
836
837		/* other headers */
838		if ((p = getenv("HTTP_USER_AGENT")) != NULL && *p != '\0')
839			_http_cmd(fd, "User-Agent: %s", p);
840		else
841			_http_cmd(fd, "User-Agent: %s " _LIBFETCH_VER, _getprogname());
842		if (url->offset)
843			_http_cmd(fd, "Range: bytes=%lld-", (long long)url->offset);
844		_http_cmd(fd, "Connection: close");
845		_http_cmd(fd, "");
846
847		/* get reply */
848		switch ((code = _http_get_reply(fd))) {
849		case HTTP_OK:
850		case HTTP_PARTIAL:
851			/* fine */
852			break;
853		case HTTP_MOVED_PERM:
854		case HTTP_MOVED_TEMP:
855		case HTTP_SEE_OTHER:
856			/*
857			 * Not so fine, but we still have to read the headers to
858			 * get the new location.
859			 */
860			break;
861		case HTTP_NEED_AUTH:
862			if (need_auth) {
863				/*
864				 * We already sent out authorization code, so there's
865				 * nothing more we can do.
866				 */
867				_http_seterr(code);
868				goto ouch;
869			}
870			/* try again, but send the password this time */
871			if (verbose)
872				_fetch_info("server requires authorization");
873			break;
874		case HTTP_NEED_PROXY_AUTH:
875			/*
876			 * If we're talking to a proxy, we already sent our proxy
877			 * authorization code, so there's nothing more we can do.
878			 */
879			_http_seterr(code);
880			goto ouch;
881		case HTTP_PROTOCOL_ERROR:
882			/* fall through */
883		case -1:
884			_fetch_syserr();
885			goto ouch;
886		default:
887			_http_seterr(code);
888			if (!verbose)
889				goto ouch;
890			/* fall through so we can get the full error message */
891		}
892
893		/* get headers */
894		do {
895			switch ((h = _http_next_header(fd, &p))) {
896			case hdr_syserror:
897				_fetch_syserr();
898				goto ouch;
899			case hdr_error:
900				_http_seterr(HTTP_PROTOCOL_ERROR);
901				goto ouch;
902			case hdr_content_length:
903				_http_parse_length(p, &clength);
904				break;
905			case hdr_content_range:
906				_http_parse_range(p, &offset, &length, &size);
907				break;
908			case hdr_last_modified:
909				_http_parse_mtime(p, &mtime);
910				break;
911			case hdr_location:
912				if (!HTTP_REDIRECT(code))
913					break;
914				if (new)
915					free(new);
916				if (verbose)
917					_fetch_info("%d redirect to %s", code, p);
918				if (*p == '/')
919					/* absolute path */
920					new = fetchMakeURL(url->scheme, url->host, url->port, p,
921					    url->user, url->pwd);
922				else
923					new = fetchParseURL(p);
924				if (new == NULL) {
925					/* XXX should set an error code */
926					DEBUG(fprintf(stderr, "failed to parse new URL\n"));
927					goto ouch;
928				}
929				if (!*new->user && !*new->pwd) {
930					strcpy(new->user, url->user);
931					strcpy(new->pwd, url->pwd);
932				}
933				new->offset = url->offset;
934				new->length = url->length;
935				break;
936			case hdr_transfer_encoding:
937				/* XXX weak test*/
938				chunked = (strcasecmp(p, "chunked") == 0);
939				break;
940			case hdr_www_authenticate:
941				if (code != HTTP_NEED_AUTH)
942					break;
943				/* if we were smarter, we'd check the method and realm */
944				break;
945			case hdr_end:
946				/* fall through */
947			case hdr_unknown:
948				/* ignore */
949				break;
950			}
951		} while (h > hdr_end);
952
953		/* we have a hit or an error */
954		if (code == HTTP_OK || code == HTTP_PARTIAL || HTTP_ERROR(code))
955			break;
956
957		/* we need to provide authentication */
958		if (code == HTTP_NEED_AUTH) {
959			need_auth = 1;
960			close(fd);
961			fd = -1;
962			continue;
963		}
964
965		/* all other cases: we got a redirect */
966		need_auth = 0;
967		close(fd);
968		fd = -1;
969		if (!new) {
970			DEBUG(fprintf(stderr, "redirect with no new location\n"));
971			break;
972		}
973		if (url != URL)
974			fetchFreeURL(url);
975		url = new;
976	} while (++i < n);
977
978	/* we failed, or ran out of retries */
979	if (fd == -1) {
980		_http_seterr(code);
981		goto ouch;
982	}
983
984	DEBUG(fprintf(stderr, "offset %lld, length %lld,"
985		  " size %lld, clength %lld\n",
986		  (long long)offset, (long long)length,
987		  (long long)size, (long long)clength));
988
989	/* check for inconsistencies */
990	if (clength != -1 && length != -1 && clength != length) {
991		_http_seterr(HTTP_PROTOCOL_ERROR);
992		goto ouch;
993	}
994	if (clength == -1)
995		clength = length;
996	if (clength != -1)
997		length = offset + clength;
998	if (length != -1 && size != -1 && length != size) {
999		_http_seterr(HTTP_PROTOCOL_ERROR);
1000		goto ouch;
1001	}
1002	if (size == -1)
1003		size = length;
1004
1005	/* fill in stats */
1006	if (us) {
1007		us->size = size;
1008		us->atime = us->mtime = mtime;
1009	}
1010
1011	/* too far? */
1012	if (offset > URL->offset) {
1013		_http_seterr(HTTP_PROTOCOL_ERROR);
1014		goto ouch;
1015	}
1016
1017	/* report back real offset and size */
1018	URL->offset = offset;
1019	URL->length = clength;
1020
1021	/* wrap it up in a FILE */
1022	if ((f = chunked ? _http_funopen(fd) : fdopen(fd, "r")) == NULL) {
1023		_fetch_syserr();
1024		goto ouch;
1025	}
1026
1027	if (url != URL)
1028		fetchFreeURL(url);
1029	if (purl)
1030		fetchFreeURL(purl);
1031
1032	if (HTTP_ERROR(code)) {
1033		_http_print_html(stderr, f);
1034		fclose(f);
1035		f = NULL;
1036	}
1037
1038	return (f);
1039
1040ouch:
1041	if (url != URL)
1042		fetchFreeURL(url);
1043	if (purl)
1044		fetchFreeURL(purl);
1045	if (fd != -1)
1046		close(fd);
1047	return (NULL);
1048}
1049
1050
1051/*****************************************************************************
1052 * Entry points
1053 */
1054
1055/*
1056 * Retrieve and stat a file by HTTP
1057 */
1058FILE *
1059fetchXGetHTTP(struct url *URL, struct url_stat *us, const char *flags)
1060{
1061	return (_http_request(URL, "GET", us, _http_get_proxy(), flags));
1062}
1063
1064/*
1065 * Retrieve a file by HTTP
1066 */
1067FILE *
1068fetchGetHTTP(struct url *URL, const char *flags)
1069{
1070	return (fetchXGetHTTP(URL, NULL, flags));
1071}
1072
1073/*
1074 * Store a file by HTTP
1075 */
1076FILE *
1077fetchPutHTTP(struct url *URL __unused, const char *flags __unused)
1078{
1079	warnx("fetchPutHTTP(): not implemented");
1080	return (NULL);
1081}
1082
1083/*
1084 * Get an HTTP document's metadata
1085 */
1086int
1087fetchStatHTTP(struct url *URL, struct url_stat *us, const char *flags)
1088{
1089	FILE *f;
1090
1091	if ((f = _http_request(URL, "HEAD", us, _http_get_proxy(), flags)) == NULL)
1092		return (-1);
1093	fclose(f);
1094	return (0);
1095}
1096
1097/*
1098 * List a directory
1099 */
1100struct url_ent *
1101fetchListHTTP(struct url *url __unused, const char *flags __unused)
1102{
1103	warnx("fetchListHTTP(): not implemented");
1104	return (NULL);
1105}
1106