http.c revision 106044 
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 106044 2002-10-27 15:43:40Z 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
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 httpio
108{
109	conn_t		*conn;		/* connection */
110	int		 chunked;	/* chunked mode */
111	char		*buf;		/* chunk buffer */
112	size_t		 bufsize;	/* size of chunk buffer */
113	ssize_t		 buflen;	/* amount of data currently in buffer */
114	int		 bufpos;	/* current read offset in buffer */
115	int		 eof;		/* end-of-file flag */
116	int		 error;		/* error flag */
117	size_t		 chunksize;	/* remaining size of current chunk */
118#ifndef NDEBUG
119	size_t		 total;
120#endif
121};
122
123/*
124 * Get next chunk header
125 */
126static int
127_http_new_chunk(struct httpio *io)
128{
129	char *p;
130
131	if (_fetch_getln(io->conn) == -1)
132		return (-1);
133
134	if (io->conn->buflen < 2 || !ishexnumber(*io->conn->buf))
135		return (-1);
136
137	for (p = io->conn->buf; *p && !isspace(*p); ++p) {
138		if (*p == ';')
139			break;
140		if (!ishexnumber(*p))
141			return (-1);
142		if (isdigit(*p)) {
143			io->chunksize = io->chunksize * 16 +
144			    *p - '0';
145		} else {
146			io->chunksize = io->chunksize * 16 +
147			    10 + tolower(*p) - 'a';
148		}
149	}
150
151#ifndef NDEBUG
152	if (fetchDebug) {
153		io->total += io->chunksize;
154		if (io->chunksize == 0)
155			fprintf(stderr, "_http_fillbuf(): "
156			    "end of last chunk\n");
157		else
158			fprintf(stderr, "_http_fillbuf(): "
159			    "new chunk: %lu (%lu)\n",
160			    (unsigned long)io->chunksize, (unsigned long)io->total);
161	}
162#endif
163
164	return (io->chunksize);
165}
166
167/*
168 * Grow the input buffer to at least len bytes
169 */
170static inline int
171_http_growbuf(struct httpio *io, size_t len)
172{
173	char *tmp;
174
175	if (io->bufsize >= len)
176		return (0);
177
178	if ((tmp = realloc(io->buf, len)) == NULL)
179		return (-1);
180	io->buf = tmp;
181	io->bufsize = len;
182	return (0);
183}
184
185/*
186 * Fill the input buffer, do chunk decoding on the fly
187 */
188static int
189_http_fillbuf(struct httpio *io, size_t len)
190{
191	if (io->error)
192		return (-1);
193	if (io->eof)
194		return (0);
195
196	if (io->chunked == 0) {
197		if (_http_growbuf(io, len) == -1)
198			return (-1);
199		if ((io->buflen = _fetch_read(io->conn, io->buf, len)) == -1)
200			return (-1);
201		io->bufpos = 0;
202		return (io->buflen);
203	}
204
205	if (io->chunksize == 0) {
206		switch (_http_new_chunk(io)) {
207		case -1:
208			io->error = 1;
209			return (-1);
210		case 0:
211			io->eof = 1;
212			return (0);
213		}
214	}
215
216	if (len > io->chunksize)
217		len = io->chunksize;
218	if (_http_growbuf(io, len) == -1)
219		return (-1);
220	if ((io->buflen = _fetch_read(io->conn, io->buf, len)) == -1)
221		return (-1);
222	io->chunksize -= io->buflen;
223
224	if (io->chunksize == 0) {
225		char endl[2];
226
227		if (_fetch_read(io->conn, endl, 2) != 2 ||
228		    endl[0] != '\r' || endl[1] != '\n')
229			return (-1);
230	}
231
232	io->bufpos = 0;
233
234	return (io->buflen);
235}
236
237/*
238 * Read function
239 */
240static int
241_http_readfn(void *v, char *buf, int len)
242{
243	struct httpio *io = (struct httpio *)v;
244	int l, pos;
245
246	if (io->error)
247		return (-1);
248	if (io->eof)
249		return (0);
250
251	for (pos = 0; len > 0; pos += l, len -= l) {
252		/* empty buffer */
253		if (!io->buf || io->bufpos == io->buflen)
254			if (_http_fillbuf(io, len) < 1)
255				break;
256		l = io->buflen - io->bufpos;
257		if (len < l)
258			l = len;
259		bcopy(io->buf + io->bufpos, buf + pos, l);
260		io->bufpos += l;
261	}
262
263	if (!pos && io->error)
264		return (-1);
265	return (pos);
266}
267
268/*
269 * Write function
270 */
271static int
272_http_writefn(void *v, const char *buf, int len)
273{
274	struct httpio *io = (struct httpio *)v;
275
276	return (_fetch_write(io->conn, buf, len));
277}
278
279/*
280 * Close function
281 */
282static int
283_http_closefn(void *v)
284{
285	struct httpio *io = (struct httpio *)v;
286	int r;
287
288	r = _fetch_close(io->conn);
289	if (io->buf)
290		free(io->buf);
291	free(io);
292	return (r);
293}
294
295/*
296 * Wrap a file descriptor up
297 */
298static FILE *
299_http_funopen(conn_t *conn, int chunked)
300{
301	struct httpio *io;
302	FILE *f;
303
304	if ((io = calloc(1, sizeof *io)) == NULL) {
305		_fetch_syserr();
306		return (NULL);
307	}
308	io->conn = conn;
309	io->chunked = chunked;
310	f = funopen(io, _http_readfn, _http_writefn, NULL, _http_closefn);
311	if (f == NULL) {
312		_fetch_syserr();
313		free(io);
314		return (NULL);
315	}
316	return (f);
317}
318
319
320/*****************************************************************************
321 * Helper functions for talking to the server and parsing its replies
322 */
323
324/* Header types */
325typedef enum {
326	hdr_syserror = -2,
327	hdr_error = -1,
328	hdr_end = 0,
329	hdr_unknown = 1,
330	hdr_content_length,
331	hdr_content_range,
332	hdr_last_modified,
333	hdr_location,
334	hdr_transfer_encoding,
335	hdr_www_authenticate
336} hdr_t;
337
338/* Names of interesting headers */
339static struct {
340	hdr_t		 num;
341	const char	*name;
342} hdr_names[] = {
343	{ hdr_content_length,		"Content-Length" },
344	{ hdr_content_range,		"Content-Range" },
345	{ hdr_last_modified,		"Last-Modified" },
346	{ hdr_location,			"Location" },
347	{ hdr_transfer_encoding,	"Transfer-Encoding" },
348	{ hdr_www_authenticate,		"WWW-Authenticate" },
349	{ hdr_unknown,			NULL },
350};
351
352/*
353 * Send a formatted line; optionally echo to terminal
354 */
355static int
356_http_cmd(conn_t *conn, const char *fmt, ...)
357{
358	va_list ap;
359	size_t len;
360	char *msg;
361	int r;
362
363	va_start(ap, fmt);
364	len = vasprintf(&msg, fmt, ap);
365	va_end(ap);
366
367	if (msg == NULL) {
368		errno = ENOMEM;
369		_fetch_syserr();
370		return (-1);
371	}
372
373	r = _fetch_putln(conn, msg, len);
374	free(msg);
375
376	if (r == -1) {
377		_fetch_syserr();
378		return (-1);
379	}
380
381	return (0);
382}
383
384/*
385 * Get and parse status line
386 */
387static int
388_http_get_reply(conn_t *conn)
389{
390	char *p;
391
392	if (_fetch_getln(conn) == -1)
393		return (-1);
394	/*
395	 * A valid status line looks like "HTTP/m.n xyz reason" where m
396	 * and n are the major and minor protocol version numbers and xyz
397	 * is the reply code.
398	 * Unfortunately, there are servers out there (NCSA 1.5.1, to name
399	 * just one) that do not send a version number, so we can't rely
400	 * on finding one, but if we do, insist on it being 1.0 or 1.1.
401	 * We don't care about the reason phrase.
402	 */
403	if (strncmp(conn->buf, "HTTP", 4) != 0)
404		return (HTTP_PROTOCOL_ERROR);
405	p = conn->buf + 4;
406	if (*p == '/') {
407		if (p[1] != '1' || p[2] != '.' || (p[3] != '0' && p[3] != '1'))
408			return (HTTP_PROTOCOL_ERROR);
409		p += 4;
410	}
411	if (*p != ' ' || !isdigit(p[1]) || !isdigit(p[2]) || !isdigit(p[3]))
412		return (HTTP_PROTOCOL_ERROR);
413
414	conn->err = (p[1] - '0') * 100 + (p[2] - '0') * 10 + (p[3] - '0');
415	return (conn->err);
416}
417
418/*
419 * Check a header; if the type matches the given string, return a pointer
420 * to the beginning of the value.
421 */
422static const char *
423_http_match(const char *str, const char *hdr)
424{
425	while (*str && *hdr && tolower(*str++) == tolower(*hdr++))
426		/* nothing */;
427	if (*str || *hdr != ':')
428		return (NULL);
429	while (*hdr && isspace(*++hdr))
430		/* nothing */;
431	return (hdr);
432}
433
434/*
435 * Get the next header and return the appropriate symbolic code.
436 */
437static hdr_t
438_http_next_header(conn_t *conn, const char **p)
439{
440	int i;
441
442	if (_fetch_getln(conn) == -1)
443		return (hdr_syserror);
444	while (conn->buflen && isspace(conn->buf[conn->buflen - 1]))
445		conn->buflen--;
446	conn->buf[conn->buflen] = '\0';
447	if (conn->buflen == 0)
448		return (hdr_end);
449	/*
450	 * We could check for malformed headers but we don't really care.
451	 * A valid header starts with a token immediately followed by a
452	 * colon; a token is any sequence of non-control, non-whitespace
453	 * characters except "()<>@,;:\\\"{}".
454	 */
455	for (i = 0; hdr_names[i].num != hdr_unknown; i++)
456		if ((*p = _http_match(hdr_names[i].name, conn->buf)) != NULL)
457			return (hdr_names[i].num);
458	return (hdr_unknown);
459}
460
461/*
462 * Parse a last-modified header
463 */
464static int
465_http_parse_mtime(const char *p, time_t *mtime)
466{
467	char locale[64], *r;
468	struct tm tm;
469
470	strncpy(locale, setlocale(LC_TIME, NULL), sizeof locale);
471	setlocale(LC_TIME, "C");
472	r = strptime(p, "%a, %d %b %Y %H:%M:%S GMT", &tm);
473	/* XXX should add support for date-2 and date-3 */
474	setlocale(LC_TIME, locale);
475	if (r == NULL)
476		return (-1);
477	DEBUG(fprintf(stderr, "last modified: [%04d-%02d-%02d "
478		  "%02d:%02d:%02d]\n",
479		  tm.tm_year + 1900, tm.tm_mon + 1, tm.tm_mday,
480		  tm.tm_hour, tm.tm_min, tm.tm_sec));
481	*mtime = timegm(&tm);
482	return (0);
483}
484
485/*
486 * Parse a content-length header
487 */
488static int
489_http_parse_length(const char *p, off_t *length)
490{
491	off_t len;
492
493	for (len = 0; *p && isdigit(*p); ++p)
494		len = len * 10 + (*p - '0');
495	if (*p)
496		return (-1);
497	DEBUG(fprintf(stderr, "content length: [%lld]\n",
498	    (long long)len));
499	*length = len;
500	return (0);
501}
502
503/*
504 * Parse a content-range header
505 */
506static int
507_http_parse_range(const char *p, off_t *offset, off_t *length, off_t *size)
508{
509	off_t first, last, len;
510
511	if (strncasecmp(p, "bytes ", 6) != 0)
512		return (-1);
513	for (first = 0, p += 6; *p && isdigit(*p); ++p)
514		first = first * 10 + *p - '0';
515	if (*p != '-')
516		return (-1);
517	for (last = 0, ++p; *p && isdigit(*p); ++p)
518		last = last * 10 + *p - '0';
519	if (first > last || *p != '/')
520		return (-1);
521	for (len = 0, ++p; *p && isdigit(*p); ++p)
522		len = len * 10 + *p - '0';
523	if (*p || len < last - first + 1)
524		return (-1);
525	DEBUG(fprintf(stderr, "content range: [%lld-%lld/%lld]\n",
526	    (long long)first, (long long)last, (long long)len));
527	*offset = first;
528	*length = last - first + 1;
529	*size = len;
530	return (0);
531}
532
533
534/*****************************************************************************
535 * Helper functions for authorization
536 */
537
538/*
539 * Base64 encoding
540 */
541static char *
542_http_base64(const char *src)
543{
544	static const char base64[] =
545	    "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
546	    "abcdefghijklmnopqrstuvwxyz"
547	    "0123456789+/";
548	char *str, *dst;
549	size_t l;
550	int t, r;
551
552	l = strlen(src);
553	if ((str = malloc(((l + 2) / 3) * 4)) == NULL)
554		return (NULL);
555	dst = str;
556	r = 0;
557
558	while (l >= 3) {
559		t = (src[0] << 16) | (src[1] << 8) | src[2];
560		dst[0] = base64[(t >> 18) & 0x3f];
561		dst[1] = base64[(t >> 12) & 0x3f];
562		dst[2] = base64[(t >> 6) & 0x3f];
563		dst[3] = base64[(t >> 0) & 0x3f];
564		src += 3; l -= 3;
565		dst += 4; r += 4;
566	}
567
568	switch (l) {
569	case 2:
570		t = (src[0] << 16) | (src[1] << 8);
571		dst[0] = base64[(t >> 18) & 0x3f];
572		dst[1] = base64[(t >> 12) & 0x3f];
573		dst[2] = base64[(t >> 6) & 0x3f];
574		dst[3] = '=';
575		dst += 4;
576		r += 4;
577		break;
578	case 1:
579		t = src[0] << 16;
580		dst[0] = base64[(t >> 18) & 0x3f];
581		dst[1] = base64[(t >> 12) & 0x3f];
582		dst[2] = dst[3] = '=';
583		dst += 4;
584		r += 4;
585		break;
586	case 0:
587		break;
588	}
589
590	*dst = 0;
591	return (str);
592}
593
594/*
595 * Encode username and password
596 */
597static int
598_http_basic_auth(conn_t *conn, const char *hdr, const char *usr, const char *pwd)
599{
600	char *upw, *auth;
601	int r;
602
603	DEBUG(fprintf(stderr, "usr: [%s]\n", usr));
604	DEBUG(fprintf(stderr, "pwd: [%s]\n", pwd));
605	if (asprintf(&upw, "%s:%s", usr, pwd) == -1)
606		return (-1);
607	auth = _http_base64(upw);
608	free(upw);
609	if (auth == NULL)
610		return (-1);
611	r = _http_cmd(conn, "%s: Basic %s", hdr, auth);
612	free(auth);
613	return (r);
614}
615
616/*
617 * Send an authorization header
618 */
619static int
620_http_authorize(conn_t *conn, const char *hdr, const char *p)
621{
622	/* basic authorization */
623	if (strncasecmp(p, "basic:", 6) == 0) {
624		char *user, *pwd, *str;
625		int r;
626
627		/* skip realm */
628		for (p += 6; *p && *p != ':'; ++p)
629			/* nothing */ ;
630		if (!*p || strchr(++p, ':') == NULL)
631			return (-1);
632		if ((str = strdup(p)) == NULL)
633			return (-1); /* XXX */
634		user = str;
635		pwd = strchr(str, ':');
636		*pwd++ = '\0';
637		r = _http_basic_auth(conn, hdr, user, pwd);
638		free(str);
639		return (r);
640	}
641	return (-1);
642}
643
644
645/*****************************************************************************
646 * Helper functions for connecting to a server or proxy
647 */
648
649/*
650 * Connect to the correct HTTP server or proxy.
651 */
652static conn_t *
653_http_connect(struct url *URL, struct url *purl, const char *flags)
654{
655	conn_t *conn;
656	int verbose;
657	int af;
658
659#ifdef INET6
660	af = AF_UNSPEC;
661#else
662	af = AF_INET;
663#endif
664
665	verbose = CHECK_FLAG('v');
666	if (CHECK_FLAG('4'))
667		af = AF_INET;
668#ifdef INET6
669	else if (CHECK_FLAG('6'))
670		af = AF_INET6;
671#endif
672
673	if (purl && strcasecmp(URL->scheme, SCHEME_HTTPS) != 0) {
674		URL = purl;
675	} else if (strcasecmp(URL->scheme, SCHEME_FTP) == 0) {
676		/* can't talk http to an ftp server */
677		/* XXX should set an error code */
678		return (NULL);
679	}
680
681	if ((conn = _fetch_connect(URL->host, URL->port, af, verbose)) == NULL)
682		/* _fetch_connect() has already set an error code */
683		return (NULL);
684	if (strcasecmp(URL->scheme, SCHEME_HTTPS) == 0 &&
685	    _fetch_ssl(conn, verbose) == -1) {
686		_fetch_close(conn);
687		/* grrr */
688		errno = EAUTH;
689		_fetch_syserr();
690		return (NULL);
691	}
692	return (conn);
693}
694
695static struct url *
696_http_get_proxy(void)
697{
698	struct url *purl;
699	char *p;
700
701	if (((p = getenv("HTTP_PROXY")) || (p = getenv("http_proxy"))) &&
702	    (purl = fetchParseURL(p))) {
703		if (!*purl->scheme)
704			strcpy(purl->scheme, SCHEME_HTTP);
705		if (!purl->port)
706			purl->port = _fetch_default_proxy_port(purl->scheme);
707		if (strcasecmp(purl->scheme, SCHEME_HTTP) == 0)
708			return (purl);
709		fetchFreeURL(purl);
710	}
711	return (NULL);
712}
713
714static void
715_http_print_html(FILE *out, FILE *in)
716{
717	size_t len;
718	char *line, *p, *q;
719	int comment, tag;
720
721	comment = tag = 0;
722	while ((line = fgetln(in, &len)) != NULL) {
723		while (len && isspace(line[len - 1]))
724			--len;
725		for (p = q = line; q < line + len; ++q) {
726			if (comment && *q == '-') {
727				if (q + 2 < line + len &&
728				    strcmp(q, "-->") == 0) {
729					tag = comment = 0;
730					q += 2;
731				}
732			} else if (tag && !comment && *q == '>') {
733				p = q + 1;
734				tag = 0;
735			} else if (!tag && *q == '<') {
736				if (q > p)
737					fwrite(p, q - p, 1, out);
738				tag = 1;
739				if (q + 3 < line + len &&
740				    strcmp(q, "<!--") == 0) {
741					comment = 1;
742					q += 3;
743				}
744			}
745		}
746		if (!tag && q > p)
747			fwrite(p, q - p, 1, out);
748		fputc('\n', out);
749	}
750}
751
752
753/*****************************************************************************
754 * Core
755 */
756
757/*
758 * Send a request and process the reply
759 *
760 * XXX This function is way too long, the do..while loop should be split
761 * XXX off into a separate function.
762 */
763FILE *
764_http_request(struct url *URL, const char *op, struct url_stat *us,
765    struct url *purl, const char *flags)
766{
767	conn_t *conn;
768	struct url *url, *new;
769	int chunked, direct, need_auth, noredirect, verbose;
770	int e, i, n;
771	off_t offset, clength, length, size;
772	time_t mtime;
773	const char *p;
774	FILE *f;
775	hdr_t h;
776	char *host;
777#ifdef INET6
778	char hbuf[MAXHOSTNAMELEN + 1];
779#endif
780
781	direct = CHECK_FLAG('d');
782	noredirect = CHECK_FLAG('A');
783	verbose = CHECK_FLAG('v');
784
785	if (direct && purl) {
786		fetchFreeURL(purl);
787		purl = NULL;
788	}
789
790	/* try the provided URL first */
791	url = URL;
792
793	/* if the A flag is set, we only get one try */
794	n = noredirect ? 1 : MAX_REDIRECT;
795	i = 0;
796
797	e = HTTP_PROTOCOL_ERROR;
798	need_auth = 0;
799	do {
800		new = NULL;
801		chunked = 0;
802		offset = 0;
803		clength = -1;
804		length = -1;
805		size = -1;
806		mtime = 0;
807
808		/* check port */
809		if (!url->port)
810			url->port = _fetch_default_port(url->scheme);
811
812		/* were we redirected to an FTP URL? */
813		if (purl == NULL && strcmp(url->scheme, SCHEME_FTP) == 0) {
814			if (strcmp(op, "GET") == 0)
815				return (_ftp_request(url, "RETR", us, purl, flags));
816			else if (strcmp(op, "HEAD") == 0)
817				return (_ftp_request(url, "STAT", us, purl, flags));
818		}
819
820		/* connect to server or proxy */
821		if ((conn = _http_connect(url, purl, flags)) == NULL)
822			goto ouch;
823
824		host = url->host;
825#ifdef INET6
826		if (strchr(url->host, ':')) {
827			snprintf(hbuf, sizeof(hbuf), "[%s]", url->host);
828			host = hbuf;
829		}
830#endif
831
832		/* send request */
833		if (verbose)
834			_fetch_info("requesting %s://%s:%d%s",
835			    url->scheme, host, url->port, url->doc);
836		if (purl) {
837			_http_cmd(conn, "%s %s://%s:%d%s HTTP/1.1",
838			    op, url->scheme, host, url->port, url->doc);
839		} else {
840			_http_cmd(conn, "%s %s HTTP/1.1",
841			    op, url->doc);
842		}
843
844		/* virtual host */
845		if (url->port == _fetch_default_port(url->scheme))
846			_http_cmd(conn, "Host: %s", host);
847		else
848			_http_cmd(conn, "Host: %s:%d", host, url->port);
849
850		/* proxy authorization */
851		if (purl) {
852			if (*purl->user || *purl->pwd)
853				_http_basic_auth(conn, "Proxy-Authorization",
854				    purl->user, purl->pwd);
855			else if ((p = getenv("HTTP_PROXY_AUTH")) != NULL && *p != '\0')
856				_http_authorize(conn, "Proxy-Authorization", p);
857		}
858
859		/* server authorization */
860		if (need_auth || *url->user || *url->pwd) {
861			if (*url->user || *url->pwd)
862				_http_basic_auth(conn, "Authorization", url->user, url->pwd);
863			else if ((p = getenv("HTTP_AUTH")) != NULL && *p != '\0')
864				_http_authorize(conn, "Authorization", p);
865			else if (fetchAuthMethod && fetchAuthMethod(url) == 0) {
866				_http_basic_auth(conn, "Authorization", url->user, url->pwd);
867			} else {
868				_http_seterr(HTTP_NEED_AUTH);
869				goto ouch;
870			}
871		}
872
873		/* other headers */
874		if ((p = getenv("HTTP_USER_AGENT")) != NULL && *p != '\0')
875			_http_cmd(conn, "User-Agent: %s", p);
876		else
877			_http_cmd(conn, "User-Agent: %s " _LIBFETCH_VER, getprogname());
878		if (url->offset)
879			_http_cmd(conn, "Range: bytes=%lld-", (long long)url->offset);
880		_http_cmd(conn, "Connection: close");
881		_http_cmd(conn, "");
882
883		/* get reply */
884		switch (_http_get_reply(conn)) {
885		case HTTP_OK:
886		case HTTP_PARTIAL:
887			/* fine */
888			break;
889		case HTTP_MOVED_PERM:
890		case HTTP_MOVED_TEMP:
891		case HTTP_SEE_OTHER:
892			/*
893			 * Not so fine, but we still have to read the headers to
894			 * get the new location.
895			 */
896			break;
897		case HTTP_NEED_AUTH:
898			if (need_auth) {
899				/*
900				 * We already sent out authorization code, so there's
901				 * nothing more we can do.
902				 */
903				_http_seterr(conn->err);
904				goto ouch;
905			}
906			/* try again, but send the password this time */
907			if (verbose)
908				_fetch_info("server requires authorization");
909			break;
910		case HTTP_NEED_PROXY_AUTH:
911			/*
912			 * If we're talking to a proxy, we already sent our proxy
913			 * authorization code, so there's nothing more we can do.
914			 */
915			_http_seterr(conn->err);
916			goto ouch;
917		case HTTP_PROTOCOL_ERROR:
918			/* fall through */
919		case -1:
920			_fetch_syserr();
921			goto ouch;
922		default:
923			_http_seterr(conn->err);
924			if (!verbose)
925				goto ouch;
926			/* fall through so we can get the full error message */
927		}
928
929		/* get headers */
930		do {
931			switch ((h = _http_next_header(conn, &p))) {
932			case hdr_syserror:
933				_fetch_syserr();
934				goto ouch;
935			case hdr_error:
936				_http_seterr(HTTP_PROTOCOL_ERROR);
937				goto ouch;
938			case hdr_content_length:
939				_http_parse_length(p, &clength);
940				break;
941			case hdr_content_range:
942				_http_parse_range(p, &offset, &length, &size);
943				break;
944			case hdr_last_modified:
945				_http_parse_mtime(p, &mtime);
946				break;
947			case hdr_location:
948				if (!HTTP_REDIRECT(conn->err))
949					break;
950				if (new)
951					free(new);
952				if (verbose)
953					_fetch_info("%d redirect to %s", conn->err, p);
954				if (*p == '/')
955					/* absolute path */
956					new = fetchMakeURL(url->scheme, url->host, url->port, p,
957					    url->user, url->pwd);
958				else
959					new = fetchParseURL(p);
960				if (new == NULL) {
961					/* XXX should set an error code */
962					DEBUG(fprintf(stderr, "failed to parse new URL\n"));
963					goto ouch;
964				}
965				if (!*new->user && !*new->pwd) {
966					strcpy(new->user, url->user);
967					strcpy(new->pwd, url->pwd);
968				}
969				new->offset = url->offset;
970				new->length = url->length;
971				break;
972			case hdr_transfer_encoding:
973				/* XXX weak test*/
974				chunked = (strcasecmp(p, "chunked") == 0);
975				break;
976			case hdr_www_authenticate:
977				if (conn->err != HTTP_NEED_AUTH)
978					break;
979				/* if we were smarter, we'd check the method and realm */
980				break;
981			case hdr_end:
982				/* fall through */
983			case hdr_unknown:
984				/* ignore */
985				break;
986			}
987		} while (h > hdr_end);
988
989		/* we need to provide authentication */
990		if (conn->err == HTTP_NEED_AUTH) {
991			e = conn->err;
992			need_auth = 1;
993			_fetch_close(conn);
994			conn = NULL;
995			continue;
996		}
997
998		/* we have a hit or an error */
999		if (conn->err == HTTP_OK || conn->err == HTTP_PARTIAL || HTTP_ERROR(conn->err))
1000			break;
1001
1002		/* all other cases: we got a redirect */
1003		e = conn->err;
1004		need_auth = 0;
1005		_fetch_close(conn);
1006		conn = NULL;
1007		if (!new) {
1008			DEBUG(fprintf(stderr, "redirect with no new location\n"));
1009			break;
1010		}
1011		if (url != URL)
1012			fetchFreeURL(url);
1013		url = new;
1014	} while (++i < n);
1015
1016	/* we failed, or ran out of retries */
1017	if (conn == NULL) {
1018		_http_seterr(e);
1019		goto ouch;
1020	}
1021
1022	DEBUG(fprintf(stderr, "offset %lld, length %lld,"
1023		  " size %lld, clength %lld\n",
1024		  (long long)offset, (long long)length,
1025		  (long long)size, (long long)clength));
1026
1027	/* check for inconsistencies */
1028	if (clength != -1 && length != -1 && clength != length) {
1029		_http_seterr(HTTP_PROTOCOL_ERROR);
1030		goto ouch;
1031	}
1032	if (clength == -1)
1033		clength = length;
1034	if (clength != -1)
1035		length = offset + clength;
1036	if (length != -1 && size != -1 && length != size) {
1037		_http_seterr(HTTP_PROTOCOL_ERROR);
1038		goto ouch;
1039	}
1040	if (size == -1)
1041		size = length;
1042
1043	/* fill in stats */
1044	if (us) {
1045		us->size = size;
1046		us->atime = us->mtime = mtime;
1047	}
1048
1049	/* too far? */
1050	if (offset > URL->offset) {
1051		_http_seterr(HTTP_PROTOCOL_ERROR);
1052		goto ouch;
1053	}
1054
1055	/* report back real offset and size */
1056	URL->offset = offset;
1057	URL->length = clength;
1058
1059	/* wrap it up in a FILE */
1060	if ((f = _http_funopen(conn, chunked)) == NULL) {
1061		_fetch_syserr();
1062		goto ouch;
1063	}
1064
1065	if (url != URL)
1066		fetchFreeURL(url);
1067	if (purl)
1068		fetchFreeURL(purl);
1069
1070	if (HTTP_ERROR(conn->err)) {
1071		_http_print_html(stderr, f);
1072		fclose(f);
1073		f = NULL;
1074	}
1075
1076	return (f);
1077
1078ouch:
1079	if (url != URL)
1080		fetchFreeURL(url);
1081	if (purl)
1082		fetchFreeURL(purl);
1083	if (conn != NULL)
1084		_fetch_close(conn);
1085	return (NULL);
1086}
1087
1088
1089/*****************************************************************************
1090 * Entry points
1091 */
1092
1093/*
1094 * Retrieve and stat a file by HTTP
1095 */
1096FILE *
1097fetchXGetHTTP(struct url *URL, struct url_stat *us, const char *flags)
1098{
1099	return (_http_request(URL, "GET", us, _http_get_proxy(), flags));
1100}
1101
1102/*
1103 * Retrieve a file by HTTP
1104 */
1105FILE *
1106fetchGetHTTP(struct url *URL, const char *flags)
1107{
1108	return (fetchXGetHTTP(URL, NULL, flags));
1109}
1110
1111/*
1112 * Store a file by HTTP
1113 */
1114FILE *
1115fetchPutHTTP(struct url *URL __unused, const char *flags __unused)
1116{
1117	warnx("fetchPutHTTP(): not implemented");
1118	return (NULL);
1119}
1120
1121/*
1122 * Get an HTTP document's metadata
1123 */
1124int
1125fetchStatHTTP(struct url *URL, struct url_stat *us, const char *flags)
1126{
1127	FILE *f;
1128
1129	if ((f = _http_request(URL, "HEAD", us, _http_get_proxy(), flags)) == NULL)
1130		return (-1);
1131	fclose(f);
1132	return (0);
1133}
1134
1135/*
1136 * List a directory
1137 */
1138struct url_ent *
1139fetchListHTTP(struct url *url __unused, const char *flags __unused)
1140{
1141	warnx("fetchListHTTP(): not implemented");
1142	return (NULL);
1143}
1144