1 /* vi: set sw=4 ts=4: */
3 * wget - retrieve a file using HTTP or FTP
5 * Chip Rosenthal Covad Communications <chip@laserlink.net>
9 /* We want libc to give us xxx64 functions also */
10 /* http://www.unix.org/version2/whatsnew/lfs20mar.html */
11 #define _LARGEFILE64_SOURCE 1
14 #include <getopt.h> /* for struct option */
17 // May be used if we ever will want to free() all xstrdup()s...
18 /* char *allocated; */
26 static void parse_url(char *url, struct host_info *h);
27 static FILE *open_socket(struct sockaddr_in *s_in);
28 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
29 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf);
31 /* Globals (can be accessed from signal handlers */
32 static off_t content_len; /* Content-length of the file */
33 static off_t beg_range; /* Range at which continue begins */
34 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
35 static off_t transferred; /* Number of bytes transferred so far */
37 static int chunked; /* chunked transfer encoding */
38 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
39 static void progressmeter(int flag);
40 static char *curfile; /* Name of current file being transferred */
41 static struct timeval start; /* Time a transfer started */
43 STALLTIME = 5 /* Seconds when xfer considered "stalled" */
46 static void progressmeter(int flag) {}
49 /* Read NMEMB elements of SIZE bytes into PTR from STREAM. Returns the
50 * number of elements read, and a short count if an eof or non-interrupt
51 * error is encountered. */
52 static size_t safe_fread(void *ptr, size_t size, size_t nmemb, FILE *stream)
58 ret += fread((char *)ptr + (ret * size), size, nmemb - ret, stream);
59 } while (ret < nmemb && ferror(stream) && errno == EINTR);
64 /* Read a line or SIZE - 1 bytes into S, whichever is less, from STREAM.
65 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
66 static char *safe_fgets(char *s, int size, FILE *stream)
72 ret = fgets(s, size, stream);
73 } while (ret == NULL && ferror(stream) && errno == EINTR);
78 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
80 * Base64-encode character string and return the string.
82 static char *base64enc(unsigned char *p, char *buf, int len)
84 bb_uuencode(p, buf, len, bb_uuenc_tbl_base64);
89 int wget_main(int argc, char **argv)
92 struct host_info server, target;
93 struct sockaddr_in s_in;
100 char *dir_prefix = NULL;
101 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
102 char *extra_headers = NULL;
103 llist_t *headers_llist = NULL;
106 /* server.allocated = target.allocated = NULL; */
108 FILE *sfp = NULL; /* socket to web/ftp server */
109 FILE *dfp = NULL; /* socket to ftp server (data) */
110 char *fname_out = NULL; /* where to direct output (-O) */
111 int got_clen = 0; /* got content-length: from server */
113 int use_proxy = 1; /* Use proxies if env vars are set */
114 const char *proxy_flag = "on"; /* Use proxies if env vars are set */
115 const char *user_agent = "Wget";/* Content of the "User-Agent" header field */
118 * Crack command line.
121 WGET_OPT_CONTINUE = 0x1,
122 WGET_OPT_QUIET = 0x2,
123 WGET_OPT_OUTNAME = 0x4,
124 WGET_OPT_PREFIX = 0x8,
125 WGET_OPT_PROXY = 0x10,
126 WGET_OPT_USER_AGENT = 0x20,
127 WGET_OPT_PASSIVE = 0x40,
128 WGET_OPT_HEADER = 0x80,
130 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
131 static const struct option wget_long_options[] = {
132 // name, has_arg, flag, val
133 { "continue", no_argument, NULL, 'c' },
134 { "quiet", no_argument, NULL, 'q' },
135 { "output-document", required_argument, NULL, 'O' },
136 { "directory-prefix", required_argument, NULL, 'P' },
137 { "proxy", required_argument, NULL, 'Y' },
138 { "user-agent", required_argument, NULL, 'U' },
139 { "passive-ftp", no_argument, NULL, 0xff },
140 { "header", required_argument, NULL, 0xfe },
143 applet_long_options = wget_long_options;
145 opt_complementary = "-1" USE_FEATURE_WGET_LONG_OPTIONS(":\xfe::");
146 opt = getopt32(argc, argv, "cqO:P:Y:U:",
147 &fname_out, &dir_prefix,
148 &proxy_flag, &user_agent
149 USE_FEATURE_WGET_LONG_OPTIONS(, &headers_llist)
151 if (strcmp(proxy_flag, "off") == 0) {
152 /* Use the proxy if necessary. */
155 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
159 llist_t *ll = headers_llist = rev_llist(headers_llist);
161 size += strlen(ll->data) + 2;
164 extra_headers = cp = xmalloc(size);
165 while (headers_llist) {
166 cp += sprintf(cp, "%s\r\n", headers_llist->data);
167 headers_llist = headers_llist->link;
172 parse_url(argv[optind], &target);
173 server.host = target.host;
174 server.port = target.port;
177 * Use the proxy if necessary.
180 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
181 if (proxy && *proxy) {
182 parse_url(proxy, &server);
188 /* Guess an output filename */
190 // Dirty hack. Needed because bb_get_last_path_component
191 // will destroy trailing / by storing '\0' in last byte!
192 if (*target.path && target.path[strlen(target.path)-1] != '/') {
194 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
197 bb_get_last_path_component(target.path);
199 if (!fname_out || !fname_out[0]) {
201 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
206 if (dir_prefix != NULL)
207 fname_out = concat_path_file(dir_prefix, fname_out);
208 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
210 curfile = bb_get_last_path_component(fname_out);
213 if ((opt & WGET_OPT_CONTINUE) && !fname_out)
214 bb_error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
217 * Determine where to start transfer.
219 if (!strcmp(fname_out, "-")) {
221 opt |= WGET_OPT_QUIET;
222 opt &= ~WGET_OPT_CONTINUE;
223 } else if (opt & WGET_OPT_CONTINUE) {
224 output_fd = open(fname_out, O_WRONLY);
225 if (output_fd >= 0) {
226 beg_range = xlseek(output_fd, 0, SEEK_END);
228 /* File doesn't exist. We do not create file here yet.
229 We are not sure it exists on remove side */
232 /* We want to do exactly _one_ DNS lookup, since some
233 * sites (i.e. ftp.us.debian.org) use round-robin DNS
234 * and we want to connect to only one IP... */
235 bb_lookup_host(&s_in, server.host);
236 s_in.sin_port = server.port;
237 if (!(opt & WGET_OPT_QUIET)) {
238 printf("Connecting to %s[%s]:%d\n",
239 server.host, inet_ntoa(s_in.sin_addr), ntohs(server.port));
242 if (use_proxy || !target.is_ftp) {
247 got_clen = chunked = 0;
250 bb_error_msg_and_die("too many redirections");
253 * Open socket to http server
255 if (sfp) fclose(sfp);
256 sfp = open_socket(&s_in);
262 const char *format = "GET %stp://%s:%d/%s HTTP/1.1\r\n";
263 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
264 if (strchr(target.host, ':'))
265 format = "GET %stp://[%s]:%d/%s HTTP/1.1\r\n";
268 target.is_ftp ? "f" : "ht", target.host,
269 ntohs(target.port), target.path);
271 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
274 fprintf(sfp, "Host: %s\r\nUser-Agent: %s\r\n", target.host,
277 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
279 fprintf(sfp, "Authorization: Basic %s\r\n",
280 base64enc((unsigned char*)target.user, buf, sizeof(buf)));
282 if (use_proxy && server.user) {
283 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
284 base64enc((unsigned char*)server.user, buf, sizeof(buf)));
289 fprintf(sfp, "Range: bytes="OFF_FMT"-\r\n", beg_range);
290 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
292 fputs(extra_headers, sfp);
294 fprintf(sfp, "Connection: close\r\n\r\n");
297 * Retrieve HTTP response line and check for "200" status code.
300 if (fgets(buf, sizeof(buf), sfp) == NULL)
301 bb_error_msg_and_die("no response from server");
304 while (*s != '\0' && !isspace(*s)) ++s;
305 s = skip_whitespace(s);
306 // FIXME: no error check
307 // xatou wouldn't work: "200 OK"
312 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL)
313 /* eat all remaining headers */;
317 case 300: /* redirection */
328 bb_error_msg_and_die("server returned error %s: %s", s, buf);
332 * Retrieve HTTP headers.
334 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
335 if (strcasecmp(buf, "content-length") == 0) {
336 if (SAFE_STRTOOFF(s, &content_len) || content_len < 0) {
337 bb_error_msg_and_die("content-length %s is garbage", s);
342 if (strcasecmp(buf, "transfer-encoding") == 0) {
343 if (strcasecmp(s, "chunked") != 0)
344 bb_error_msg_and_die("server wants to do %s transfer encoding", s);
345 chunked = got_clen = 1;
347 if (strcasecmp(buf, "location") == 0) {
349 /* free(target.allocated); */
350 target.path = /* target.allocated = */ xstrdup(s+1);
352 parse_url(s, &target);
353 if (use_proxy == 0) {
354 server.host = target.host;
355 server.port = target.port;
357 bb_lookup_host(&s_in, server.host);
358 s_in.sin_port = server.port;
363 } while(status >= 300);
373 target.user = xstrdup("anonymous:busybox@");
375 sfp = open_socket(&s_in);
376 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
377 bb_error_msg_and_die("%s", buf+4);
380 * Splitting username:password pair,
383 s = strchr(target.user, ':');
386 switch (ftpcmd("USER ", target.user, sfp, buf)) {
390 if (ftpcmd("PASS ", s, sfp, buf) == 230)
392 /* FALLTHRU (failed login) */
394 bb_error_msg_and_die("ftp login: %s", buf+4);
397 ftpcmd("TYPE I", NULL, sfp, buf);
402 if (ftpcmd("SIZE ", target.path, sfp, buf) == 213) {
403 if (SAFE_STRTOOFF(buf+4, &content_len) || content_len < 0) {
404 bb_error_msg_and_die("SIZE value is garbage");
410 * Entering passive mode
412 if (ftpcmd("PASV", NULL, sfp, buf) != 227) {
414 bb_error_msg_and_die("bad response to %s: %s", "PASV", buf);
416 // Response is "227 garbageN1,N2,N3,N4,P1,P2
417 // Server's IP is N1.N2.N3.N4 (we ignore it)
418 // Server's port for data connection is P1*256+P2
419 s = strrchr(buf, ',');
420 if (!s) goto pasv_error;
421 port = xatol_range(s+1, 0, 255);
423 s = strrchr(buf, ',');
424 if (!s) goto pasv_error;
425 port += xatol_range(s+1, 0, 255) * 256;
426 s_in.sin_port = htons(port);
427 dfp = open_socket(&s_in);
430 sprintf(buf, "REST "OFF_FMT, beg_range);
431 if (ftpcmd(buf, NULL, sfp, buf) == 350)
432 content_len -= beg_range;
435 if (ftpcmd("RETR ", target.path, sfp, buf) > 150)
436 bb_error_msg_and_die("bad response to %s: %s", "RETR", buf);
444 fgets(buf, sizeof(buf), dfp);
445 content_len = STRTOOFF(buf, (char **) NULL, 16);
446 /* FIXME: error check?? */
449 /* Do it before progressmeter (want to have nice error message) */
451 output_fd = xopen3(fname_out,
452 O_WRONLY|O_CREAT|O_EXCL|O_TRUNC, 0666);
454 if (!(opt & WGET_OPT_QUIET))
458 while (content_len > 0 || !got_clen) {
459 unsigned rdsz = sizeof(buf);
460 if (content_len < sizeof(buf) && (chunked || got_clen))
461 rdsz = (unsigned)content_len;
462 n = safe_fread(buf, 1, rdsz, dfp);
465 if (full_write(output_fd, buf, n) != n) {
466 bb_perror_msg_and_die(bb_msg_write_error);
468 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
477 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
478 safe_fgets(buf, sizeof(buf), dfp);
479 content_len = STRTOOFF(buf, (char **) NULL, 16);
480 /* FIXME: error check? */
481 if (content_len == 0) {
482 chunked = 0; /* all done! */
486 if (n == 0 && ferror(dfp)) {
487 bb_perror_msg_and_die(bb_msg_read_error);
491 if (!(opt & WGET_OPT_QUIET))
494 if ((use_proxy == 0) && target.is_ftp) {
496 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
497 bb_error_msg_and_die("ftp error: %s", buf+4);
498 ftpcmd("QUIT", NULL, sfp, buf);
504 static void parse_url(char *src_url, struct host_info *h)
506 char *url, *p, *cp, *sp, *up, *pp;
508 /* h->allocated = */ url = xstrdup(src_url);
510 if (strncmp(url, "http://", 7) == 0) {
511 h->port = bb_lookup_port("http", "tcp", 80);
514 } else if (strncmp(url, "ftp://", 6) == 0) {
515 h->port = bb_lookup_port("ftp", "tcp", 21);
519 bb_error_msg_and_die("not an http or ftp url: %s", url);
522 // "Real" wget 'http://busybox.net?var=a/b' sends this request:
523 // 'GET /?var=a/b HTTP 1.0'
524 // and saves 'index.html?var=a%2Fb' (we save 'b')
525 // wget 'http://busybox.net?login=john@doe':
526 // request: 'GET /?login=john@doe HTTP/1.0'
527 // saves: 'index.html?login=john@doe' (we save '?login=john@doe')
528 // wget 'http://busybox.net#test/test':
529 // request: 'GET / HTTP/1.0'
530 // saves: 'index.html' (we save 'test')
532 // We also don't add unique .N suffix if file exists...
533 sp = strchr(h->host, '/');
534 p = strchr(h->host, '?'); if (!sp || (p && sp > p)) sp = p;
535 p = strchr(h->host, '#'); if (!sp || (p && sp > p)) sp = p;
538 } else if (*sp == '/') {
541 } else { // '#' or '?'
542 // http://busybox.net?login=john@doe is a valid URL
543 // memmove converts to:
544 // http:/busybox.nett?login=john@doe...
545 memmove(h->host-1, h->host, sp - h->host);
551 up = strrchr(h->host, '@');
561 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
562 if (h->host[0] == '[') {
566 while (*ep == ':' || isxdigit(*ep))
576 cp = strchr(pp, ':');
579 h->port = htons(xatou16(cp));
584 static FILE *open_socket(struct sockaddr_in *s_in)
588 fp = fdopen(xconnect(s_in), "r+");
590 bb_perror_msg_and_die("fdopen");
596 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
603 /* retrieve header line */
604 if (fgets(buf, bufsiz, fp) == NULL)
607 /* see if we are at the end of the headers */
608 for (s = buf ; *s == '\r' ; ++s)
613 /* convert the header name to lower case */
614 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
617 /* verify we are at the end of the header name */
619 bb_error_msg_and_die("bad header line: %s", buf);
621 /* locate the start of the header value */
622 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
626 /* locate the end of header */
627 while (*s != '\0' && *s != '\r' && *s != '\n')
630 /* end of header found */
636 /* Rats! The buffer isn't big enough to hold the entire header value. */
637 while (c = getc(fp), c != EOF && c != '\n')
643 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
648 fprintf(fp, "%s%s\r\n", s1, s2);
655 if (fgets(buf, 510, fp) == NULL) {
656 bb_perror_msg_and_die("error getting response");
658 buf_ptr = strstr(buf, "\r\n");
662 } while (!isdigit(buf[0]) || buf[3] != ' ');
665 result = xatoi_u(buf);
670 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
671 /* Stuff below is from BSD rcp util.c, as added to openshh.
672 * Original copyright notice is retained at the end of this file.
678 get_terminal_width_height(0, &width, NULL);
683 updateprogressmeter(int ignore)
685 int save_errno = errno;
691 static void alarmtimer(int iwait)
693 struct itimerval itv;
695 itv.it_value.tv_sec = iwait;
696 itv.it_value.tv_usec = 0;
697 itv.it_interval = itv.it_value;
698 setitimer(ITIMER_REAL, &itv, NULL);
703 progressmeter(int flag)
705 static struct timeval lastupdate;
706 static off_t lastsize, totalsize;
708 struct timeval now, td, tvwait;
710 int elapsed, ratio, barlength, i;
713 if (flag == -1) { /* first call to progressmeter */
714 (void) gettimeofday(&start, (struct timezone *) 0);
717 totalsize = content_len + beg_range; /* as content_len changes.. */
720 (void) gettimeofday(&now, (struct timezone *) 0);
722 if (totalsize != 0 && !chunked) {
723 ratio = (int) (100 * (transferred+beg_range) / totalsize);
724 ratio = MIN(ratio, 100);
727 fprintf(stderr, "\r%-20.20s%4d%% ", curfile, ratio);
729 barlength = getttywidth() - 51;
730 if (barlength > 0 && barlength < sizeof(buf)) {
731 i = barlength * ratio / 100;
733 memset(buf + i, ' ', barlength - i);
734 buf[barlength] = '\0';
735 fprintf(stderr, "|%s|", buf);
738 abbrevsize = transferred + beg_range;
739 while (abbrevsize >= 100000) {
743 /* See http://en.wikipedia.org/wiki/Tera */
744 fprintf(stderr, "%6d %c%c ", (int)abbrevsize, " KMGTPEZY"[i], i?'B':' ');
746 timersub(&now, &lastupdate, &tvwait);
747 if (transferred > lastsize) {
749 lastsize = transferred;
750 if (tvwait.tv_sec >= STALLTIME)
751 timeradd(&start, &tvwait, &start);
754 timersub(&now, &start, &td);
757 if (tvwait.tv_sec >= STALLTIME) {
758 fprintf(stderr, " - stalled -");
760 off_t to_download = totalsize - beg_range;
761 if (transferred <= 0 || elapsed <= 0 || transferred > to_download || chunked) {
762 fprintf(stderr, "--:--:-- ETA");
764 /* to_download / (transferred/elapsed) - elapsed: */
765 int eta = (int) (to_download*elapsed/transferred - elapsed);
767 fprintf(stderr, "%02d:%02d:%02d ETA", eta / 3600, i / 60, i % 60);
771 if (flag == -1) { /* first call to progressmeter */
773 sa.sa_handler = updateprogressmeter;
774 sigemptyset(&sa.sa_mask);
775 sa.sa_flags = SA_RESTART;
776 sigaction(SIGALRM, &sa, NULL);
778 } else if (flag == 1) { /* last call to progressmeter */
786 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
787 * much of which was blatantly stolen from openssh. */
790 * Copyright (c) 1992, 1993
791 * The Regents of the University of California. All rights reserved.
793 * Redistribution and use in source and binary forms, with or without
794 * modification, are permitted provided that the following conditions
796 * 1. Redistributions of source code must retain the above copyright
797 * notice, this list of conditions and the following disclaimer.
798 * 2. Redistributions in binary form must reproduce the above copyright
799 * notice, this list of conditions and the following disclaimer in the
800 * documentation and/or other materials provided with the distribution.
802 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
803 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
805 * 4. Neither the name of the University nor the names of its contributors
806 * may be used to endorse or promote products derived from this software
807 * without specific prior written permission.
809 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
810 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
811 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
812 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
813 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
814 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
815 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
816 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
817 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
818 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
821 * $Id: wget.c,v 1.75 2004/10/08 08:27:40 andersen Exp $