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(len_and_sockaddr *lsa);
28 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
29 static int ftpcmd(const char *s1, const 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 #if ENABLE_FEATURE_WGET_STATUSBAR
35 static off_t transferred; /* Number of bytes transferred so far */
37 static int chunked; /* chunked transfer encoding */
38 #if ENABLE_FEATURE_WGET_STATUSBAR
39 static void progressmeter(int flag);
40 static const 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 #if ENABLE_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);
90 int wget_main(int argc, char **argv)
93 struct host_info server, target;
94 len_and_sockaddr *lsa;
101 char *dir_prefix = NULL;
102 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
103 char *extra_headers = NULL;
104 llist_t *headers_llist = NULL;
107 /* server.allocated = target.allocated = NULL; */
109 FILE *sfp = NULL; /* socket to web/ftp server */
110 FILE *dfp = NULL; /* socket to ftp server (data) */
111 char *fname_out = NULL; /* where to direct output (-O) */
112 int got_clen = 0; /* got content-length: from server */
114 int use_proxy = 1; /* Use proxies if env vars are set */
115 const char *proxy_flag = "on"; /* Use proxies if env vars are set */
116 const char *user_agent = "Wget";/* Content of the "User-Agent" header field */
119 * Crack command line.
122 WGET_OPT_CONTINUE = 0x1,
123 WGET_OPT_SPIDER = 0x2,
124 WGET_OPT_QUIET = 0x4,
125 WGET_OPT_OUTNAME = 0x8,
126 WGET_OPT_PREFIX = 0x10,
127 WGET_OPT_PROXY = 0x20,
128 WGET_OPT_USER_AGENT = 0x40,
129 WGET_OPT_PASSIVE = 0x80,
130 WGET_OPT_HEADER = 0x100,
132 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
133 static const struct option wget_long_options[] = {
134 // name, has_arg, flag, val
135 { "continue", no_argument, NULL, 'c' },
136 { "spider", no_argument, NULL, 's' },
137 { "quiet", no_argument, NULL, 'q' },
138 { "output-document", required_argument, NULL, 'O' },
139 { "directory-prefix", required_argument, NULL, 'P' },
140 { "proxy", required_argument, NULL, 'Y' },
141 { "user-agent", required_argument, NULL, 'U' },
142 { "passive-ftp", no_argument, NULL, 0xff },
143 { "header", required_argument, NULL, 0xfe },
146 applet_long_options = wget_long_options;
148 opt_complementary = "-1" USE_FEATURE_WGET_LONG_OPTIONS(":\xfe::");
149 opt = getopt32(argc, argv, "csqO:P:Y:U:",
150 &fname_out, &dir_prefix,
151 &proxy_flag, &user_agent
152 USE_FEATURE_WGET_LONG_OPTIONS(, &headers_llist)
154 if (strcmp(proxy_flag, "off") == 0) {
155 /* Use the proxy if necessary. */
158 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
162 llist_t *ll = headers_llist = llist_rev(headers_llist);
164 size += strlen(ll->data) + 2;
167 extra_headers = cp = xmalloc(size);
168 while (headers_llist) {
169 cp += sprintf(cp, "%s\r\n", headers_llist->data);
170 headers_llist = headers_llist->link;
175 parse_url(argv[optind], &target);
176 server.host = target.host;
177 server.port = target.port;
180 * Use the proxy if necessary.
183 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
184 if (proxy && *proxy) {
185 parse_url(proxy, &server);
191 /* Guess an output filename */
193 // Dirty hack. Needed because bb_get_last_path_component
194 // will destroy trailing / by storing '\0' in last byte!
195 if (!last_char_is(target.path, '/')) {
196 fname_out = bb_get_last_path_component(target.path);
197 #if ENABLE_FEATURE_WGET_STATUSBAR
201 if (!fname_out || !fname_out[0]) {
202 /* bb_get_last_path_component writes
203 * to last '/' only. We don't have one here... */
204 fname_out = (char*)"index.html";
205 #if ENABLE_FEATURE_WGET_STATUSBAR
209 if (dir_prefix != NULL)
210 fname_out = concat_path_file(dir_prefix, fname_out);
211 #if ENABLE_FEATURE_WGET_STATUSBAR
213 curfile = bb_get_last_path_component(fname_out);
217 if ((opt & WGET_OPT_CONTINUE) && !fname_out)
218 bb_error_msg_and_die("cannot specify continue (-c) without a filename (-O)"); */
221 * Determine where to start transfer.
223 if (LONE_DASH(fname_out)) {
225 opt &= ~WGET_OPT_CONTINUE;
227 if (opt & WGET_OPT_CONTINUE) {
228 output_fd = open(fname_out, O_WRONLY);
229 if (output_fd >= 0) {
230 beg_range = xlseek(output_fd, 0, SEEK_END);
232 /* File doesn't exist. We do not create file here yet.
233 We are not sure it exists on remove side */
236 /* We want to do exactly _one_ DNS lookup, since some
237 * sites (i.e. ftp.us.debian.org) use round-robin DNS
238 * and we want to connect to only one IP... */
239 lsa = xhost2sockaddr(server.host, server.port);
240 if (!(opt & WGET_OPT_QUIET)) {
241 fprintf(stderr, "Connecting to %s (%s)\n", server.host,
242 xmalloc_sockaddr2dotted(&lsa->sa, lsa->len));
243 /* We leak result of xmalloc_sockaddr2dotted */
246 if (use_proxy || !target.is_ftp) {
251 got_clen = chunked = 0;
254 bb_error_msg_and_die("too many redirections");
257 * Open socket to http server
259 if (sfp) fclose(sfp);
260 sfp = open_socket(lsa);
266 fprintf(sfp, "GET %stp://%s/%s HTTP/1.1\r\n",
267 target.is_ftp ? "f" : "ht", target.host,
270 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
273 fprintf(sfp, "Host: %s\r\nUser-Agent: %s\r\n",
274 target.host, user_agent);
276 #if ENABLE_FEATURE_WGET_AUTHENTICATION
278 fprintf(sfp, "Authorization: Basic %s\r\n",
279 base64enc((unsigned char*)target.user, buf, sizeof(buf)));
281 if (use_proxy && server.user) {
282 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
283 base64enc((unsigned char*)server.user, buf, sizeof(buf)));
288 fprintf(sfp, "Range: bytes=%"OFF_FMT"d-\r\n", beg_range);
289 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
291 fputs(extra_headers, sfp);
293 fprintf(sfp, "Connection: close\r\n\r\n");
296 * Retrieve HTTP response line and check for "200" status code.
299 if (fgets(buf, sizeof(buf), sfp) == NULL)
300 bb_error_msg_and_die("no response from server");
303 while (*s != '\0' && !isspace(*s)) ++s;
304 s = skip_whitespace(s);
305 // FIXME: no error check
306 // xatou wouldn't work: "200 OK"
311 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL)
312 /* eat all remaining headers */;
316 case 300: /* redirection */
326 /* Show first line only and kill any ESC tricks */
327 buf[strcspn(buf, "\n\r\x1b")] = '\0';
328 bb_error_msg_and_die("server returned error: %s", buf);
332 * Retrieve HTTP headers.
334 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
335 if (strcasecmp(buf, "content-length") == 0) {
336 content_len = BB_STRTOOFF(s, NULL, 10);
337 if (errno || content_len < 0) {
338 bb_error_msg_and_die("content-length %s is garbage", s);
343 if (strcasecmp(buf, "transfer-encoding") == 0) {
344 if (strcasecmp(s, "chunked") != 0)
345 bb_error_msg_and_die("server wants to do %s transfer encoding", s);
346 chunked = got_clen = 1;
348 if (strcasecmp(buf, "location") == 0) {
350 /* free(target.allocated); */
351 target.path = /* target.allocated = */ xstrdup(s+1);
353 parse_url(s, &target);
354 if (use_proxy == 0) {
355 server.host = target.host;
356 server.port = target.port;
359 lsa = xhost2sockaddr(server.host, server.port);
364 } while (status >= 300);
374 target.user = xstrdup("anonymous:busybox@");
376 sfp = open_socket(lsa);
377 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
378 bb_error_msg_and_die("%s", buf+4);
381 * Splitting username:password pair,
384 s = strchr(target.user, ':');
387 switch (ftpcmd("USER ", target.user, sfp, buf)) {
391 if (ftpcmd("PASS ", s, sfp, buf) == 230)
393 /* FALLTHRU (failed login) */
395 bb_error_msg_and_die("ftp login: %s", buf+4);
398 ftpcmd("TYPE I", NULL, sfp, buf);
403 if (ftpcmd("SIZE ", target.path, sfp, buf) == 213) {
404 content_len = BB_STRTOOFF(buf+4, NULL, 10);
405 if (errno || content_len < 0) {
406 bb_error_msg_and_die("SIZE value is garbage");
412 * Entering passive mode
414 if (ftpcmd("PASV", NULL, sfp, buf) != 227) {
416 bb_error_msg_and_die("bad response to %s: %s", "PASV", buf);
418 // Response is "227 garbageN1,N2,N3,N4,P1,P2[)garbage]
419 // Server's IP is N1.N2.N3.N4 (we ignore it)
420 // Server's port for data connection is P1*256+P2
421 s = strrchr(buf, ')');
423 s = strrchr(buf, ',');
424 if (!s) goto pasv_error;
425 port = xatou_range(s+1, 0, 255);
427 s = strrchr(buf, ',');
428 if (!s) goto pasv_error;
429 port += xatou_range(s+1, 0, 255) * 256;
430 set_nport(lsa, htons(port));
431 dfp = open_socket(lsa);
434 sprintf(buf, "REST %"OFF_FMT"d", beg_range);
435 if (ftpcmd(buf, NULL, sfp, buf) == 350)
436 content_len -= beg_range;
439 if (ftpcmd("RETR ", target.path, sfp, buf) > 150)
440 bb_error_msg_and_die("bad response to RETR: %s", buf);
442 if (opt & WGET_OPT_SPIDER) {
443 if (ENABLE_FEATURE_CLEAN_UP)
452 fgets(buf, sizeof(buf), dfp);
453 content_len = STRTOOFF(buf, NULL, 16);
454 /* FIXME: error check?? */
457 /* Do it before progressmeter (want to have nice error message) */
459 output_fd = xopen(fname_out,
460 O_WRONLY|O_CREAT|O_EXCL|O_TRUNC);
462 if (!(opt & WGET_OPT_QUIET))
466 while (content_len > 0 || !got_clen) {
467 unsigned rdsz = sizeof(buf);
468 if (content_len < sizeof(buf) && (chunked || got_clen))
469 rdsz = (unsigned)content_len;
470 n = safe_fread(buf, 1, rdsz, dfp);
473 if (full_write(output_fd, buf, n) != n) {
474 bb_perror_msg_and_die(bb_msg_write_error);
476 #if ENABLE_FEATURE_WGET_STATUSBAR
485 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
486 safe_fgets(buf, sizeof(buf), dfp);
487 content_len = STRTOOFF(buf, NULL, 16);
488 /* FIXME: error check? */
489 if (content_len == 0) {
490 chunked = 0; /* all done! */
494 if (n == 0 && ferror(dfp)) {
495 bb_perror_msg_and_die(bb_msg_read_error);
499 if (!(opt & WGET_OPT_QUIET))
502 if ((use_proxy == 0) && target.is_ftp) {
504 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
505 bb_error_msg_and_die("ftp error: %s", buf+4);
506 ftpcmd("QUIT", NULL, sfp, buf);
513 static void parse_url(char *src_url, struct host_info *h)
517 /* h->allocated = */ url = xstrdup(src_url);
519 if (strncmp(url, "http://", 7) == 0) {
520 h->port = bb_lookup_port("http", "tcp", 80);
523 } else if (strncmp(url, "ftp://", 6) == 0) {
524 h->port = bb_lookup_port("ftp", "tcp", 21);
528 bb_error_msg_and_die("not an http or ftp url: %s", url);
531 // "Real" wget 'http://busybox.net?var=a/b' sends this request:
532 // 'GET /?var=a/b HTTP 1.0'
533 // and saves 'index.html?var=a%2Fb' (we save 'b')
534 // wget 'http://busybox.net?login=john@doe':
535 // request: 'GET /?login=john@doe HTTP/1.0'
536 // saves: 'index.html?login=john@doe' (we save '?login=john@doe')
537 // wget 'http://busybox.net#test/test':
538 // request: 'GET / HTTP/1.0'
539 // saves: 'index.html' (we save 'test')
541 // We also don't add unique .N suffix if file exists...
542 sp = strchr(h->host, '/');
543 p = strchr(h->host, '?'); if (!sp || (p && sp > p)) sp = p;
544 p = strchr(h->host, '#'); if (!sp || (p && sp > p)) sp = p;
546 /* must be writable because of bb_get_last_path_component() */
547 static char nullstr[] = "";
549 } else if (*sp == '/') {
552 } else { // '#' or '?'
553 // http://busybox.net?login=john@doe is a valid URL
554 // memmove converts to:
555 // http:/busybox.nett?login=john@doe...
556 memmove(h->host-1, h->host, sp - h->host);
562 sp = strrchr(h->host, '@');
574 static FILE *open_socket(len_and_sockaddr *lsa)
578 /* glibc 2.4 seems to try seeking on it - ??! */
579 /* hopefully it understands what ESPIPE means... */
580 fp = fdopen(xconnect_stream(lsa), "r+");
582 bb_perror_msg_and_die("fdopen");
588 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
595 /* retrieve header line */
596 if (fgets(buf, bufsiz, fp) == NULL)
599 /* see if we are at the end of the headers */
600 for (s = buf; *s == '\r'; ++s)
605 /* convert the header name to lower case */
606 for (s = buf; isalnum(*s) || *s == '-'; ++s)
609 /* verify we are at the end of the header name */
611 bb_error_msg_and_die("bad header line: %s", buf);
613 /* locate the start of the header value */
614 for (*s++ = '\0'; *s == ' ' || *s == '\t'; ++s)
618 /* locate the end of header */
619 while (*s != '\0' && *s != '\r' && *s != '\n')
622 /* end of header found */
628 /* Rats! The buffer isn't big enough to hold the entire header value. */
629 while (c = getc(fp), c != EOF && c != '\n')
635 static int ftpcmd(const char *s1, const char *s2, FILE *fp, char *buf)
640 fprintf(fp, "%s%s\r\n", s1, s2);
647 if (fgets(buf, 510, fp) == NULL) {
648 bb_perror_msg_and_die("error getting response");
650 buf_ptr = strstr(buf, "\r\n");
654 } while (!isdigit(buf[0]) || buf[3] != ' ');
657 result = xatoi_u(buf);
662 #if ENABLE_FEATURE_WGET_STATUSBAR
663 /* Stuff below is from BSD rcp util.c, as added to openshh.
664 * Original copyright notice is retained at the end of this file.
670 get_terminal_width_height(0, &width, NULL);
675 updateprogressmeter(int ignore)
677 int save_errno = errno;
683 static void alarmtimer(int iwait)
685 struct itimerval itv;
687 itv.it_value.tv_sec = iwait;
688 itv.it_value.tv_usec = 0;
689 itv.it_interval = itv.it_value;
690 setitimer(ITIMER_REAL, &itv, NULL);
695 progressmeter(int flag)
697 static struct timeval lastupdate;
698 static off_t lastsize, totalsize;
700 struct timeval now, td, tvwait;
702 int elapsed, ratio, barlength, i;
705 if (flag == -1) { /* first call to progressmeter */
706 gettimeofday(&start, (struct timezone *) 0);
709 totalsize = content_len + beg_range; /* as content_len changes.. */
712 gettimeofday(&now, (struct timezone *) 0);
714 if (totalsize != 0 && !chunked) {
715 /* long long helps to have working ETA even if !LFS */
716 ratio = (int) (100 * (unsigned long long)(transferred+beg_range) / totalsize);
717 ratio = MIN(ratio, 100);
720 fprintf(stderr, "\r%-20.20s%4d%% ", curfile, ratio);
722 barlength = getttywidth() - 51;
723 if (barlength > 0 && barlength < sizeof(buf)) {
724 i = barlength * ratio / 100;
726 memset(buf + i, ' ', barlength - i);
727 buf[barlength] = '\0';
728 fprintf(stderr, "|%s|", buf);
731 abbrevsize = transferred + beg_range;
732 while (abbrevsize >= 100000) {
736 /* see http://en.wikipedia.org/wiki/Tera */
737 fprintf(stderr, "%6d %c%c ", (int)abbrevsize, " KMGTPEZY"[i], i?'B':' ');
739 timersub(&now, &lastupdate, &tvwait);
740 if (transferred > lastsize) {
742 lastsize = transferred;
743 if (tvwait.tv_sec >= STALLTIME)
744 timeradd(&start, &tvwait, &start);
747 timersub(&now, &start, &td);
750 if (tvwait.tv_sec >= STALLTIME) {
751 fprintf(stderr, " - stalled -");
753 off_t to_download = totalsize - beg_range;
754 if (transferred <= 0 || elapsed <= 0 || transferred > to_download || chunked) {
755 fprintf(stderr, "--:--:-- ETA");
757 /* to_download / (transferred/elapsed) - elapsed: */
758 int eta = (int) ((unsigned long long)to_download*elapsed/transferred - elapsed);
759 /* (long long helps to have working ETA even if !LFS) */
761 fprintf(stderr, "%02d:%02d:%02d ETA", eta / 3600, i / 60, i % 60);
765 if (flag == -1) { /* first call to progressmeter */
767 sa.sa_handler = updateprogressmeter;
768 sigemptyset(&sa.sa_mask);
769 sa.sa_flags = SA_RESTART;
770 sigaction(SIGALRM, &sa, NULL);
772 } else if (flag == 1) { /* last call to progressmeter */
780 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
781 * much of which was blatantly stolen from openssh. */
784 * Copyright (c) 1992, 1993
785 * The Regents of the University of California. All rights reserved.
787 * Redistribution and use in source and binary forms, with or without
788 * modification, are permitted provided that the following conditions
790 * 1. Redistributions of source code must retain the above copyright
791 * notice, this list of conditions and the following disclaimer.
792 * 2. Redistributions in binary form must reproduce the above copyright
793 * notice, this list of conditions and the following disclaimer in the
794 * documentation and/or other materials provided with the distribution.
796 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
797 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
799 * 4. Neither the name of the University nor the names of its contributors
800 * may be used to endorse or promote products derived from this software
801 * without specific prior written permission.
803 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
804 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
805 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
806 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
807 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
808 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
809 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
810 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
811 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
812 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF