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
13 #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 bool 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 */
42 STALLTIME = 5 /* Seconds when xfer considered "stalled" */
45 static ALWAYS_INLINE void progressmeter(int flag) {}
48 /* Read NMEMB bytes into PTR from STREAM. Returns the number of bytes read,
49 * and a short count if an eof or non-interrupt error is encountered. */
50 static size_t safe_fread(void *ptr, size_t nmemb, FILE *stream)
57 ret = fread(p, 1, nmemb, stream);
60 } while (nmemb && ferror(stream) && errno == EINTR);
62 return p - (char*)ptr;
65 /* Read a line or SIZE-1 bytes into S, whichever is less, from STREAM.
66 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
67 static char *safe_fgets(char *s, int size, FILE *stream)
73 ret = fgets(s, size, stream);
74 } while (ret == NULL && ferror(stream) && errno == EINTR);
79 #if ENABLE_FEATURE_WGET_AUTHENTICATION
80 /* Base64-encode character string. buf is assumed to be char buf[512]. */
81 static char *base64enc_512(char buf[512], const char *str)
83 unsigned len = strlen(str);
84 if (len > 512/4*3 - 10) /* paranoia */
86 bb_uuencode(buf, str, len, bb_uuenc_tbl_base64);
91 int wget_main(int argc, char **argv);
92 int wget_main(int argc, char **argv)
95 struct host_info server, target;
96 len_and_sockaddr *lsa;
103 char *dir_prefix = NULL;
104 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
105 char *extra_headers = NULL;
106 llist_t *headers_llist = 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 bool got_clen = 0; /* got content-length: from server */
114 bool 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";/* "User-Agent" header field */
117 static const char * const keywords[] = {
118 "content-length", "transfer-encoding", "chunked", "location", NULL
121 KEY_content_length = 1, KEY_transfer_encoding, KEY_chunked, KEY_location
124 WGET_OPT_CONTINUE = 0x1,
125 WGET_OPT_SPIDER = 0x2,
126 WGET_OPT_QUIET = 0x4,
127 WGET_OPT_OUTNAME = 0x8,
128 WGET_OPT_PREFIX = 0x10,
129 WGET_OPT_PROXY = 0x20,
130 WGET_OPT_USER_AGENT = 0x40,
131 WGET_OPT_PASSIVE = 0x80,
132 WGET_OPT_HEADER = 0x100,
134 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
135 static const char wget_longopts[] =
136 /* name, has_arg, val */
137 "continue\0" No_argument "c"
138 "spider\0" No_argument "s"
139 "quiet\0" No_argument "q"
140 "output-document\0" Required_argument "O"
141 "directory-prefix\0" Required_argument "P"
142 "proxy\0" Required_argument "Y"
143 "user-agent\0" Required_argument "U"
144 "passive-ftp\0" No_argument "\xff"
145 "header\0" Required_argument "\xfe"
147 applet_long_options = wget_longopts;
149 /* server.allocated = target.allocated = NULL; */
150 opt_complementary = "-1" USE_FEATURE_WGET_LONG_OPTIONS(":\xfe::");
151 opt = getopt32(argc, argv, "csqO:P:Y:U:",
152 &fname_out, &dir_prefix,
153 &proxy_flag, &user_agent
154 USE_FEATURE_WGET_LONG_OPTIONS(, &headers_llist)
156 if (strcmp(proxy_flag, "off") == 0) {
157 /* Use the proxy if necessary */
160 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
164 llist_t *ll = headers_llist;
166 size += strlen(ll->data) + 2;
169 extra_headers = cp = xmalloc(size);
170 while (headers_llist) {
171 cp += sprintf(cp, "%s\r\n", headers_llist->data);
172 headers_llist = headers_llist->link;
177 parse_url(argv[optind], &target);
178 server.host = target.host;
179 server.port = target.port;
181 /* 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)"); */
220 /* Determine where to start transfer */
221 if (LONE_DASH(fname_out)) {
223 opt &= ~WGET_OPT_CONTINUE;
225 if (opt & WGET_OPT_CONTINUE) {
226 output_fd = open(fname_out, O_WRONLY);
227 if (output_fd >= 0) {
228 beg_range = xlseek(output_fd, 0, SEEK_END);
230 /* File doesn't exist. We do not create file here yet.
231 We are not sure it exists on remove side */
234 /* We want to do exactly _one_ DNS lookup, since some
235 * sites (i.e. ftp.us.debian.org) use round-robin DNS
236 * and we want to connect to only one IP... */
237 lsa = xhost2sockaddr(server.host, server.port);
238 if (!(opt & WGET_OPT_QUIET)) {
239 fprintf(stderr, "Connecting to %s (%s)\n", server.host,
240 xmalloc_sockaddr2dotted(&lsa->sa, lsa->len));
241 /* We leak result of xmalloc_sockaddr2dotted */
244 if (use_proxy || !target.is_ftp) {
249 got_clen = chunked = 0;
252 bb_error_msg_and_die("too many redirections");
254 /* Open socket to http server */
255 if (sfp) fclose(sfp);
256 sfp = open_socket(lsa);
258 /* Send HTTP request. */
260 fprintf(sfp, "GET %stp://%s/%s HTTP/1.1\r\n",
261 target.is_ftp ? "f" : "ht", target.host,
264 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
267 fprintf(sfp, "Host: %s\r\nUser-Agent: %s\r\n",
268 target.host, user_agent);
270 #if ENABLE_FEATURE_WGET_AUTHENTICATION
272 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n"+6,
273 base64enc_512(buf, target.user));
275 if (use_proxy && server.user) {
276 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
277 base64enc_512(buf, server.user));
282 fprintf(sfp, "Range: bytes=%"OFF_FMT"d-\r\n", beg_range);
283 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
285 fputs(extra_headers, sfp);
287 fprintf(sfp, "Connection: close\r\n\r\n");
290 * Retrieve HTTP response line and check for "200" status code.
293 if (fgets(buf, sizeof(buf), sfp) == NULL)
294 bb_error_msg_and_die("no response from server");
297 str = skip_non_whitespace(str);
298 str = skip_whitespace(str);
299 // FIXME: no error check
300 // xatou wouldn't work: "200 OK"
305 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL)
306 /* eat all remaining headers */;
310 case 300: /* redirection */
320 /* Show first line only and kill any ESC tricks */
321 buf[strcspn(buf, "\n\r\x1b")] = '\0';
322 bb_error_msg_and_die("server returned error: %s", buf);
326 * Retrieve HTTP headers.
328 while ((str = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
329 /* gethdr did already convert the "FOO:" string to lowercase */
330 smalluint key = index_in_str_array(keywords, *&buf) + 1;
331 if (key == KEY_content_length) {
332 content_len = BB_STRTOOFF(str, NULL, 10);
333 if (errno || content_len < 0) {
334 bb_error_msg_and_die("content-length %s is garbage", str);
339 if (key == KEY_transfer_encoding) {
340 if (index_in_str_array(keywords, str_tolower(str)) + 1 != KEY_chunked)
341 bb_error_msg_and_die("server wants to do %s transfer encoding", str);
342 chunked = got_clen = 1;
344 if (key == KEY_location) {
346 /* free(target.allocated); */
347 target.path = /* target.allocated = */ xstrdup(str+1);
349 parse_url(str, &target);
350 if (use_proxy == 0) {
351 server.host = target.host;
352 server.port = target.port;
355 lsa = xhost2sockaddr(server.host, server.port);
360 } while (status >= 300);
370 target.user = xstrdup("anonymous:busybox@");
372 sfp = open_socket(lsa);
373 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
374 bb_error_msg_and_die("%s", buf+4);
377 * Splitting username:password pair,
380 str = strchr(target.user, ':');
383 switch (ftpcmd("USER ", target.user, sfp, buf)) {
387 if (ftpcmd("PASS ", str, sfp, buf) == 230)
389 /* FALLTHRU (failed login) */
391 bb_error_msg_and_die("ftp login: %s", buf+4);
394 ftpcmd("TYPE I", NULL, sfp, buf);
399 if (ftpcmd("SIZE ", target.path, sfp, buf) == 213) {
400 content_len = BB_STRTOOFF(buf+4, NULL, 10);
401 if (errno || content_len < 0) {
402 bb_error_msg_and_die("SIZE value is garbage");
408 * Entering passive mode
410 if (ftpcmd("PASV", NULL, sfp, buf) != 227) {
412 bb_error_msg_and_die("bad response to %s: %s", "PASV", buf);
414 // Response is "227 garbageN1,N2,N3,N4,P1,P2[)garbage]
415 // Server's IP is N1.N2.N3.N4 (we ignore it)
416 // Server's port for data connection is P1*256+P2
417 str = strrchr(buf, ')');
418 if (str) str[0] = '\0';
419 str = strrchr(buf, ',');
420 if (!str) goto pasv_error;
421 port = xatou_range(str+1, 0, 255);
423 str = strrchr(buf, ',');
424 if (!str) goto pasv_error;
425 port += xatou_range(str+1, 0, 255) * 256;
426 set_nport(lsa, htons(port));
427 dfp = open_socket(lsa);
430 sprintf(buf, "REST %"OFF_FMT"d", 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 RETR: %s", buf);
438 if (opt & WGET_OPT_SPIDER) {
439 if (ENABLE_FEATURE_CLEAN_UP)
448 fgets(buf, sizeof(buf), dfp);
449 content_len = STRTOOFF(buf, NULL, 16);
450 /* FIXME: error check?? */
453 /* Do it before progressmeter (want to have nice error message) */
455 output_fd = xopen(fname_out,
456 O_WRONLY|O_CREAT|O_EXCL|O_TRUNC);
458 if (!(opt & WGET_OPT_QUIET))
462 while (content_len > 0 || !got_clen) {
463 unsigned rdsz = sizeof(buf);
464 if (content_len < sizeof(buf) && (chunked || got_clen))
465 rdsz = (unsigned)content_len;
466 n = safe_fread(buf, rdsz, dfp);
469 if (full_write(output_fd, buf, n) != n) {
470 bb_perror_msg_and_die(bb_msg_write_error);
472 #if ENABLE_FEATURE_WGET_STATUSBAR
481 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
482 safe_fgets(buf, sizeof(buf), dfp);
483 content_len = STRTOOFF(buf, NULL, 16);
484 /* FIXME: error check? */
485 if (content_len == 0) {
486 chunked = 0; /* all done! */
490 if (n == 0 && ferror(dfp)) {
491 bb_perror_msg_and_die(bb_msg_read_error);
495 if (!(opt & WGET_OPT_QUIET))
498 if ((use_proxy == 0) && target.is_ftp) {
500 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
501 bb_error_msg_and_die("ftp error: %s", buf+4);
502 ftpcmd("QUIT", NULL, sfp, buf);
509 static void parse_url(char *src_url, struct host_info *h)
513 /* h->allocated = */ url = xstrdup(src_url);
515 if (strncmp(url, "http://", 7) == 0) {
516 h->port = bb_lookup_port("http", "tcp", 80);
519 } else if (strncmp(url, "ftp://", 6) == 0) {
520 h->port = bb_lookup_port("ftp", "tcp", 21);
524 bb_error_msg_and_die("not an http or ftp url: %s", url);
527 // "Real" wget 'http://busybox.net?var=a/b' sends this request:
528 // 'GET /?var=a/b HTTP 1.0'
529 // and saves 'index.html?var=a%2Fb' (we save 'b')
530 // wget 'http://busybox.net?login=john@doe':
531 // request: 'GET /?login=john@doe HTTP/1.0'
532 // saves: 'index.html?login=john@doe' (we save '?login=john@doe')
533 // wget 'http://busybox.net#test/test':
534 // request: 'GET / HTTP/1.0'
535 // saves: 'index.html' (we save 'test')
537 // We also don't add unique .N suffix if file exists...
538 sp = strchr(h->host, '/');
539 p = strchr(h->host, '?'); if (!sp || (p && sp > p)) sp = p;
540 p = strchr(h->host, '#'); if (!sp || (p && sp > p)) sp = p;
542 /* must be writable because of bb_get_last_path_component() */
543 static char nullstr[] = "";
545 } else if (*sp == '/') {
548 } else { // '#' or '?'
549 // http://busybox.net?login=john@doe is a valid URL
550 // memmove converts to:
551 // http:/busybox.nett?login=john@doe...
552 memmove(h->host-1, h->host, sp - h->host);
558 sp = strrchr(h->host, '@');
570 static FILE *open_socket(len_and_sockaddr *lsa)
574 /* glibc 2.4 seems to try seeking on it - ??! */
575 /* hopefully it understands what ESPIPE means... */
576 fp = fdopen(xconnect_stream(lsa), "r+");
578 bb_perror_msg_and_die("fdopen");
584 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
591 /* retrieve header line */
592 if (fgets(buf, bufsiz, fp) == NULL)
595 /* see if we are at the end of the headers */
596 for (s = buf; *s == '\r'; ++s)
601 /* convert the header name to lower case */
602 for (s = buf; isalnum(*s) || *s == '-'; ++s)
605 /* verify we are at the end of the header name */
607 bb_error_msg_and_die("bad header line: %s", buf);
609 /* locate the start of the header value */
610 for (*s++ = '\0'; *s == ' ' || *s == '\t'; ++s)
614 /* locate the end of header */
615 while (*s != '\0' && *s != '\r' && *s != '\n')
618 /* end of header found */
624 /* Rats! The buffer isn't big enough to hold the entire header value. */
625 while (c = getc(fp), c != EOF && c != '\n')
631 static int ftpcmd(const char *s1, const char *s2, FILE *fp, char *buf)
636 fprintf(fp, "%s%s\r\n", s1, s2);
643 if (fgets(buf, 510, fp) == NULL) {
644 bb_perror_msg_and_die("error getting response");
646 buf_ptr = strstr(buf, "\r\n");
650 } while (!isdigit(buf[0]) || buf[3] != ' ');
653 result = xatoi_u(buf);
658 #if ENABLE_FEATURE_WGET_STATUSBAR
659 /* Stuff below is from BSD rcp util.c, as added to openshh.
660 * Original copyright notice is retained at the end of this file.
666 get_terminal_width_height(0, &width, NULL);
671 updateprogressmeter(int ignore)
673 int save_errno = errno;
679 static void alarmtimer(int iwait)
681 struct itimerval itv;
683 itv.it_value.tv_sec = iwait;
684 itv.it_value.tv_usec = 0;
685 itv.it_interval = itv.it_value;
686 setitimer(ITIMER_REAL, &itv, NULL);
690 progressmeter(int flag)
692 static unsigned lastupdate_sec;
693 static unsigned start_sec;
694 static off_t lastsize, totalsize;
697 unsigned since_last_update, elapsed;
701 if (flag == -1) { /* first call to progressmeter */
702 start_sec = monotonic_sec();
703 lastupdate_sec = start_sec;
705 totalsize = content_len + beg_range; /* as content_len changes.. */
709 if (totalsize != 0 && !chunked) {
710 /* long long helps to have it working even if !LFS */
711 ratio = (unsigned) (100ULL * (transferred+beg_range) / totalsize);
712 if (ratio > 100) ratio = 100;
715 fprintf(stderr, "\r%-20.20s%4d%% ", curfile, ratio);
717 barlength = getttywidth() - 49;
719 /* god bless gcc for variable arrays :) */
720 i = barlength * ratio / 100;
725 fprintf(stderr, "|%s%*s|", buf, barlength - i, "");
729 abbrevsize = transferred + beg_range;
730 while (abbrevsize >= 100000) {
734 /* see http://en.wikipedia.org/wiki/Tera */
735 fprintf(stderr, "%6d%c ", (int)abbrevsize, " kMGTPEZY"[i]);
737 // Nuts! Ain't it easier to update progress meter ONLY when we transferred++?
738 // FIXME: get rid of alarmtimer + updateprogressmeter mess
740 elapsed = monotonic_sec();
741 since_last_update = elapsed - lastupdate_sec;
742 if (transferred > lastsize) {
743 lastupdate_sec = elapsed;
744 lastsize = transferred;
745 if (since_last_update >= STALLTIME) {
746 /* We "cut off" these seconds from elapsed time
747 * by adjusting start time */
748 start_sec += since_last_update;
750 since_last_update = 0; /* we are un-stalled now */
752 elapsed -= start_sec; /* now it's "elapsed since start" */
754 if (since_last_update >= STALLTIME) {
755 fprintf(stderr, " - stalled -");
757 off_t to_download = totalsize - beg_range;
758 if (transferred <= 0 || (int)elapsed <= 0 || transferred > to_download || chunked) {
759 fprintf(stderr, "--:--:-- ETA");
761 /* to_download / (transferred/elapsed) - elapsed: */
762 int eta = (int) ((unsigned long long)to_download*elapsed/transferred - elapsed);
763 /* (long long helps to have working ETA even if !LFS) */
765 fprintf(stderr, "%02d:%02d:%02d ETA", eta / 3600, i / 60, i % 60);
769 if (flag == -1) { /* first call to progressmeter */
771 sa.sa_handler = updateprogressmeter;
772 sigemptyset(&sa.sa_mask);
773 sa.sa_flags = SA_RESTART;
774 sigaction(SIGALRM, &sa, NULL);
776 } else if (flag == 1) { /* last call to progressmeter */
782 #endif /* FEATURE_WGET_STATUSBAR */
784 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
785 * much of which was blatantly stolen from openssh. */
788 * Copyright (c) 1992, 1993
789 * The Regents of the University of California. All rights reserved.
791 * Redistribution and use in source and binary forms, with or without
792 * modification, are permitted provided that the following conditions
794 * 1. Redistributions of source code must retain the above copyright
795 * notice, this list of conditions and the following disclaimer.
796 * 2. Redistributions in binary form must reproduce the above copyright
797 * notice, this list of conditions and the following disclaimer in the
798 * documentation and/or other materials provided with the distribution.
800 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
801 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
803 * 4. Neither the name of the University nor the names of its contributors
804 * may be used to endorse or promote products derived from this software
805 * without specific prior written permission.
807 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
808 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
809 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
810 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
811 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
812 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
813 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
814 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
815 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
816 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF