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 FILEOFF_TYPE content_len; /* Content-length of the file */
33 static FILEOFF_TYPE beg_range; /* Range at which continue begins */
34 static FILEOFF_TYPE transferred; /* Number of bytes transferred so far */
35 static int chunked; /* chunked transfer encoding */
36 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
37 static void progressmeter(int flag);
38 static char *curfile; /* Name of current file being transferred */
39 static struct timeval start; /* Time a transfer started */
41 STALLTIME = 5 /* Seconds when xfer considered "stalled" */
44 static void progressmeter(int flag) {}
47 /* Read NMEMB elements of SIZE bytes into PTR from STREAM. Returns the
48 * number of elements read, and a short count if an eof or non-interrupt
49 * error is encountered. */
50 static size_t safe_fread(void *ptr, size_t size, size_t nmemb, FILE *stream)
56 ret += fread((char *)ptr + (ret * size), size, nmemb - ret, stream);
57 } while (ret < nmemb && ferror(stream) && errno == EINTR);
62 /* Read a line or SIZE - 1 bytes into S, whichever is less, from STREAM.
63 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
64 static char *safe_fgets(char *s, int size, FILE *stream)
70 ret = fgets(s, size, stream);
71 } while (ret == NULL && ferror(stream) && errno == EINTR);
76 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
78 * Base64-encode character string and return the string.
80 static char *base64enc(unsigned char *p, char *buf, int len)
82 bb_uuencode(p, buf, len, bb_uuenc_tbl_base64);
87 #define WGET_OPT_CONTINUE 1
88 #define WGET_OPT_QUIET 2
89 #define WGET_OPT_PASSIVE 4
90 #define WGET_OPT_OUTNAME 8
91 #define WGET_OPT_HEADER 16
92 #define WGET_OPT_PREFIX 32
93 #define WGET_OPT_PROXY 64
94 #define WGET_OPT_USER_AGENT 128
96 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
97 static const struct option wget_long_options[] = {
98 { "continue", 0, NULL, 'c' },
99 { "quiet", 0, NULL, 'q' },
100 { "passive-ftp", 0, NULL, 139 }, /* FIXME: what is this - 139?? */
101 { "output-document", 1, NULL, 'O' },
102 { "header", 1, NULL, 131 },
103 { "directory-prefix",1, NULL, 'P' },
104 { "proxy", 1, NULL, 'Y' },
105 { "user-agent", 1, NULL, 'U' },
110 int wget_main(int argc, char **argv)
112 int n, try=5, status;
116 char *dir_prefix=NULL;
118 char extra_headers[1024];
119 char *extra_headers_ptr = extra_headers;
120 int extra_headers_left = sizeof(extra_headers);
121 struct host_info server, target;
122 struct sockaddr_in s_in;
123 llist_t *headers_llist = NULL;
125 /* server.allocated = target.allocated = NULL; */
127 FILE *sfp = NULL; /* socket to web/ftp server */
128 FILE *dfp = NULL; /* socket to ftp server (data) */
129 char *fname_out = NULL; /* where to direct output (-O) */
130 int got_clen = 0; /* got content-length: from server */
132 int use_proxy = 1; /* Use proxies if env vars are set */
133 const char *proxy_flag = "on"; /* Use proxies if env vars are set */
134 const char *user_agent = "Wget";/* Content of the "User-Agent" header field */
137 * Crack command line.
139 opt_complementary = "-1:\203::";
140 #if ENABLE_FEATURE_WGET_LONG_OPTIONS
141 applet_long_options = wget_long_options;
143 opt = getopt32(argc, argv, "cq\213O:\203:P:Y:U:",
144 &fname_out, &headers_llist,
145 &dir_prefix, &proxy_flag, &user_agent);
146 if (strcmp(proxy_flag, "off") == 0) {
147 /* Use the proxy if necessary. */
150 if (opt & WGET_OPT_HEADER) {
151 while (headers_llist) {
152 int arglen = strlen(headers_llist->data);
153 if (extra_headers_left - arglen - 2 <= 0)
154 bb_error_msg_and_die("extra_headers buffer too small "
155 "(need %i)", extra_headers_left - arglen);
156 strcpy(extra_headers_ptr, headers_llist->data);
157 extra_headers_ptr += arglen;
158 extra_headers_left -= ( arglen + 2 );
159 *extra_headers_ptr++ = '\r';
160 *extra_headers_ptr++ = '\n';
161 *(extra_headers_ptr + 1) = 0;
162 headers_llist = headers_llist->link;
166 parse_url(argv[optind], &target);
167 server.host = target.host;
168 server.port = target.port;
171 * Use the proxy if necessary.
174 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
175 if (proxy && *proxy) {
176 parse_url(proxy, &server);
182 /* Guess an output filename */
184 // Dirty hack. Needed because bb_get_last_path_component
185 // will destroy trailing / by storing '\0' in last byte!
186 if (*target.path && target.path[strlen(target.path)-1] != '/') {
188 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
191 bb_get_last_path_component(target.path);
193 if (!fname_out || !fname_out[0]) {
195 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
200 if (dir_prefix != NULL)
201 fname_out = concat_path_file(dir_prefix, fname_out);
202 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
204 curfile = bb_get_last_path_component(fname_out);
207 if ((opt & WGET_OPT_CONTINUE) && !fname_out)
208 bb_error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
211 * Determine where to start transfer.
213 if (!strcmp(fname_out, "-")) {
215 opt |= WGET_OPT_QUIET;
216 opt &= ~WGET_OPT_CONTINUE;
217 } else if (opt & WGET_OPT_CONTINUE) {
218 output_fd = open(fname_out, O_WRONLY|O_LARGEFILE);
219 if (output_fd >= 0) {
220 beg_range = LSEEK(output_fd, 0, SEEK_END);
221 if (beg_range == (FILEOFF_TYPE)-1)
222 bb_perror_msg_and_die("lseek");
224 /* File doesn't exist. We do not create file here yet.
225 We are not sure it exists on remove side */
228 /* We want to do exactly _one_ DNS lookup, since some
229 * sites (i.e. ftp.us.debian.org) use round-robin DNS
230 * and we want to connect to only one IP... */
231 bb_lookup_host(&s_in, server.host);
232 s_in.sin_port = server.port;
233 if (!(opt & WGET_OPT_QUIET)) {
234 printf("Connecting to %s[%s]:%d\n",
235 server.host, inet_ntoa(s_in.sin_addr), ntohs(server.port));
238 if (use_proxy || !target.is_ftp) {
243 got_clen = chunked = 0;
246 bb_error_msg_and_die("too many redirections");
249 * Open socket to http server
251 if (sfp) fclose(sfp);
252 sfp = open_socket(&s_in);
258 const char *format = "GET %stp://%s:%d/%s HTTP/1.1\r\n";
259 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
260 if (strchr(target.host, ':'))
261 format = "GET %stp://[%s]:%d/%s HTTP/1.1\r\n";
264 target.is_ftp ? "f" : "ht", target.host,
265 ntohs(target.port), target.path);
267 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
270 fprintf(sfp, "Host: %s\r\nUser-Agent: %s\r\n", target.host,
273 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
275 fprintf(sfp, "Authorization: Basic %s\r\n",
276 base64enc((unsigned char*)target.user, buf, sizeof(buf)));
278 if (use_proxy && server.user) {
279 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
280 base64enc((unsigned char*)server.user, buf, sizeof(buf)));
285 fprintf(sfp, "Range: bytes="FILEOFF_FMT"-\r\n", beg_range);
286 if(extra_headers_left < sizeof(extra_headers))
287 fputs(extra_headers,sfp);
288 fprintf(sfp,"Connection: close\r\n\r\n");
291 * Retrieve HTTP response line and check for "200" status code.
294 if (fgets(buf, sizeof(buf), sfp) == NULL)
295 bb_error_msg_and_die("no response from server");
298 while (*s != '\0' && !isspace(*s)) ++s;
299 while (isspace(*s)) ++s;
300 switch (status = atoi(s)) {
303 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL)
304 /* eat all remaining headers */;
308 case 300: /* redirection */
319 bb_error_msg_and_die("server returned error %s: %s", s, buf);
323 * Retrieve HTTP headers.
325 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
326 if (strcasecmp(buf, "content-length") == 0) {
327 if (SAFE_STRTOOFF(s, &content_len) || content_len < 0) {
328 bb_error_msg_and_die("content-length %s is garbage", s);
333 if (strcasecmp(buf, "transfer-encoding") == 0) {
334 if (strcasecmp(s, "chunked") == 0) {
335 chunked = got_clen = 1;
337 bb_error_msg_and_die("server wants to do %s transfer encoding", s);
340 if (strcasecmp(buf, "location") == 0) {
342 /* free(target.allocated); */
343 target.path = /* target.allocated = */ xstrdup(s+1);
345 parse_url(s, &target);
346 if (use_proxy == 0) {
347 server.host = target.host;
348 server.port = target.port;
350 bb_lookup_host(&s_in, server.host);
351 s_in.sin_port = server.port;
356 } while(status >= 300);
366 target.user = xstrdup("anonymous:busybox@");
368 sfp = open_socket(&s_in);
369 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
370 bb_error_msg_and_die("%s", buf+4);
373 * Splitting username:password pair,
376 s = strchr(target.user, ':');
379 switch (ftpcmd("USER ", target.user, sfp, buf)) {
383 if (ftpcmd("PASS ", s, sfp, buf) == 230)
385 /* FALLTHRU (failed login) */
387 bb_error_msg_and_die("ftp login: %s", buf+4);
390 ftpcmd("TYPE I", NULL, sfp, buf);
395 if (ftpcmd("SIZE ", target.path, sfp, buf) == 213) {
396 if (SAFE_STRTOOFF(buf+4, &content_len) || content_len < 0) {
397 bb_error_msg_and_die("SIZE value is garbage");
403 * Entering passive mode
405 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
406 bb_error_msg_and_die("PASV: %s", buf+4);
407 s = strrchr(buf, ',');
410 s = strrchr(buf, ',');
411 port += atoi(s+1) * 256;
412 s_in.sin_port = htons(port);
413 dfp = open_socket(&s_in);
416 sprintf(buf, "REST "FILEOFF_FMT, beg_range);
417 if (ftpcmd(buf, NULL, sfp, buf) == 350)
418 content_len -= beg_range;
421 if (ftpcmd("RETR ", target.path, sfp, buf) > 150)
422 bb_error_msg_and_die("RETR: %s", buf+4);
430 fgets(buf, sizeof(buf), dfp);
431 content_len = STRTOOFF(buf, (char **) NULL, 16);
432 /* FIXME: error check?? */
435 /* Do it before progressmeter (want to have nice error message) */
437 output_fd = xopen3(fname_out,
438 O_WRONLY|O_CREAT|O_EXCL|O_TRUNC|O_LARGEFILE, 0666);
440 if (!(opt & WGET_OPT_QUIET))
444 while (content_len > 0 || !got_clen) {
445 unsigned rdsz = sizeof(buf);
446 if (content_len < sizeof(buf) && (chunked || got_clen))
447 rdsz = (unsigned)content_len;
448 n = safe_fread(buf, 1, rdsz, dfp);
451 if (full_write(output_fd, buf, n) != n) {
452 bb_perror_msg_and_die(bb_msg_write_error);
454 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
463 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
464 safe_fgets(buf, sizeof(buf), dfp);
465 content_len = STRTOOFF(buf, (char **) NULL, 16);
466 /* FIXME: error check? */
467 if (content_len == 0) {
468 chunked = 0; /* all done! */
472 if (n == 0 && ferror(dfp)) {
473 bb_perror_msg_and_die(bb_msg_read_error);
477 if (!(opt & WGET_OPT_QUIET))
480 if ((use_proxy == 0) && target.is_ftp) {
482 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
483 bb_error_msg_and_die("ftp error: %s", buf+4);
484 ftpcmd("QUIT", NULL, sfp, buf);
490 static void parse_url(char *src_url, struct host_info *h)
492 char *url, *p, *cp, *sp, *up, *pp;
494 /* h->allocated = */ url = xstrdup(src_url);
496 if (strncmp(url, "http://", 7) == 0) {
497 h->port = bb_lookup_port("http", "tcp", 80);
500 } else if (strncmp(url, "ftp://", 6) == 0) {
501 h->port = bb_lookup_port("ftp", "tcp", 21);
505 bb_error_msg_and_die("not an http or ftp url: %s", url);
508 // "Real" wget 'http://busybox.net?var=a/b' sends this request:
509 // 'GET /?var=a/b HTTP 1.0'
510 // and saves 'index.html?var=a%2Fb' (we save 'b')
511 // wget 'http://busybox.net?login=john@doe':
512 // request: 'GET /?login=john@doe HTTP/1.0'
513 // saves: 'index.html?login=john@doe' (we save ?login=john@doe)
514 // wget 'http://busybox.net#test/test':
515 // request: 'GET / HTTP/1.0'
516 // saves: 'index.html' (we save 'test')
518 // We also don't add unique .N suffix if file exists...
519 sp = strchr(h->host, '/');
520 p = strchr(h->host, '?'); if (!sp || (p && sp > p)) sp = p;
521 p = strchr(h->host, '#'); if (!sp || (p && sp > p)) sp = p;
524 } else if (*sp == '/') {
527 } else { // '#' or '?'
528 // http://busybox.net?login=john@doe is a valid URL
529 // memmove converts to:
530 // http:/busybox.nett?login=john@doe...
531 memmove(h->host-1, h->host, sp - h->host);
537 up = strrchr(h->host, '@');
547 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
548 if (h->host[0] == '[') {
552 while (*ep == ':' || isxdigit(*ep))
562 cp = strchr(pp, ':');
565 h->port = htons(atoi(cp));
570 static FILE *open_socket(struct sockaddr_in *s_in)
574 fp = fdopen(xconnect(s_in), "r+");
576 bb_perror_msg_and_die("fdopen");
582 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
589 /* retrieve header line */
590 if (fgets(buf, bufsiz, fp) == NULL)
593 /* see if we are at the end of the headers */
594 for (s = buf ; *s == '\r' ; ++s)
599 /* convert the header name to lower case */
600 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
603 /* verify we are at the end of the header name */
605 bb_error_msg_and_die("bad header line: %s", buf);
607 /* locate the start of the header value */
608 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
612 /* locate the end of header */
613 while (*s != '\0' && *s != '\r' && *s != '\n')
616 /* end of header found */
622 /* Rats! The buffer isn't big enough to hold the entire header value. */
623 while (c = getc(fp), c != EOF && c != '\n')
629 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
633 fprintf(fp, "%s%s\r\n", s1, s2);
640 if (fgets(buf, 510, fp) == NULL) {
641 bb_perror_msg_and_die("fgets");
643 buf_ptr = strstr(buf, "\r\n");
647 } while (!isdigit(buf[0]) || buf[3] != ' ');
652 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
653 /* Stuff below is from BSD rcp util.c, as added to openshh.
654 * Original copyright notice is retained at the end of this file.
660 get_terminal_width_height(0, &width, NULL);
665 updateprogressmeter(int ignore)
667 int save_errno = errno;
673 static void alarmtimer(int iwait)
675 struct itimerval itv;
677 itv.it_value.tv_sec = iwait;
678 itv.it_value.tv_usec = 0;
679 itv.it_interval = itv.it_value;
680 setitimer(ITIMER_REAL, &itv, NULL);
685 progressmeter(int flag)
687 static struct timeval lastupdate;
688 static FILEOFF_TYPE lastsize, totalsize;
690 struct timeval now, td, tvwait;
691 FILEOFF_TYPE abbrevsize;
692 int elapsed, ratio, barlength, i;
695 if (flag == -1) { /* first call to progressmeter */
696 (void) gettimeofday(&start, (struct timezone *) 0);
699 totalsize = content_len + beg_range; /* as content_len changes.. */
702 (void) gettimeofday(&now, (struct timezone *) 0);
704 if (totalsize != 0 && !chunked) {
705 ratio = (int) (100 * (transferred+beg_range) / totalsize);
706 ratio = MIN(ratio, 100);
709 fprintf(stderr, "\r%-20.20s%4d%% ", curfile, ratio);
711 barlength = getttywidth() - 51;
712 if (barlength > 0 && barlength < sizeof(buf)) {
713 i = barlength * ratio / 100;
715 memset(buf + i, ' ', barlength - i);
716 buf[barlength] = '\0';
717 fprintf(stderr, "|%s|", buf);
720 abbrevsize = transferred + beg_range;
721 while (abbrevsize >= 100000) {
725 /* See http://en.wikipedia.org/wiki/Tera */
726 fprintf(stderr, "%6d %c%c ", (int)abbrevsize, " KMGTPEZY"[i], i?'B':' ');
728 timersub(&now, &lastupdate, &tvwait);
729 if (transferred > lastsize) {
731 lastsize = transferred;
732 if (tvwait.tv_sec >= STALLTIME)
733 timeradd(&start, &tvwait, &start);
736 timersub(&now, &start, &td);
739 if (tvwait.tv_sec >= STALLTIME) {
740 fprintf(stderr, " - stalled -");
742 FILEOFF_TYPE to_download = totalsize - beg_range;
743 if (transferred <= 0 || elapsed <= 0 || transferred > to_download || chunked) {
744 fprintf(stderr, "--:--:-- ETA");
746 /* to_download / (transferred/elapsed) - elapsed: */
747 int eta = (int) (to_download*elapsed/transferred - elapsed);
749 fprintf(stderr, "%02d:%02d:%02d ETA", eta / 3600, i / 60, i % 60);
753 if (flag == -1) { /* first call to progressmeter */
755 sa.sa_handler = updateprogressmeter;
756 sigemptyset(&sa.sa_mask);
757 sa.sa_flags = SA_RESTART;
758 sigaction(SIGALRM, &sa, NULL);
760 } else if (flag == 1) { /* last call to progressmeter */
768 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
769 * much of which was blatantly stolen from openssh. */
772 * Copyright (c) 1992, 1993
773 * The Regents of the University of California. All rights reserved.
775 * Redistribution and use in source and binary forms, with or without
776 * modification, are permitted provided that the following conditions
778 * 1. Redistributions of source code must retain the above copyright
779 * notice, this list of conditions and the following disclaimer.
780 * 2. Redistributions in binary form must reproduce the above copyright
781 * notice, this list of conditions and the following disclaimer in the
782 * documentation and/or other materials provided with the distribution.
784 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
785 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
787 * 4. Neither the name of the University nor the names of its contributors
788 * may be used to endorse or promote products derived from this software
789 * without specific prior written permission.
791 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
792 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
793 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
794 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
795 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
796 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
797 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
798 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
799 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
800 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
803 * $Id: wget.c,v 1.75 2004/10/08 08:27:40 andersen Exp $