1 /* vi: set sw=4 ts=4: */
3 * wget - retrieve a file using HTTP or FTP
5 * Chip Rosenthal Covad Communications <chip@laserlink.net>
17 #include <sys/ioctl.h>
20 #include <sys/types.h>
22 #include <sys/socket.h>
23 #include <netinet/in.h>
24 #include <arpa/inet.h>
42 static void parse_url(char *url, struct host_info *h);
43 static FILE *open_socket(struct sockaddr_in *s_in);
44 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
45 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf);
47 /* Globals (can be accessed from signal handlers */
48 static off_t filesize = 0; /* content-length of the file */
49 static int chunked = 0; /* chunked transfer encoding */
50 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
51 static void progressmeter(int flag);
52 static char *curfile; /* Name of current file being transferred. */
53 static struct timeval start; /* Time a transfer started. */
54 static volatile unsigned long statbytes = 0; /* Number of bytes transferred so far. */
55 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
56 static const int STALLTIME = 5;
59 static void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
61 if (output != stdout && do_continue==0) {
67 /* Read NMEMB elements of SIZE bytes into PTR from STREAM. Returns the
68 * number of elements read, and a short count if an eof or non-interrupt
69 * error is encountered. */
70 static size_t safe_fread(void *ptr, size_t size, size_t nmemb, FILE *stream)
76 ret += fread((char *)ptr + (ret * size), size, nmemb - ret, stream);
77 } while (ret < nmemb && ferror(stream) && errno == EINTR);
82 /* Write NMEMB elements of SIZE bytes from PTR to STREAM. Returns the
83 * number of elements written, and a short count if an eof or non-interrupt
84 * error is encountered. */
85 static size_t safe_fwrite(void *ptr, size_t size, size_t nmemb, FILE *stream)
91 ret += fwrite((char *)ptr + (ret * size), size, nmemb - ret, stream);
92 } while (ret < nmemb && ferror(stream) && errno == EINTR);
97 /* Read a line or SIZE - 1 bytes into S, whichever is less, from STREAM.
98 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
99 static char *safe_fgets(char *s, int size, FILE *stream)
105 ret = fgets(s, size, stream);
106 } while (ret == NULL && ferror(stream) && errno == EINTR);
111 #define close_delete_and_die(s...) { \
112 close_and_delete_outfile(output, fname_out, do_continue); \
113 bb_error_msg_and_die(s); }
116 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
118 * Base64-encode character string
119 * oops... isn't something similar in uuencode.c?
120 * It would be better to use already existing code
122 char *base64enc(char *p, char *buf, int len) {
124 char al[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
130 bb_error_msg_and_die("buffer overflow");
131 *(s++) = al[(*p >> 2) & 0x3F];
132 *(s++) = al[((*p << 4) & 0x30) | ((*(p+1) >> 4) & 0x0F)];
136 *(s++) = al[((*p << 2) & 0x3C) | ((*(p+1) >> 6) & 0x03)];
138 *(s++) = al[*(p++) & 0x3F];
145 #define WGET_OPT_CONTINUE 1
146 #define WGET_OPT_QUIET 2
147 #define WGET_OPT_PASSIVE 4
148 #define WGET_OPT_OUTNAME 8
149 #define WGET_OPT_HEADER 16
150 #define WGET_OPT_PREFIX 32
151 #define WGET_OPT_PROXY 64
153 static const struct option wget_long_options[] = {
154 { "continue", 0, NULL, 'c' },
155 { "quiet", 0, NULL, 'q' },
156 { "passive-ftp", 0, NULL, 139 },
157 { "output-document", 1, NULL, 'O' },
158 { "header", 1, NULL, 131 },
159 { "directory-prefix",1, NULL, 'P' },
160 { "proxy", 1, NULL, 'Y' },
164 int wget_main(int argc, char **argv)
166 int n, try=5, status;
170 char *dir_prefix=NULL;
173 char extra_headers[1024];
174 char *extra_headers_ptr = extra_headers;
175 int extra_headers_left = sizeof(extra_headers);
176 struct host_info server, target;
177 struct sockaddr_in s_in;
178 llist_t *headers_llist = NULL;
180 FILE *sfp = NULL; /* socket to web/ftp server */
181 FILE *dfp = NULL; /* socket to ftp server (data) */
182 char *fname_out = NULL; /* where to direct output (-O) */
183 int do_continue = 0; /* continue a prev transfer (-c) */
184 long beg_range = 0L; /* range at which continue begins */
185 int got_clen = 0; /* got content-length: from server */
186 FILE *output; /* socket to web server */
187 int quiet_flag = FALSE; /* Be verry, verry quiet... */
188 int noproxy = 0; /* Use proxies if env vars are set */
189 char *proxy_flag = "on"; /* Use proxies if env vars are set */
192 * Crack command line.
194 bb_opt_complementaly = "\203*";
195 bb_applet_long_options = wget_long_options;
196 opt = bb_getopt_ulflags(argc, argv, "cq\213O:\203:P:Y:", &fname_out, &headers_llist, &dir_prefix, &proxy_flag);
197 if (opt & WGET_OPT_CONTINUE) {
200 if (opt & WGET_OPT_QUIET) {
203 if (strcmp(proxy_flag, "on") == 0) {
204 /* Use the proxy if necessary. */
205 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
207 parse_url(bb_xstrdup(proxy), &server);
209 if (opt & WGET_OPT_HEADER) {
210 while (headers_llist) {
211 int arglen = strlen(headers_llist->data);
212 if (extra_headers_left - arglen - 2 <= 0)
213 bb_error_msg_and_die("extra_headers buffer too small(need %i)", extra_headers_left - arglen);
214 strcpy(extra_headers_ptr, headers_llist->data);
215 extra_headers_ptr += arglen;
216 extra_headers_left -= ( arglen + 2 );
217 *extra_headers_ptr++ = '\r';
218 *extra_headers_ptr++ = '\n';
219 *(extra_headers_ptr + 1) = 0;
220 headers_llist = headers_llist->link;
223 if (argc - optind != 1)
226 parse_url(argv[optind], &target);
227 server.host = target.host;
228 server.port = target.port;
231 * Use the proxy if necessary.
234 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
236 parse_url(bb_xstrdup(proxy), &server);
239 /* Guess an output filename */
242 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
245 bb_get_last_path_component(target.path);
246 if (fname_out==NULL || strlen(fname_out)<1) {
248 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
253 if (dir_prefix != NULL)
254 fname_out = concat_path_file(dir_prefix, fname_out);
255 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
257 curfile = bb_get_last_path_component(fname_out);
260 if (do_continue && !fname_out)
261 bb_error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
265 * Open the output file stream.
267 if (strcmp(fname_out, "-") == 0) {
271 output = bb_xfopen(fname_out, (do_continue ? "a" : "w"));
275 * Determine where to start transfer.
278 if (fstat(fileno(output), &sbuf) < 0)
279 bb_perror_msg_and_die("fstat()");
280 if (sbuf.st_size > 0)
281 beg_range = sbuf.st_size;
286 /* We want to do exactly _one_ DNS lookup, since some
287 * sites (i.e. ftp.us.debian.org) use round-robin DNS
288 * and we want to connect to only one IP... */
289 bb_lookup_host(&s_in, server.host);
290 s_in.sin_port = server.port;
291 if (quiet_flag==FALSE) {
292 fprintf(stdout, "Connecting to %s[%s]:%d\n",
293 server.host, inet_ntoa(s_in.sin_addr), ntohs(server.port));
296 if (proxy || !target.is_ftp) {
301 got_clen = chunked = 0;
304 close_delete_and_die("too many redirections");
307 * Open socket to http server
309 if (sfp) fclose(sfp);
310 sfp = open_socket(&s_in);
316 const char *format = "GET %stp://%s:%d/%s HTTP/1.1\r\n";
317 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
318 if (strchr (target.host, ':'))
319 format = "GET %stp://[%s]:%d/%s HTTP/1.1\r\n";
322 target.is_ftp ? "f" : "ht", target.host,
323 target.port, target.path);
325 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
328 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
330 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
332 fprintf(sfp, "Authorization: Basic %s\r\n",
333 base64enc(target.user, buf, sizeof(buf)));
335 if (proxy && server.user) {
336 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
337 base64enc(server.user, buf, sizeof(buf)));
342 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
343 if(extra_headers_left < sizeof(extra_headers))
344 fputs(extra_headers,sfp);
345 fprintf(sfp,"Connection: close\r\n\r\n");
348 * Retrieve HTTP response line and check for "200" status code.
351 if (fgets(buf, sizeof(buf), sfp) == NULL)
352 close_delete_and_die("no response from server");
354 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
356 for ( ; isspace(*s) ; ++s)
358 switch (status = atoi(s)) {
361 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
364 if (do_continue && output != stdout)
365 output = freopen(fname_out, "w", output);
368 case 300: /* redirection */
379 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
383 * Retrieve HTTP headers.
385 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
386 if (strcasecmp(buf, "content-length") == 0) {
391 if (strcasecmp(buf, "transfer-encoding") == 0) {
392 if (strcasecmp(s, "chunked") == 0) {
393 chunked = got_clen = 1;
395 close_delete_and_die("server wants to do %s transfer encoding", s);
398 if (strcasecmp(buf, "location") == 0) {
400 target.path = bb_xstrdup(s+1);
402 parse_url(bb_xstrdup(s), &target);
404 server.host = target.host;
405 server.port = target.port;
410 } while(status >= 300);
420 target.user = bb_xstrdup("anonymous:busybox@");
422 sfp = open_socket(&s_in);
423 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
424 close_delete_and_die("%s", buf+4);
427 * Splitting username:password pair,
430 s = strchr(target.user, ':');
433 switch(ftpcmd("USER ", target.user, sfp, buf)) {
437 if (ftpcmd("PASS ", s, sfp, buf) == 230)
439 /* FALLTHRU (failed login) */
441 close_delete_and_die("ftp login: %s", buf+4);
444 ftpcmd("CDUP", NULL, sfp, buf);
445 ftpcmd("TYPE I", NULL, sfp, buf);
450 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
451 filesize = atol(buf+4);
456 * Entering passive mode
458 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
459 close_delete_and_die("PASV: %s", buf+4);
460 s = strrchr(buf, ',');
463 s = strrchr(buf, ',');
464 port += atoi(s+1) * 256;
465 s_in.sin_port = htons(port);
466 dfp = open_socket(&s_in);
469 sprintf(buf, "REST %ld", beg_range);
470 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
471 if (output != stdout)
472 output = freopen(fname_out, "w", output);
475 filesize -= beg_range;
478 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
479 close_delete_and_die("RETR: %s", buf+4);
488 fgets(buf, sizeof(buf), dfp);
489 filesize = strtol(buf, (char **) NULL, 16);
491 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
492 if (quiet_flag==FALSE)
496 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, ((chunked || got_clen) && (filesize < sizeof(buf)) ? filesize : sizeof(buf)), dfp)) > 0) {
497 if (safe_fwrite(buf, 1, n, output) != n) {
498 bb_perror_msg_and_die("write error");
500 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
509 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
510 safe_fgets(buf, sizeof(buf), dfp);
511 filesize = strtol(buf, (char **) NULL, 16);
513 chunked = 0; /* all done! */
517 if (n == 0 && ferror(dfp)) {
518 bb_perror_msg_and_die("network read error");
521 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
522 if (quiet_flag==FALSE)
525 if (!proxy && target.is_ftp) {
527 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
528 bb_error_msg_and_die("ftp error: %s", buf+4);
529 ftpcmd("QUIT", NULL, sfp, buf);
535 void parse_url(char *url, struct host_info *h)
537 char *cp, *sp, *up, *pp;
539 if (strncmp(url, "http://", 7) == 0) {
540 h->port = bb_lookup_port("http", "tcp", 80);
543 } else if (strncmp(url, "ftp://", 6) == 0) {
544 h->port = bb_lookup_port("ftp", "tfp", 21);
548 bb_error_msg_and_die("not an http or ftp url: %s", url);
550 sp = strchr(h->host, '/');
555 h->path = bb_xstrdup("");
557 up = strrchr(h->host, '@');
567 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
568 if (h->host[0] == '[') {
572 while (*ep == ':' || isxdigit (*ep))
582 cp = strchr(pp, ':');
585 h->port = htons(atoi(cp));
590 FILE *open_socket(struct sockaddr_in *s_in)
594 fp = fdopen(xconnect(s_in), "r+");
596 bb_perror_msg_and_die("fdopen()");
602 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
609 /* retrieve header line */
610 if (fgets(buf, bufsiz, fp) == NULL)
613 /* see if we are at the end of the headers */
614 for (s = buf ; *s == '\r' ; ++s)
619 /* convert the header name to lower case */
620 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
623 /* verify we are at the end of the header name */
625 bb_error_msg_and_die("bad header line: %s", buf);
627 /* locate the start of the header value */
628 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
632 /* locate the end of header */
633 while (*s != '\0' && *s != '\r' && *s != '\n')
636 /* end of header found */
642 /* Rats! The buffer isn't big enough to hold the entire header value. */
643 while (c = getc(fp), c != EOF && c != '\n')
649 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
655 fprintf(fp, "%s%s\r\n", s1, s2);
660 p = fgets(buf, 510, fp);
662 bb_perror_msg_and_die("fgets()");
663 } while (! isdigit(buf[0]) || buf[3] != ' ');
668 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
669 /* Stuff below is from BSD rcp util.c, as added to openshh.
670 * Original copyright notice is retained at the end of this file.
679 get_terminal_width_height(0, &width, NULL);
684 updateprogressmeter(int ignore)
686 int save_errno = errno;
695 struct itimerval itv;
697 itv.it_value.tv_sec = wait;
698 itv.it_value.tv_usec = 0;
699 itv.it_interval = itv.it_value;
700 setitimer(ITIMER_REAL, &itv, NULL);
705 progressmeter(int flag)
707 static const char prefixes[] = " KMGTP";
708 static struct timeval lastupdate;
709 static off_t lastsize, totalsize;
710 struct timeval now, td, wait;
711 off_t cursize, abbrevsize;
713 int ratio, barlength, i, remaining;
717 (void) gettimeofday(&start, (struct timezone *) 0);
720 totalsize = filesize; /* as filesize changes.. */
723 (void) gettimeofday(&now, (struct timezone *) 0);
725 if (totalsize != 0 && !chunked) {
726 ratio = 100.0 * cursize / totalsize;
727 ratio = MAX(ratio, 0);
728 ratio = MIN(ratio, 100);
732 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
734 barlength = getttywidth() - 51;
736 i = barlength * ratio / 100;
737 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
739 "*****************************************************************************"
740 "*****************************************************************************",
744 abbrevsize = cursize;
745 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
749 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
750 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
753 timersub(&now, &lastupdate, &wait);
754 if (cursize > lastsize) {
757 if (wait.tv_sec >= STALLTIME) {
758 start.tv_sec += wait.tv_sec;
759 start.tv_usec += wait.tv_usec;
763 timersub(&now, &start, &td);
764 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
766 if (wait.tv_sec >= STALLTIME) {
767 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
769 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
770 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
773 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
774 i = remaining / 3600;
776 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
779 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
781 i = remaining % 3600;
782 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
783 "%02d:%02d ETA", i / 60, i % 60);
785 write(fileno(stderr), buf, strlen(buf));
789 sa.sa_handler = updateprogressmeter;
790 sigemptyset(&sa.sa_mask);
791 sa.sa_flags = SA_RESTART;
792 sigaction(SIGALRM, &sa, NULL);
794 } else if (flag == 1) {
802 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
803 * much of which was blatently stolen from openssh. */
806 * Copyright (c) 1992, 1993
807 * The Regents of the University of California. All rights reserved.
809 * Redistribution and use in source and binary forms, with or without
810 * modification, are permitted provided that the following conditions
812 * 1. Redistributions of source code must retain the above copyright
813 * notice, this list of conditions and the following disclaimer.
814 * 2. Redistributions in binary form must reproduce the above copyright
815 * notice, this list of conditions and the following disclaimer in the
816 * documentation and/or other materials provided with the distribution.
818 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
819 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
821 * 4. Neither the name of the University nor the names of its contributors
822 * may be used to endorse or promote products derived from this software
823 * without specific prior written permission.
825 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
826 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
827 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
828 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
829 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
830 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
831 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
832 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
833 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
834 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
837 * $Id: wget.c,v 1.65 2004/01/17 05:03:31 bug1 Exp $
844 c-file-style: "linux"