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>
29 /* Stupid libc5 doesn't define this... */
31 #define timersub(a, b, result) \
33 (result)->tv_sec = (a)->tv_sec - (b)->tv_sec; \
34 (result)->tv_usec = (a)->tv_usec - (b)->tv_usec; \
35 if ((result)->tv_usec < 0) { \
37 (result)->tv_usec += 1000000; \
50 static void parse_url(char *url, struct host_info *h);
51 static FILE *open_socket(char *host, int port);
52 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
53 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf);
54 static void progressmeter(int flag);
56 /* Globals (can be accessed from signal handlers */
57 static off_t filesize = 0; /* content-length of the file */
58 static int chunked = 0; /* chunked transfer encoding */
59 #ifdef BB_FEATURE_WGET_STATUSBAR
60 static char *curfile; /* Name of current file being transferred. */
61 static struct timeval start; /* Time a transfer started. */
62 static volatile unsigned long statbytes = 0; /* Number of bytes transferred so far. */
63 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
64 static const int STALLTIME = 5;
67 static void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
69 if (output != stdout && do_continue==0) {
75 /* Read NMEMB elements of SIZE bytes into PTR from STREAM. Returns the
76 * number of elements read, and a short count if an eof or non-interrupt
77 * error is encountered. */
78 static size_t safe_fread(void *ptr, size_t size, size_t nmemb, FILE *stream)
84 ret += fread((char *)ptr + (ret * size), size, nmemb - ret, stream);
85 } while (ret < nmemb && ferror(stream) && errno == EINTR);
90 /* Write NMEMB elements of SIZE bytes from PTR to STREAM. Returns the
91 * number of elements written, and a short count if an eof or non-interrupt
92 * error is encountered. */
93 static size_t safe_fwrite(void *ptr, size_t size, size_t nmemb, FILE *stream)
99 ret += fwrite((char *)ptr + (ret * size), size, nmemb - ret, stream);
100 } while (ret < nmemb && ferror(stream) && errno == EINTR);
105 /* Read a line or SIZE - 1 bytes into S, whichever is less, from STREAM.
106 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
107 static char *safe_fgets(char *s, int size, FILE *stream)
113 ret = fgets(s, size, stream);
114 } while (ret == NULL && ferror(stream) && errno == EINTR);
119 #define close_delete_and_die(s...) { \
120 close_and_delete_outfile(output, fname_out, do_continue); \
121 error_msg_and_die(s); }
124 #ifdef BB_FEATURE_WGET_AUTHENTICATION
126 * Base64-encode character string
127 * oops... isn't something similar in uuencode.c?
128 * It would be better to use already existing code
130 char *base64enc(char *p, char *buf, int len) {
132 char al[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
138 error_msg_and_die("buffer overflow");
139 *(s++) = al[(*p >> 2) & 0x3F];
140 *(s++) = al[((*p << 4) & 0x30) | ((*(p+1) >> 4) & 0x0F)];
144 *(s++) = al[((*p << 2) & 0x3C) | ((*(p+1) >> 6) & 0x03)];
146 *(s++) = al[*(p++) & 0x3F];
153 int wget_main(int argc, char **argv)
155 int n, try=5, status;
161 struct host_info server, target;
163 FILE *sfp = NULL; /* socket to web/ftp server */
164 FILE *dfp = NULL; /* socket to ftp server (data) */
165 char *fname_out = NULL; /* where to direct output (-O) */
166 int do_continue = 0; /* continue a prev transfer (-c) */
167 long beg_range = 0L; /* range at which continue begins */
168 int got_clen = 0; /* got content-length: from server */
169 FILE *output; /* socket to web server */
170 int quiet_flag = FALSE; /* Be verry, verry quiet... */
173 * Crack command line.
175 while ((n = getopt(argc, argv, "cqO:")) != EOF) {
184 /* can't set fname_out to NULL if outputting to stdout, because
185 * this gets interpreted as the auto-gen output filename
186 * case below - tausq@debian.org
195 if (argc - optind != 1)
198 parse_url(argv[optind], &target);
199 server.host = target.host;
200 server.port = target.port;
203 * Use the proxy if necessary.
205 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
207 parse_url(xstrdup(proxy), &server);
209 /* Guess an output filename */
212 #ifdef BB_FEATURE_WGET_STATUSBAR
215 get_last_path_component(target.path);
216 if (fname_out==NULL || strlen(fname_out)<1) {
218 #ifdef BB_FEATURE_WGET_STATUSBAR
223 #ifdef BB_FEATURE_WGET_STATUSBAR
225 curfile = get_last_path_component(fname_out);
228 if (do_continue && !fname_out)
229 error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
233 * Open the output file stream.
235 if (strcmp(fname_out, "-") == 0) {
239 output = xfopen(fname_out, (do_continue ? "a" : "w"));
243 * Determine where to start transfer.
246 if (fstat(fileno(output), &sbuf) < 0)
247 perror_msg_and_die("fstat()");
248 if (sbuf.st_size > 0)
249 beg_range = sbuf.st_size;
254 if (proxy || !target.is_ftp) {
260 close_delete_and_die("too many redirections");
263 * Open socket to http server
265 if (sfp) fclose(sfp);
266 sfp = open_socket(server.host, server.port);
272 fprintf(sfp, "GET %stp://%s:%d/%s HTTP/1.1\r\n",
273 target.is_ftp ? "f" : "ht", target.host,
274 target.port, target.path);
276 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
279 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
281 #ifdef BB_FEATURE_WGET_AUTHENTICATION
283 fprintf(sfp, "Authorization: Basic %s\r\n",
284 base64enc(target.user, buf, sizeof(buf)));
286 if (proxy && server.user) {
287 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
288 base64enc(server.user, buf, sizeof(buf)));
293 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
294 fprintf(sfp,"Connection: close\r\n\r\n");
297 * Retrieve HTTP response line and check for "200" status code.
299 read_response: if (fgets(buf, sizeof(buf), sfp) == NULL)
300 close_delete_and_die("no response from server");
302 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
304 for ( ; isspace(*s) ; ++s)
306 switch (status = atoi(s)) {
309 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
312 if (do_continue && output != stdout)
313 output = freopen(fname_out, "w", output);
316 case 300: /* redirection */
327 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
331 * Retrieve HTTP headers.
333 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
334 if (strcasecmp(buf, "content-length") == 0) {
339 if (strcasecmp(buf, "transfer-encoding") == 0) {
340 if (strcasecmp(s, "chunked") == 0) {
341 chunked = got_clen = 1;
343 close_delete_and_die("server wants to do %s transfer encoding", s);
346 if (strcasecmp(buf, "location") == 0) {
348 target.path = xstrdup(s+1);
350 parse_url(xstrdup(s), &target);
352 server.host = target.host;
353 server.port = target.port;
358 } while(status >= 300);
368 target.user = xstrdup("anonymous:busybox@");
370 sfp = open_socket(server.host, server.port);
371 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
372 close_delete_and_die("%s", buf+4);
375 * Splitting username:password pair,
378 s = strchr(target.user, ':');
381 switch(ftpcmd("USER ", target.user, sfp, buf)) {
385 if (ftpcmd("PASS ", s, sfp, buf) == 230)
387 /* FALLTHRU (failed login) */
389 close_delete_and_die("ftp login: %s", buf+4);
392 ftpcmd("CDUP", NULL, sfp, buf);
393 ftpcmd("TYPE I", NULL, sfp, buf);
398 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
399 filesize = atol(buf+4);
404 * Entering passive mode
406 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
407 close_delete_and_die("PASV: %s", buf+4);
408 s = strrchr(buf, ',');
411 s = strrchr(buf, ',');
412 port += atoi(s+1) * 256;
413 dfp = open_socket(server.host, port);
416 sprintf(buf, "REST %ld", beg_range);
417 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
418 if (output != stdout)
419 output = freopen(fname_out, "w", output);
422 filesize -= beg_range;
425 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
426 close_delete_and_die("RETR: %s", buf+4);
435 fgets(buf, sizeof(buf), dfp);
436 filesize = strtol(buf, (char **) NULL, 16);
438 #ifdef BB_FEATURE_WGET_STATUSBAR
439 if (quiet_flag==FALSE)
443 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, chunked ? (filesize > sizeof(buf) ? sizeof(buf) : filesize) : sizeof(buf), dfp)) > 0) {
444 safe_fwrite(buf, 1, n, output);
445 #ifdef BB_FEATURE_WGET_STATUSBAR
453 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
454 safe_fgets(buf, sizeof(buf), dfp);
455 filesize = strtol(buf, (char **) NULL, 16);
456 if (filesize==0) chunked = 0; /* all done! */
459 if (n == 0 && ferror(dfp))
460 perror_msg_and_die("network read error");
462 #ifdef BB_FEATURE_WGET_STATUSBAR
463 if (quiet_flag==FALSE)
466 if (!proxy && target.is_ftp) {
468 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
469 error_msg_and_die("ftp error: %s", buf+4);
470 ftpcmd("QUIT", NULL, sfp, buf);
476 void parse_url(char *url, struct host_info *h)
480 if (strncmp(url, "http://", 7) == 0) {
484 } else if (strncmp(url, "ftp://", 6) == 0) {
489 error_msg_and_die("not an http or ftp url: %s", url);
491 sp = strchr(h->host, '/');
498 up = strrchr(h->host, '@');
506 cp = strchr(h->host, ':');
515 FILE *open_socket(char *host, int port)
517 struct sockaddr_in s_in;
522 memset(&s_in, 0, sizeof(s_in));
523 s_in.sin_family = AF_INET;
524 if ((hp = (struct hostent *) gethostbyname(host)) == NULL)
525 error_msg_and_die("cannot resolve %s", host);
526 memcpy(&s_in.sin_addr, hp->h_addr_list[0], hp->h_length);
527 s_in.sin_port = htons(port);
530 * Get the server onto a stdio stream.
532 if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0)
533 perror_msg_and_die("socket()");
534 if (connect(fd, (struct sockaddr *) &s_in, sizeof(s_in)) < 0)
535 perror_msg_and_die("connect(%s)", host);
536 if ((fp = fdopen(fd, "r+")) == NULL)
537 perror_msg_and_die("fdopen()");
543 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
550 /* retrieve header line */
551 if (fgets(buf, bufsiz, fp) == NULL)
554 /* see if we are at the end of the headers */
555 for (s = buf ; *s == '\r' ; ++s)
560 /* convert the header name to lower case */
561 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
564 /* verify we are at the end of the header name */
566 error_msg_and_die("bad header line: %s", buf);
568 /* locate the start of the header value */
569 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
573 /* locate the end of header */
574 while (*s != '\0' && *s != '\r' && *s != '\n')
577 /* end of header found */
583 /* Rats! The buffer isn't big enough to hold the entire header value. */
584 while (c = getc(fp), c != EOF && c != '\n')
590 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
596 fprintf(fp, "%s%s\n", s1, s2);
601 p = fgets(buf, 510, fp);
603 perror_msg_and_die("fgets()");
604 } while (! isdigit(buf[0]) || buf[3] != ' ');
609 #ifdef BB_FEATURE_WGET_STATUSBAR
610 /* Stuff below is from BSD rcp util.c, as added to openshh.
611 * Original copyright notice is retained at the end of this file.
619 struct winsize winsize;
621 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
622 return (winsize.ws_col ? winsize.ws_col : 80);
628 updateprogressmeter(int ignore)
630 int save_errno = errno;
639 struct itimerval itv;
641 itv.it_value.tv_sec = wait;
642 itv.it_value.tv_usec = 0;
643 itv.it_interval = itv.it_value;
644 setitimer(ITIMER_REAL, &itv, NULL);
649 progressmeter(int flag)
651 static const char prefixes[] = " KMGTP";
652 static struct timeval lastupdate;
653 static off_t lastsize, totalsize;
654 struct timeval now, td, wait;
655 off_t cursize, abbrevsize;
657 int ratio, barlength, i, remaining;
661 (void) gettimeofday(&start, (struct timezone *) 0);
664 totalsize = filesize; /* as filesize changes.. */
667 (void) gettimeofday(&now, (struct timezone *) 0);
669 if (totalsize != 0 && !chunked) {
670 ratio = 100.0 * cursize / totalsize;
671 ratio = MAX(ratio, 0);
672 ratio = MIN(ratio, 100);
676 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
678 barlength = getttywidth() - 51;
680 i = barlength * ratio / 100;
681 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
683 "*****************************************************************************"
684 "*****************************************************************************",
688 abbrevsize = cursize;
689 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
693 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
694 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
697 timersub(&now, &lastupdate, &wait);
698 if (cursize > lastsize) {
701 if (wait.tv_sec >= STALLTIME) {
702 start.tv_sec += wait.tv_sec;
703 start.tv_usec += wait.tv_usec;
707 timersub(&now, &start, &td);
708 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
710 if (wait.tv_sec >= STALLTIME) {
711 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
713 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
714 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
717 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
718 i = remaining / 3600;
720 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
723 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
725 i = remaining % 3600;
726 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
727 "%02d:%02d ETA", i / 60, i % 60);
729 write(fileno(stderr), buf, strlen(buf));
733 sa.sa_handler = updateprogressmeter;
734 sigemptyset(&sa.sa_mask);
735 sa.sa_flags = SA_RESTART;
736 sigaction(SIGALRM, &sa, NULL);
738 } else if (flag == 1) {
746 /* Original copyright notice which applies to the BB_FEATURE_WGET_STATUSBAR stuff,
747 * much of which was blatently stolen from openssh. */
750 * Copyright (c) 1992, 1993
751 * The Regents of the University of California. All rights reserved.
753 * Redistribution and use in source and binary forms, with or without
754 * modification, are permitted provided that the following conditions
756 * 1. Redistributions of source code must retain the above copyright
757 * notice, this list of conditions and the following disclaimer.
758 * 2. Redistributions in binary form must reproduce the above copyright
759 * notice, this list of conditions and the following disclaimer in the
760 * documentation and/or other materials provided with the distribution.
762 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
763 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
765 * 4. Neither the name of the University nor the names of its contributors
766 * may be used to endorse or promote products derived from this software
767 * without specific prior written permission.
769 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
770 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
771 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
772 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
773 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
774 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
775 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
776 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
777 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
778 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
781 * $Id: wget.c,v 1.38 2001/05/13 00:55:54 andersen Exp $
788 c-file-style: "linux"