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) {
238 output = xfopen(fname_out, (do_continue ? "a" : "w"));
242 * Determine where to start transfer.
245 if (fstat(fileno(output), &sbuf) < 0)
246 perror_msg_and_die("fstat()");
247 if (sbuf.st_size > 0)
248 beg_range = sbuf.st_size;
253 if (proxy || !target.is_ftp) {
259 close_delete_and_die("too many redirections");
262 * Open socket to http server
264 if (sfp) fclose(sfp);
265 sfp = open_socket(server.host, server.port);
271 fprintf(sfp, "GET %stp://%s:%d/%s HTTP/1.1\r\n",
272 target.is_ftp ? "f" : "ht", target.host,
273 target.port, target.path);
275 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
278 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
280 #ifdef BB_FEATURE_WGET_AUTHENTICATION
282 fprintf(sfp, "Authorization: Basic %s\r\n",
283 base64enc(target.user, buf, sizeof(buf)));
285 if (proxy && server.user) {
286 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
287 base64enc(server.user, buf, sizeof(buf)));
292 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
293 fprintf(sfp,"Connection: close\r\n\r\n");
296 * Retrieve HTTP response line and check for "200" status code.
298 read_response: if (fgets(buf, sizeof(buf), sfp) == NULL)
299 close_delete_and_die("no response from server");
301 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
303 for ( ; isspace(*s) ; ++s)
305 switch (status = atoi(s)) {
308 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
311 if (do_continue && output != stdout)
312 output = freopen(fname_out, "w", output);
315 case 300: /* redirection */
326 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
330 * Retrieve HTTP headers.
332 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
333 if (strcasecmp(buf, "content-length") == 0) {
338 if (strcasecmp(buf, "transfer-encoding") == 0) {
339 if (strcasecmp(s, "chunked") == 0) {
340 chunked = got_clen = 1;
342 close_delete_and_die("server wants to do %s transfer encoding", s);
345 if (strcasecmp(buf, "location") == 0) {
347 target.path = xstrdup(s+1);
349 parse_url(xstrdup(s), &target);
351 server.host = target.host;
352 server.port = target.port;
357 } while(status >= 300);
367 target.user = xstrdup("anonymous:busybox@");
369 sfp = open_socket(server.host, server.port);
370 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
371 close_delete_and_die("%s", buf+4);
374 * Splitting username:password pair,
377 s = strchr(target.user, ':');
380 switch(ftpcmd("USER ", target.user, sfp, buf)) {
384 if (ftpcmd("PASS ", s, sfp, buf) == 230)
386 /* FALLTHRU (failed login) */
388 close_delete_and_die("ftp login: %s", buf+4);
391 ftpcmd("CDUP", NULL, sfp, buf);
392 ftpcmd("TYPE I", NULL, sfp, buf);
397 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
398 filesize = atol(buf+4);
403 * Entering passive mode
405 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
406 close_delete_and_die("PASV: %s", buf+4);
407 s = strrchr(buf, ',');
410 s = strrchr(buf, ',');
411 port += atoi(s+1) * 256;
412 dfp = open_socket(server.host, port);
415 sprintf(buf, "REST %ld", beg_range);
416 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
417 if (output != stdout)
418 output = freopen(fname_out, "w", output);
421 filesize -= beg_range;
424 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
425 close_delete_and_die("RETR: %s", buf+4);
434 fgets(buf, sizeof(buf), dfp);
435 filesize = strtol(buf, (char **) NULL, 16);
437 #ifdef BB_FEATURE_WGET_STATUSBAR
438 if (quiet_flag==FALSE)
442 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, chunked ? (filesize > sizeof(buf) ? sizeof(buf) : filesize) : sizeof(buf), dfp)) > 0) {
443 safe_fwrite(buf, 1, n, output);
444 #ifdef BB_FEATURE_WGET_STATUSBAR
452 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
453 safe_fgets(buf, sizeof(buf), dfp);
454 filesize = strtol(buf, (char **) NULL, 16);
455 if (filesize==0) chunked = 0; /* all done! */
458 if (n == 0 && ferror(dfp))
459 perror_msg_and_die("network read error");
461 #ifdef BB_FEATURE_WGET_STATUSBAR
462 if (quiet_flag==FALSE)
465 if (!proxy && target.is_ftp) {
467 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
468 error_msg_and_die("ftp error: %s", buf+4);
469 ftpcmd("QUIT", NULL, sfp, buf);
475 void parse_url(char *url, struct host_info *h)
479 if (strncmp(url, "http://", 7) == 0) {
483 } else if (strncmp(url, "ftp://", 6) == 0) {
488 error_msg_and_die("not an http or ftp url: %s", url);
490 sp = strchr(h->host, '/');
497 up = strrchr(h->host, '@');
505 cp = strchr(h->host, ':');
514 FILE *open_socket(char *host, int port)
516 struct sockaddr_in s_in;
521 memset(&s_in, 0, sizeof(s_in));
522 s_in.sin_family = AF_INET;
523 if ((hp = (struct hostent *) gethostbyname(host)) == NULL)
524 error_msg_and_die("cannot resolve %s", host);
525 memcpy(&s_in.sin_addr, hp->h_addr_list[0], hp->h_length);
526 s_in.sin_port = htons(port);
529 * Get the server onto a stdio stream.
531 if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0)
532 perror_msg_and_die("socket()");
533 if (connect(fd, (struct sockaddr *) &s_in, sizeof(s_in)) < 0)
534 perror_msg_and_die("connect(%s)", host);
535 if ((fp = fdopen(fd, "r+")) == NULL)
536 perror_msg_and_die("fdopen()");
542 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
549 /* retrieve header line */
550 if (fgets(buf, bufsiz, fp) == NULL)
553 /* see if we are at the end of the headers */
554 for (s = buf ; *s == '\r' ; ++s)
559 /* convert the header name to lower case */
560 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
563 /* verify we are at the end of the header name */
565 error_msg_and_die("bad header line: %s", buf);
567 /* locate the start of the header value */
568 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
572 /* locate the end of header */
573 while (*s != '\0' && *s != '\r' && *s != '\n')
576 /* end of header found */
582 /* Rats! The buffer isn't big enough to hold the entire header value. */
583 while (c = getc(fp), c != EOF && c != '\n')
589 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
595 fprintf(fp, "%s%s\n", s1, s2);
600 p = fgets(buf, 510, fp);
602 perror_msg_and_die("fgets()");
603 } while (! isdigit(buf[0]) || buf[3] != ' ');
608 #ifdef BB_FEATURE_WGET_STATUSBAR
609 /* Stuff below is from BSD rcp util.c, as added to openshh.
610 * Original copyright notice is retained at the end of this file.
618 struct winsize winsize;
620 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
621 return (winsize.ws_col ? winsize.ws_col : 80);
627 updateprogressmeter(int ignore)
629 int save_errno = errno;
638 struct itimerval itv;
640 itv.it_value.tv_sec = wait;
641 itv.it_value.tv_usec = 0;
642 itv.it_interval = itv.it_value;
643 setitimer(ITIMER_REAL, &itv, NULL);
648 progressmeter(int flag)
650 static const char prefixes[] = " KMGTP";
651 static struct timeval lastupdate;
652 static off_t lastsize, totalsize;
653 struct timeval now, td, wait;
654 off_t cursize, abbrevsize;
656 int ratio, barlength, i, remaining;
660 (void) gettimeofday(&start, (struct timezone *) 0);
663 totalsize = filesize; /* as filesize changes.. */
666 (void) gettimeofday(&now, (struct timezone *) 0);
668 if (totalsize != 0 && !chunked) {
669 ratio = 100.0 * cursize / totalsize;
670 ratio = MAX(ratio, 0);
671 ratio = MIN(ratio, 100);
675 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
677 barlength = getttywidth() - 51;
679 i = barlength * ratio / 100;
680 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
682 "*****************************************************************************"
683 "*****************************************************************************",
687 abbrevsize = cursize;
688 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
692 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
693 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
696 timersub(&now, &lastupdate, &wait);
697 if (cursize > lastsize) {
700 if (wait.tv_sec >= STALLTIME) {
701 start.tv_sec += wait.tv_sec;
702 start.tv_usec += wait.tv_usec;
706 timersub(&now, &start, &td);
707 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
709 if (wait.tv_sec >= STALLTIME) {
710 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
712 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
713 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
716 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
717 i = remaining / 3600;
719 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
722 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
724 i = remaining % 3600;
725 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
726 "%02d:%02d ETA", i / 60, i % 60);
728 write(fileno(stderr), buf, strlen(buf));
732 sa.sa_handler = updateprogressmeter;
733 sigemptyset(&sa.sa_mask);
734 sa.sa_flags = SA_RESTART;
735 sigaction(SIGALRM, &sa, NULL);
737 } else if (flag == 1) {
745 /* Original copyright notice which applies to the BB_FEATURE_WGET_STATUSBAR stuff,
746 * much of which was blatently stolen from openssh. */
749 * Copyright (c) 1992, 1993
750 * The Regents of the University of California. All rights reserved.
752 * Redistribution and use in source and binary forms, with or without
753 * modification, are permitted provided that the following conditions
755 * 1. Redistributions of source code must retain the above copyright
756 * notice, this list of conditions and the following disclaimer.
757 * 2. Redistributions in binary form must reproduce the above copyright
758 * notice, this list of conditions and the following disclaimer in the
759 * documentation and/or other materials provided with the distribution.
761 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
762 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
764 * 4. Neither the name of the University nor the names of its contributors
765 * may be used to endorse or promote products derived from this software
766 * without specific prior written permission.
768 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
769 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
770 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
771 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
772 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
773 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
774 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
775 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
776 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
777 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
780 * $Id: wget.c,v 1.37 2001/05/09 19:15:46 kraai Exp $
787 c-file-style: "linux"