1 /* vi: set sw=4 ts=4: */
3 * wget - retrieve a file using HTTP or FTP
5 * Chip Rosenthal Covad Communications <chip@laserlink.net>
7 * Note: According to RFC2616 section 3.6.1, "All HTTP/1.1 applications MUST be
8 * able to receive and decode the "chunked" transfer-coding, and MUST ignore
9 * chunk-extension extensions they do not understand."
11 * This prevents this particular wget app from completely RFC compliant, and as
12 * such, prevents it from being used as a general purpose web browser... This
13 * is a design decision, since it makes the code smaller.
25 #include <sys/ioctl.h>
28 #include <sys/types.h>
30 #include <sys/socket.h>
31 #include <netinet/in.h>
32 #include <arpa/inet.h>
37 /* Stupid libc5 doesn't define this... */
39 #define timersub(a, b, result) \
41 (result)->tv_sec = (a)->tv_sec - (b)->tv_sec; \
42 (result)->tv_usec = (a)->tv_usec - (b)->tv_usec; \
43 if ((result)->tv_usec < 0) { \
45 (result)->tv_usec += 1000000; \
58 static void parse_url(char *url, struct host_info *h);
59 static FILE *open_socket(char *host, int port);
60 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
61 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf);
62 static void progressmeter(int flag);
64 /* Globals (can be accessed from signal handlers */
65 static off_t filesize = 0; /* content-length of the file */
66 #ifdef BB_FEATURE_WGET_STATUSBAR
67 static char *curfile; /* Name of current file being transferred. */
68 static struct timeval start; /* Time a transfer started. */
69 static volatile unsigned long statbytes; /* Number of bytes transferred so far. */
70 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
71 static const int STALLTIME = 5;
74 static void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
76 if (output != stdout && do_continue==0) {
82 #define close_delete_and_die(s...) { \
83 close_and_delete_outfile(output, fname_out, do_continue); \
84 error_msg_and_die(s); }
87 #ifdef BB_FEATURE_WGET_AUTHENTICATION
89 * Base64-encode character string
90 * oops... isn't something similar in uuencode.c?
91 * It would be better to use already existing code
93 char *base64enc(char *p, char *buf, int len) {
95 char al[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
101 error_msg_and_die("buffer overflow");
102 *(s++) = al[(*p >> 2) & 0x3F];
103 *(s++) = al[((*p << 4) & 0x30) | ((*(p+1) >> 4) & 0x0F)];
107 *(s++) = al[((*p << 2) & 0x3C) | ((*(p+1) >> 6) & 0x03)];
109 *(s++) = al[*(p++) & 0x3F];
116 int wget_main(int argc, char **argv)
118 int n, try=5, status;
124 struct host_info server, target;
126 FILE *sfp = NULL; /* socket to web/ftp server */
127 FILE *dfp = NULL; /* socket to ftp server (data) */
128 char *fname_out = NULL; /* where to direct output (-O) */
129 int do_continue = 0; /* continue a prev transfer (-c) */
130 long beg_range = 0L; /* range at which continue begins */
131 int got_clen = 0; /* got content-length: from server */
132 FILE *output; /* socket to web server */
133 int quiet_flag = FALSE; /* Be verry, verry quiet... */
136 * Crack command line.
138 while ((n = getopt(argc, argv, "cqO:")) != EOF) {
147 /* can't set fname_out to NULL if outputting to stdout, because
148 * this gets interpreted as the auto-gen output filename
149 * case below - tausq@debian.org
151 fname_out = (strcmp(optarg, "-") == 0 ? (char *)1 : optarg);
158 if (argc - optind != 1)
161 parse_url(argv[optind], &target);
162 server.host = target.host;
163 server.port = target.port;
166 * Use the proxy if necessary.
168 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
170 parse_url(xstrdup(proxy), &server);
172 /* Guess an output filename */
175 #ifdef BB_FEATURE_WGET_STATUSBAR
178 get_last_path_component(target.path);
179 if (fname_out==NULL || strlen(fname_out)<1) {
181 #ifdef BB_FEATURE_WGET_STATUSBAR
186 #ifdef BB_FEATURE_WGET_STATUSBAR
188 curfile = get_last_path_component(fname_out);
191 if (do_continue && !fname_out)
192 error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
196 * Open the output file stream.
198 if (fname_out != (char *)1) {
199 output = xfopen( fname_out, (do_continue ? "a" : "w") );
205 * Determine where to start transfer.
208 if (fstat(fileno(output), &sbuf) < 0)
209 perror_msg_and_die("fstat()");
210 if (sbuf.st_size > 0)
211 beg_range = sbuf.st_size;
216 if (proxy || !target.is_ftp) {
222 close_delete_and_die("too many redirections");
225 * Open socket to http server
227 if (sfp) fclose(sfp);
228 sfp = open_socket(server.host, server.port);
234 fprintf(sfp, "GET %stp://%s:%d/%s HTTP/1.0\r\n",
235 target.is_ftp ? "f" : "ht", target.host,
236 target.port, target.path);
238 fprintf(sfp, "GET /%s HTTP/1.0\r\n", target.path);
241 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
243 #ifdef BB_FEATURE_WGET_AUTHENTICATION
245 fprintf(sfp, "Authorization: Basic %s\r\n",
246 base64enc(target.user, buf, sizeof(buf)));
248 if (proxy && server.user) {
249 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
250 base64enc(server.user, buf, sizeof(buf)));
255 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
256 fprintf(sfp,"Connection: close\r\n\r\n");
259 * Retrieve HTTP response line and check for "200" status code.
261 if (fgets(buf, sizeof(buf), sfp) == NULL)
262 close_delete_and_die("no response from server");
264 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
266 for ( ; isspace(*s) ; ++s)
268 switch (status = atoi(s)) {
271 if (do_continue && output != stdout)
272 output = freopen(fname_out, "w", output);
275 case 300: /* redirection */
286 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
290 * Retrieve HTTP headers.
292 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
293 if (strcasecmp(buf, "content-length") == 0) {
298 if (strcasecmp(buf, "transfer-encoding") == 0)
299 close_delete_and_die("server wants to do %s transfer encoding", s);
301 if (strcasecmp(buf, "location") == 0) {
303 target.path = xstrdup(s+1);
305 parse_url(xstrdup(s), &target);
307 server.host = target.host;
308 server.port = target.port;
313 } while(status >= 300);
323 target.user = xstrdup("anonymous:busybox@");
325 sfp = open_socket(server.host, server.port);
326 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
327 close_delete_and_die("%s", buf+4);
330 * Splitting username:password pair,
333 s = strchr(target.user, ':');
336 switch(ftpcmd("USER ", target.user, sfp, buf)) {
340 if (ftpcmd("PASS ", s, sfp, buf) == 230)
342 /* FALLTHRU (failed login) */
344 close_delete_and_die("ftp login: %s", buf+4);
347 ftpcmd("CDUP", NULL, sfp, buf);
348 ftpcmd("TYPE I", NULL, sfp, buf);
353 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
354 filesize = atol(buf+4);
359 * Entering passive mode
361 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
362 close_delete_and_die("PASV: %s", buf+4);
363 s = strrchr(buf, ',');
366 s = strrchr(buf, ',');
367 port += atoi(s+1) * 256;
368 dfp = open_socket(server.host, port);
371 sprintf(buf, "REST %ld", beg_range);
372 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
373 if (output != stdout)
374 output = freopen(fname_out, "w", output);
377 filesize -= beg_range;
380 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
381 close_delete_and_die("RETR: %s", buf+4);
389 #ifdef BB_FEATURE_WGET_STATUSBAR
391 if (quiet_flag==FALSE)
394 while ((filesize > 0 || !got_clen) && (n = fread(buf, 1, sizeof(buf), dfp)) > 0) {
395 fwrite(buf, 1, n, output);
396 #ifdef BB_FEATURE_WGET_STATUSBAR
398 if (quiet_flag==FALSE)
405 if (n == 0 && ferror(dfp))
406 perror_msg_and_die("network read error");
408 if (!proxy && target.is_ftp) {
410 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
411 error_msg_and_die("ftp error: %s", buf+4);
412 ftpcmd("QUIT", NULL, sfp, buf);
419 void parse_url(char *url, struct host_info *h)
423 if (strncmp(url, "http://", 7) == 0) {
427 } else if (strncmp(url, "ftp://", 6) == 0) {
432 error_msg_and_die("not an http or ftp url: %s", url);
434 sp = strchr(h->host, '/');
441 up = strrchr(h->host, '@');
449 cp = strchr(h->host, ':');
458 FILE *open_socket(char *host, int port)
460 struct sockaddr_in s_in;
465 memset(&s_in, 0, sizeof(s_in));
466 s_in.sin_family = AF_INET;
467 if ((hp = (struct hostent *) gethostbyname(host)) == NULL)
468 error_msg_and_die("cannot resolve %s", host);
469 memcpy(&s_in.sin_addr, hp->h_addr_list[0], hp->h_length);
470 s_in.sin_port = htons(port);
473 * Get the server onto a stdio stream.
475 if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0)
476 perror_msg_and_die("socket()");
477 if (connect(fd, (struct sockaddr *) &s_in, sizeof(s_in)) < 0)
478 perror_msg_and_die("connect(%s)", host);
479 if ((fp = fdopen(fd, "r+")) == NULL)
480 perror_msg_and_die("fdopen()");
486 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
493 /* retrieve header line */
494 if (fgets(buf, bufsiz, fp) == NULL)
497 /* see if we are at the end of the headers */
498 for (s = buf ; *s == '\r' ; ++s)
503 /* convert the header name to lower case */
504 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
507 /* verify we are at the end of the header name */
509 error_msg_and_die("bad header line: %s", buf);
511 /* locate the start of the header value */
512 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
516 /* locate the end of header */
517 while (*s != '\0' && *s != '\r' && *s != '\n')
520 /* end of header found */
526 /* Rats! The buffer isn't big enough to hold the entire header value. */
527 while (c = getc(fp), c != EOF && c != '\n')
533 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
539 fprintf(fp, "%s%s\n", s1, s2);
544 p = fgets(buf, 510, fp);
546 perror_msg_and_die("fgets()");
547 } while (! isdigit(buf[0]) || buf[3] != ' ');
552 #ifdef BB_FEATURE_WGET_STATUSBAR
553 /* Stuff below is from BSD rcp util.c, as added to openshh.
554 * Original copyright notice is retained at the end of this file.
562 struct winsize winsize;
564 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
565 return (winsize.ws_col ? winsize.ws_col : 80);
571 updateprogressmeter(int ignore)
573 int save_errno = errno;
582 struct itimerval itv;
584 itv.it_value.tv_sec = wait;
585 itv.it_value.tv_usec = 0;
586 itv.it_interval = itv.it_value;
587 setitimer(ITIMER_REAL, &itv, NULL);
592 progressmeter(int flag)
594 static const char prefixes[] = " KMGTP";
595 static struct timeval lastupdate;
596 static off_t lastsize;
597 struct timeval now, td, wait;
598 off_t cursize, abbrevsize;
600 int ratio, barlength, i, remaining;
604 (void) gettimeofday(&start, (struct timezone *) 0);
609 (void) gettimeofday(&now, (struct timezone *) 0);
612 ratio = 100.0 * cursize / filesize;
613 ratio = MAX(ratio, 0);
614 ratio = MIN(ratio, 100);
618 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
620 barlength = getttywidth() - 51;
622 i = barlength * ratio / 100;
623 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
625 "*****************************************************************************"
626 "*****************************************************************************",
630 abbrevsize = cursize;
631 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
635 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
636 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
639 timersub(&now, &lastupdate, &wait);
640 if (cursize > lastsize) {
643 if (wait.tv_sec >= STALLTIME) {
644 start.tv_sec += wait.tv_sec;
645 start.tv_usec += wait.tv_usec;
649 timersub(&now, &start, &td);
650 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
652 if (statbytes <= 0 || elapsed <= 0.0 || cursize > filesize) {
653 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
655 } else if (wait.tv_sec >= STALLTIME) {
656 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
659 remaining = (int) (filesize / (statbytes / elapsed) - elapsed);
660 i = remaining / 3600;
662 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
665 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
667 i = remaining % 3600;
668 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
669 "%02d:%02d ETA", i / 60, i % 60);
671 write(fileno(stderr), buf, strlen(buf));
675 sa.sa_handler = updateprogressmeter;
676 sigemptyset(&sa.sa_mask);
677 sa.sa_flags = SA_RESTART;
678 sigaction(SIGALRM, &sa, NULL);
680 } else if (flag == 1) {
687 /* Original copyright notice which applies to the BB_FEATURE_WGET_STATUSBAR stuff,
688 * much of which was blatently stolen from openssh. */
691 * Copyright (c) 1992, 1993
692 * The Regents of the University of California. All rights reserved.
694 * Redistribution and use in source and binary forms, with or without
695 * modification, are permitted provided that the following conditions
697 * 1. Redistributions of source code must retain the above copyright
698 * notice, this list of conditions and the following disclaimer.
699 * 2. Redistributions in binary form must reproduce the above copyright
700 * notice, this list of conditions and the following disclaimer in the
701 * documentation and/or other materials provided with the distribution.
703 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
704 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
706 * 4. Neither the name of the University nor the names of its contributors
707 * may be used to endorse or promote products derived from this software
708 * without specific prior written permission.
710 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
711 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
712 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
713 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
714 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
715 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
716 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
717 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
718 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
719 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
722 * $Id: wget.c,v 1.31 2001/04/05 21:45:53 andersen Exp $
729 c-file-style: "linux"