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>
34 /* Stupid libc5 doesn't define this... */
36 #define timersub(a, b, result) \
38 (result)->tv_sec = (a)->tv_sec - (b)->tv_sec; \
39 (result)->tv_usec = (a)->tv_usec - (b)->tv_usec; \
40 if ((result)->tv_usec < 0) { \
42 (result)->tv_usec += 1000000; \
55 static void parse_url(char *url, struct host_info *h);
56 static FILE *open_socket(char *host, int port);
57 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
58 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf);
59 static void progressmeter(int flag);
61 /* Globals (can be accessed from signal handlers */
62 static off_t filesize = 0; /* content-length of the file */
63 static int chunked = 0; /* chunked transfer encoding */
64 #ifdef BB_FEATURE_WGET_STATUSBAR
65 static char *curfile; /* Name of current file being transferred. */
66 static struct timeval start; /* Time a transfer started. */
67 static volatile unsigned long statbytes = 0; /* Number of bytes transferred so far. */
68 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
69 static const int STALLTIME = 5;
72 static void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
74 if (output != stdout && do_continue==0) {
80 /* Read NMEMB elements of SIZE bytes into PTR from STREAM. Returns the
81 * number of elements read, and a short count if an eof or non-interrupt
82 * error is encountered. */
83 static size_t safe_fread(void *ptr, size_t size, size_t nmemb, FILE *stream)
89 ret += fread((char *)ptr + (ret * size), size, nmemb - ret, stream);
90 } while (ret < nmemb && ferror(stream) && errno == EINTR);
95 /* Write NMEMB elements of SIZE bytes from PTR to STREAM. Returns the
96 * number of elements written, and a short count if an eof or non-interrupt
97 * error is encountered. */
98 static size_t safe_fwrite(void *ptr, size_t size, size_t nmemb, FILE *stream)
104 ret += fwrite((char *)ptr + (ret * size), size, nmemb - ret, stream);
105 } while (ret < nmemb && ferror(stream) && errno == EINTR);
110 /* Read a line or SIZE - 1 bytes into S, whichever is less, from STREAM.
111 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
112 static char *safe_fgets(char *s, int size, FILE *stream)
118 ret = fgets(s, size, stream);
119 } while (ret == NULL && ferror(stream) && errno == EINTR);
124 #define close_delete_and_die(s...) { \
125 close_and_delete_outfile(output, fname_out, do_continue); \
126 error_msg_and_die(s); }
129 #ifdef BB_FEATURE_WGET_AUTHENTICATION
131 * Base64-encode character string
132 * oops... isn't something similar in uuencode.c?
133 * It would be better to use already existing code
135 char *base64enc(char *p, char *buf, int len) {
137 char al[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
143 error_msg_and_die("buffer overflow");
144 *(s++) = al[(*p >> 2) & 0x3F];
145 *(s++) = al[((*p << 4) & 0x30) | ((*(p+1) >> 4) & 0x0F)];
149 *(s++) = al[((*p << 2) & 0x3C) | ((*(p+1) >> 6) & 0x03)];
151 *(s++) = al[*(p++) & 0x3F];
158 int wget_main(int argc, char **argv)
160 int n, try=5, status;
165 char extra_headers[1024];
166 char *extra_headers_ptr = extra_headers;
167 int extra_headers_left = sizeof(extra_headers);
168 int which_long_opt = 0, option_index = -1;
169 struct host_info server, target;
171 FILE *sfp = NULL; /* socket to web/ftp server */
172 FILE *dfp = NULL; /* socket to ftp server (data) */
173 char *fname_out = NULL; /* where to direct output (-O) */
174 int do_continue = 0; /* continue a prev transfer (-c) */
175 long beg_range = 0L; /* range at which continue begins */
176 int got_clen = 0; /* got content-length: from server */
177 FILE *output; /* socket to web server */
178 int quiet_flag = FALSE; /* Be verry, verry quiet... */
180 #define LONG_HEADER 1
181 struct option long_options[] = {
182 { "continue", 0, NULL, 'c' },
183 { "quiet", 0, NULL, 'q' },
184 { "output-document", 1, NULL, 'O' },
185 { "header", 1, &which_long_opt, LONG_HEADER },
189 * Crack command line.
191 while ((n = getopt_long(argc, argv, "cqO:", long_options, &option_index)) != EOF) {
200 /* can't set fname_out to NULL if outputting to stdout, because
201 * this gets interpreted as the auto-gen output filename
202 * case below - tausq@debian.org
207 switch (which_long_opt) {
209 int arglen = strlen(optarg);
210 if(extra_headers_left - arglen - 2 <= 0)
211 error_msg_and_die("extra_headers buffer too small(need %i)", extra_headers_left - arglen);
212 strcpy(extra_headers_ptr, optarg);
213 extra_headers_ptr += arglen;
214 extra_headers_left -= ( arglen + 2 );
215 *extra_headers_ptr++ = '\r';
216 *extra_headers_ptr++ = '\n';
217 *(extra_headers_ptr + 1) = 0;
227 fprintf(stderr, "extra_headers='%s'\n", extra_headers);
228 if (argc - optind != 1)
231 parse_url(argv[optind], &target);
232 server.host = target.host;
233 server.port = target.port;
236 * Use the proxy if necessary.
238 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
240 parse_url(xstrdup(proxy), &server);
242 /* Guess an output filename */
245 #ifdef BB_FEATURE_WGET_STATUSBAR
248 get_last_path_component(target.path);
249 if (fname_out==NULL || strlen(fname_out)<1) {
251 #ifdef BB_FEATURE_WGET_STATUSBAR
256 #ifdef BB_FEATURE_WGET_STATUSBAR
258 curfile = get_last_path_component(fname_out);
261 if (do_continue && !fname_out)
262 error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
266 * Open the output file stream.
268 if (strcmp(fname_out, "-") == 0) {
272 output = xfopen(fname_out, (do_continue ? "a" : "w"));
276 * Determine where to start transfer.
279 if (fstat(fileno(output), &sbuf) < 0)
280 perror_msg_and_die("fstat()");
281 if (sbuf.st_size > 0)
282 beg_range = sbuf.st_size;
287 if (proxy || !target.is_ftp) {
293 close_delete_and_die("too many redirections");
296 * Open socket to http server
298 if (sfp) fclose(sfp);
299 sfp = open_socket(server.host, server.port);
305 fprintf(sfp, "GET %stp://%s:%d/%s HTTP/1.1\r\n",
306 target.is_ftp ? "f" : "ht", target.host,
307 target.port, target.path);
309 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
312 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
314 #ifdef BB_FEATURE_WGET_AUTHENTICATION
316 fprintf(sfp, "Authorization: Basic %s\r\n",
317 base64enc(target.user, buf, sizeof(buf)));
319 if (proxy && server.user) {
320 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
321 base64enc(server.user, buf, sizeof(buf)));
326 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
327 if(extra_headers_left < sizeof(extra_headers))
328 fputs(extra_headers,sfp);
329 fprintf(sfp,"Connection: close\r\n\r\n");
332 * Retrieve HTTP response line and check for "200" status code.
334 read_response: if (fgets(buf, sizeof(buf), sfp) == NULL)
335 close_delete_and_die("no response from server");
337 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
339 for ( ; isspace(*s) ; ++s)
341 switch (status = atoi(s)) {
344 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
347 if (do_continue && output != stdout)
348 output = freopen(fname_out, "w", output);
351 case 300: /* redirection */
362 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
366 * Retrieve HTTP headers.
368 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
369 if (strcasecmp(buf, "content-length") == 0) {
374 if (strcasecmp(buf, "transfer-encoding") == 0) {
375 if (strcasecmp(s, "chunked") == 0) {
376 chunked = got_clen = 1;
378 close_delete_and_die("server wants to do %s transfer encoding", s);
381 if (strcasecmp(buf, "location") == 0) {
383 target.path = xstrdup(s+1);
385 parse_url(xstrdup(s), &target);
387 server.host = target.host;
388 server.port = target.port;
393 } while(status >= 300);
403 target.user = xstrdup("anonymous:busybox@");
405 sfp = open_socket(server.host, server.port);
406 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
407 close_delete_and_die("%s", buf+4);
410 * Splitting username:password pair,
413 s = strchr(target.user, ':');
416 switch(ftpcmd("USER ", target.user, sfp, buf)) {
420 if (ftpcmd("PASS ", s, sfp, buf) == 230)
422 /* FALLTHRU (failed login) */
424 close_delete_and_die("ftp login: %s", buf+4);
427 ftpcmd("CDUP", NULL, sfp, buf);
428 ftpcmd("TYPE I", NULL, sfp, buf);
433 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
434 filesize = atol(buf+4);
439 * Entering passive mode
441 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
442 close_delete_and_die("PASV: %s", buf+4);
443 s = strrchr(buf, ',');
446 s = strrchr(buf, ',');
447 port += atoi(s+1) * 256;
448 dfp = open_socket(server.host, port);
451 sprintf(buf, "REST %ld", beg_range);
452 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
453 if (output != stdout)
454 output = freopen(fname_out, "w", output);
457 filesize -= beg_range;
460 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
461 close_delete_and_die("RETR: %s", buf+4);
470 fgets(buf, sizeof(buf), dfp);
471 filesize = strtol(buf, (char **) NULL, 16);
473 #ifdef BB_FEATURE_WGET_STATUSBAR
474 if (quiet_flag==FALSE)
478 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, chunked ? (filesize > sizeof(buf) ? sizeof(buf) : filesize) : sizeof(buf), dfp)) > 0) {
479 safe_fwrite(buf, 1, n, output);
480 #ifdef BB_FEATURE_WGET_STATUSBAR
488 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
489 safe_fgets(buf, sizeof(buf), dfp);
490 filesize = strtol(buf, (char **) NULL, 16);
491 if (filesize==0) chunked = 0; /* all done! */
494 if (n == 0 && ferror(dfp))
495 perror_msg_and_die("network read error");
497 #ifdef BB_FEATURE_WGET_STATUSBAR
498 if (quiet_flag==FALSE)
501 if (!proxy && target.is_ftp) {
503 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
504 error_msg_and_die("ftp error: %s", buf+4);
505 ftpcmd("QUIT", NULL, sfp, buf);
511 void parse_url(char *url, struct host_info *h)
515 if (strncmp(url, "http://", 7) == 0) {
519 } else if (strncmp(url, "ftp://", 6) == 0) {
524 error_msg_and_die("not an http or ftp url: %s", url);
526 sp = strchr(h->host, '/');
533 up = strrchr(h->host, '@');
541 cp = strchr(h->host, ':');
550 FILE *open_socket(char *host, int port)
552 struct sockaddr_in s_in;
557 memset(&s_in, 0, sizeof(s_in));
558 s_in.sin_family = AF_INET;
559 hp = xgethostbyname(host);
560 memcpy(&s_in.sin_addr, hp->h_addr_list[0], hp->h_length);
561 s_in.sin_port = htons(port);
564 * Get the server onto a stdio stream.
566 if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0)
567 perror_msg_and_die("socket()");
568 if (connect(fd, (struct sockaddr *) &s_in, sizeof(s_in)) < 0)
569 perror_msg_and_die("connect(%s)", host);
570 if ((fp = fdopen(fd, "r+")) == NULL)
571 perror_msg_and_die("fdopen()");
577 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
584 /* retrieve header line */
585 if (fgets(buf, bufsiz, fp) == NULL)
588 /* see if we are at the end of the headers */
589 for (s = buf ; *s == '\r' ; ++s)
594 /* convert the header name to lower case */
595 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
598 /* verify we are at the end of the header name */
600 error_msg_and_die("bad header line: %s", buf);
602 /* locate the start of the header value */
603 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
607 /* locate the end of header */
608 while (*s != '\0' && *s != '\r' && *s != '\n')
611 /* end of header found */
617 /* Rats! The buffer isn't big enough to hold the entire header value. */
618 while (c = getc(fp), c != EOF && c != '\n')
624 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
630 fprintf(fp, "%s%s\n", s1, s2);
635 p = fgets(buf, 510, fp);
637 perror_msg_and_die("fgets()");
638 } while (! isdigit(buf[0]) || buf[3] != ' ');
643 #ifdef BB_FEATURE_WGET_STATUSBAR
644 /* Stuff below is from BSD rcp util.c, as added to openshh.
645 * Original copyright notice is retained at the end of this file.
653 struct winsize winsize;
655 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
656 return (winsize.ws_col ? winsize.ws_col : 80);
662 updateprogressmeter(int ignore)
664 int save_errno = errno;
673 struct itimerval itv;
675 itv.it_value.tv_sec = wait;
676 itv.it_value.tv_usec = 0;
677 itv.it_interval = itv.it_value;
678 setitimer(ITIMER_REAL, &itv, NULL);
683 progressmeter(int flag)
685 static const char prefixes[] = " KMGTP";
686 static struct timeval lastupdate;
687 static off_t lastsize, totalsize;
688 struct timeval now, td, wait;
689 off_t cursize, abbrevsize;
691 int ratio, barlength, i, remaining;
695 (void) gettimeofday(&start, (struct timezone *) 0);
698 totalsize = filesize; /* as filesize changes.. */
701 (void) gettimeofday(&now, (struct timezone *) 0);
703 if (totalsize != 0 && !chunked) {
704 ratio = 100.0 * cursize / totalsize;
705 ratio = MAX(ratio, 0);
706 ratio = MIN(ratio, 100);
710 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
712 barlength = getttywidth() - 51;
714 i = barlength * ratio / 100;
715 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
717 "*****************************************************************************"
718 "*****************************************************************************",
722 abbrevsize = cursize;
723 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
727 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
728 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
731 timersub(&now, &lastupdate, &wait);
732 if (cursize > lastsize) {
735 if (wait.tv_sec >= STALLTIME) {
736 start.tv_sec += wait.tv_sec;
737 start.tv_usec += wait.tv_usec;
741 timersub(&now, &start, &td);
742 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
744 if (wait.tv_sec >= STALLTIME) {
745 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
747 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
748 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
751 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
752 i = remaining / 3600;
754 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
757 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
759 i = remaining % 3600;
760 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
761 "%02d:%02d ETA", i / 60, i % 60);
763 write(fileno(stderr), buf, strlen(buf));
767 sa.sa_handler = updateprogressmeter;
768 sigemptyset(&sa.sa_mask);
769 sa.sa_flags = SA_RESTART;
770 sigaction(SIGALRM, &sa, NULL);
772 } else if (flag == 1) {
780 /* Original copyright notice which applies to the BB_FEATURE_WGET_STATUSBAR stuff,
781 * much of which was blatently stolen from openssh. */
784 * Copyright (c) 1992, 1993
785 * The Regents of the University of California. All rights reserved.
787 * Redistribution and use in source and binary forms, with or without
788 * modification, are permitted provided that the following conditions
790 * 1. Redistributions of source code must retain the above copyright
791 * notice, this list of conditions and the following disclaimer.
792 * 2. Redistributions in binary form must reproduce the above copyright
793 * notice, this list of conditions and the following disclaimer in the
794 * documentation and/or other materials provided with the distribution.
796 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
797 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
799 * 4. Neither the name of the University nor the names of its contributors
800 * may be used to endorse or promote products derived from this software
801 * without specific prior written permission.
803 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
804 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
805 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
806 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
807 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
808 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
809 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
810 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
811 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
812 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
815 * $Id: wget.c,v 1.41 2001/05/16 15:40:48 kraai Exp $
822 c-file-style: "linux"