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);
60 /* Globals (can be accessed from signal handlers */
61 static off_t filesize = 0; /* content-length of the file */
62 static int chunked = 0; /* chunked transfer encoding */
63 #ifdef BB_FEATURE_WGET_STATUSBAR
64 static void progressmeter(int flag);
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;
163 char *dir_prefix=NULL;
166 char extra_headers[1024];
167 char *extra_headers_ptr = extra_headers;
168 int extra_headers_left = sizeof(extra_headers);
169 int which_long_opt = 0, option_index = -1;
170 struct host_info server, target;
172 FILE *sfp = NULL; /* socket to web/ftp server */
173 FILE *dfp = NULL; /* socket to ftp server (data) */
174 char *fname_out = NULL; /* where to direct output (-O) */
175 int do_continue = 0; /* continue a prev transfer (-c) */
176 long beg_range = 0L; /* range at which continue begins */
177 int got_clen = 0; /* got content-length: from server */
178 FILE *output; /* socket to web server */
179 int quiet_flag = FALSE; /* Be verry, verry quiet... */
181 #define LONG_HEADER 1
182 struct option long_options[] = {
183 { "continue", 0, NULL, 'c' },
184 { "quiet", 0, NULL, 'q' },
185 { "output-document", 1, NULL, 'O' },
186 { "header", 1, &which_long_opt, LONG_HEADER },
190 * Crack command line.
192 while ((n = getopt_long(argc, argv, "cqO:P:", long_options, &option_index)) != EOF) {
204 /* can't set fname_out to NULL if outputting to stdout, because
205 * this gets interpreted as the auto-gen output filename
206 * case below - tausq@debian.org
211 switch (which_long_opt) {
213 int arglen = strlen(optarg);
214 if(extra_headers_left - arglen - 2 <= 0)
215 error_msg_and_die("extra_headers buffer too small(need %i)", extra_headers_left - arglen);
216 strcpy(extra_headers_ptr, optarg);
217 extra_headers_ptr += arglen;
218 extra_headers_left -= ( arglen + 2 );
219 *extra_headers_ptr++ = '\r';
220 *extra_headers_ptr++ = '\n';
221 *(extra_headers_ptr + 1) = 0;
231 if (argc - optind != 1)
234 parse_url(argv[optind], &target);
235 server.host = target.host;
236 server.port = target.port;
239 * Use the proxy if necessary.
241 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
243 parse_url(xstrdup(proxy), &server);
245 /* Guess an output filename */
248 #ifdef BB_FEATURE_WGET_STATUSBAR
251 get_last_path_component(target.path);
252 if (fname_out==NULL || strlen(fname_out)<1) {
254 #ifdef BB_FEATURE_WGET_STATUSBAR
259 if (dir_prefix != NULL)
260 fname_out = concat_path_file(dir_prefix, fname_out);
261 #ifdef BB_FEATURE_WGET_STATUSBAR
263 curfile = get_last_path_component(fname_out);
266 if (do_continue && !fname_out)
267 error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
271 * Open the output file stream.
273 if (strcmp(fname_out, "-") == 0) {
277 output = xfopen(fname_out, (do_continue ? "a" : "w"));
281 * Determine where to start transfer.
284 if (fstat(fileno(output), &sbuf) < 0)
285 perror_msg_and_die("fstat()");
286 if (sbuf.st_size > 0)
287 beg_range = sbuf.st_size;
292 if (proxy || !target.is_ftp) {
298 close_delete_and_die("too many redirections");
301 * Open socket to http server
303 if (sfp) fclose(sfp);
304 sfp = open_socket(server.host, server.port);
310 fprintf(sfp, "GET %stp://%s:%d/%s HTTP/1.1\r\n",
311 target.is_ftp ? "f" : "ht", target.host,
312 target.port, target.path);
314 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
317 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
319 #ifdef BB_FEATURE_WGET_AUTHENTICATION
321 fprintf(sfp, "Authorization: Basic %s\r\n",
322 base64enc(target.user, buf, sizeof(buf)));
324 if (proxy && server.user) {
325 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
326 base64enc(server.user, buf, sizeof(buf)));
331 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
332 if(extra_headers_left < sizeof(extra_headers))
333 fputs(extra_headers,sfp);
334 fprintf(sfp,"Connection: close\r\n\r\n");
337 * Retrieve HTTP response line and check for "200" status code.
339 read_response: if (fgets(buf, sizeof(buf), sfp) == NULL)
340 close_delete_and_die("no response from server");
342 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
344 for ( ; isspace(*s) ; ++s)
346 switch (status = atoi(s)) {
349 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
352 if (do_continue && output != stdout)
353 output = freopen(fname_out, "w", output);
356 case 300: /* redirection */
367 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
371 * Retrieve HTTP headers.
373 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
374 if (strcasecmp(buf, "content-length") == 0) {
379 if (strcasecmp(buf, "transfer-encoding") == 0) {
380 if (strcasecmp(s, "chunked") == 0) {
381 chunked = got_clen = 1;
383 close_delete_and_die("server wants to do %s transfer encoding", s);
386 if (strcasecmp(buf, "location") == 0) {
388 target.path = xstrdup(s+1);
390 parse_url(xstrdup(s), &target);
392 server.host = target.host;
393 server.port = target.port;
398 } while(status >= 300);
408 target.user = xstrdup("anonymous:busybox@");
410 sfp = open_socket(server.host, server.port);
411 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
412 close_delete_and_die("%s", buf+4);
415 * Splitting username:password pair,
418 s = strchr(target.user, ':');
421 switch(ftpcmd("USER ", target.user, sfp, buf)) {
425 if (ftpcmd("PASS ", s, sfp, buf) == 230)
427 /* FALLTHRU (failed login) */
429 close_delete_and_die("ftp login: %s", buf+4);
432 ftpcmd("CDUP", NULL, sfp, buf);
433 ftpcmd("TYPE I", NULL, sfp, buf);
438 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
439 filesize = atol(buf+4);
444 * Entering passive mode
446 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
447 close_delete_and_die("PASV: %s", buf+4);
448 s = strrchr(buf, ',');
451 s = strrchr(buf, ',');
452 port += atoi(s+1) * 256;
453 dfp = open_socket(server.host, port);
456 sprintf(buf, "REST %ld", beg_range);
457 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
458 if (output != stdout)
459 output = freopen(fname_out, "w", output);
462 filesize -= beg_range;
465 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
466 close_delete_and_die("RETR: %s", buf+4);
475 fgets(buf, sizeof(buf), dfp);
476 filesize = strtol(buf, (char **) NULL, 16);
478 #ifdef BB_FEATURE_WGET_STATUSBAR
479 if (quiet_flag==FALSE)
483 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, chunked ? (filesize > sizeof(buf) ? sizeof(buf) : filesize) : sizeof(buf), dfp)) > 0) {
484 safe_fwrite(buf, 1, n, output);
485 #ifdef BB_FEATURE_WGET_STATUSBAR
493 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
494 safe_fgets(buf, sizeof(buf), dfp);
495 filesize = strtol(buf, (char **) NULL, 16);
496 if (filesize==0) chunked = 0; /* all done! */
499 if (n == 0 && ferror(dfp))
500 perror_msg_and_die("network read error");
502 #ifdef BB_FEATURE_WGET_STATUSBAR
503 if (quiet_flag==FALSE)
506 if (!proxy && target.is_ftp) {
508 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
509 error_msg_and_die("ftp error: %s", buf+4);
510 ftpcmd("QUIT", NULL, sfp, buf);
516 void parse_url(char *url, struct host_info *h)
520 if (strncmp(url, "http://", 7) == 0) {
524 } else if (strncmp(url, "ftp://", 6) == 0) {
529 error_msg_and_die("not an http or ftp url: %s", url);
531 sp = strchr(h->host, '/');
538 up = strrchr(h->host, '@');
546 cp = strchr(h->host, ':');
555 FILE *open_socket(char *host, int port)
557 struct sockaddr_in s_in;
562 memset(&s_in, 0, sizeof(s_in));
563 s_in.sin_family = AF_INET;
564 hp = xgethostbyname(host);
565 memcpy(&s_in.sin_addr, hp->h_addr_list[0], hp->h_length);
566 s_in.sin_port = htons(port);
569 * Get the server onto a stdio stream.
571 if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0)
572 perror_msg_and_die("socket()");
573 if (connect(fd, (struct sockaddr *) &s_in, sizeof(s_in)) < 0)
574 perror_msg_and_die("connect(%s)", host);
575 if ((fp = fdopen(fd, "r+")) == NULL)
576 perror_msg_and_die("fdopen()");
582 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
589 /* retrieve header line */
590 if (fgets(buf, bufsiz, fp) == NULL)
593 /* see if we are at the end of the headers */
594 for (s = buf ; *s == '\r' ; ++s)
599 /* convert the header name to lower case */
600 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
603 /* verify we are at the end of the header name */
605 error_msg_and_die("bad header line: %s", buf);
607 /* locate the start of the header value */
608 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
612 /* locate the end of header */
613 while (*s != '\0' && *s != '\r' && *s != '\n')
616 /* end of header found */
622 /* Rats! The buffer isn't big enough to hold the entire header value. */
623 while (c = getc(fp), c != EOF && c != '\n')
629 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
635 fprintf(fp, "%s%s\n", s1, s2);
640 p = fgets(buf, 510, fp);
642 perror_msg_and_die("fgets()");
643 } while (! isdigit(buf[0]) || buf[3] != ' ');
648 #ifdef BB_FEATURE_WGET_STATUSBAR
649 /* Stuff below is from BSD rcp util.c, as added to openshh.
650 * Original copyright notice is retained at the end of this file.
658 struct winsize winsize;
660 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
661 return (winsize.ws_col ? winsize.ws_col : 80);
667 updateprogressmeter(int ignore)
669 int save_errno = errno;
678 struct itimerval itv;
680 itv.it_value.tv_sec = wait;
681 itv.it_value.tv_usec = 0;
682 itv.it_interval = itv.it_value;
683 setitimer(ITIMER_REAL, &itv, NULL);
688 progressmeter(int flag)
690 static const char prefixes[] = " KMGTP";
691 static struct timeval lastupdate;
692 static off_t lastsize, totalsize;
693 struct timeval now, td, wait;
694 off_t cursize, abbrevsize;
696 int ratio, barlength, i, remaining;
700 (void) gettimeofday(&start, (struct timezone *) 0);
703 totalsize = filesize; /* as filesize changes.. */
706 (void) gettimeofday(&now, (struct timezone *) 0);
708 if (totalsize != 0 && !chunked) {
709 ratio = 100.0 * cursize / totalsize;
710 ratio = MAX(ratio, 0);
711 ratio = MIN(ratio, 100);
715 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
717 barlength = getttywidth() - 51;
719 i = barlength * ratio / 100;
720 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
722 "*****************************************************************************"
723 "*****************************************************************************",
727 abbrevsize = cursize;
728 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
732 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
733 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
736 timersub(&now, &lastupdate, &wait);
737 if (cursize > lastsize) {
740 if (wait.tv_sec >= STALLTIME) {
741 start.tv_sec += wait.tv_sec;
742 start.tv_usec += wait.tv_usec;
746 timersub(&now, &start, &td);
747 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
749 if (wait.tv_sec >= STALLTIME) {
750 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
752 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
753 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
756 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
757 i = remaining / 3600;
759 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
762 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
764 i = remaining % 3600;
765 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
766 "%02d:%02d ETA", i / 60, i % 60);
768 write(fileno(stderr), buf, strlen(buf));
772 sa.sa_handler = updateprogressmeter;
773 sigemptyset(&sa.sa_mask);
774 sa.sa_flags = SA_RESTART;
775 sigaction(SIGALRM, &sa, NULL);
777 } else if (flag == 1) {
785 /* Original copyright notice which applies to the BB_FEATURE_WGET_STATUSBAR stuff,
786 * much of which was blatently stolen from openssh. */
789 * Copyright (c) 1992, 1993
790 * The Regents of the University of California. All rights reserved.
792 * Redistribution and use in source and binary forms, with or without
793 * modification, are permitted provided that the following conditions
795 * 1. Redistributions of source code must retain the above copyright
796 * notice, this list of conditions and the following disclaimer.
797 * 2. Redistributions in binary form must reproduce the above copyright
798 * notice, this list of conditions and the following disclaimer in the
799 * documentation and/or other materials provided with the distribution.
801 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
802 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
804 * 4. Neither the name of the University nor the names of its contributors
805 * may be used to endorse or promote products derived from this software
806 * without specific prior written permission.
808 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
809 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
810 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
811 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
812 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
813 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
814 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
815 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
816 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
817 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
820 * $Id: wget.c,v 1.45 2001/07/19 22:28:01 andersen Exp $
827 c-file-style: "linux"