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 CONFIG_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 CONFIG_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... */
180 int noproxy = 0; /* Use proxies if env vars are set */
182 #define LONG_HEADER 1
183 #define LONG_PASSIVE 2
185 struct option long_options[] = {
186 { "continue", 0, NULL, 'c' },
187 { "quiet", 0, NULL, 'q' },
188 { "output-document", 1, NULL, 'O' },
189 { "header", 1, &which_long_opt, LONG_HEADER },
190 { "proxy", 1, NULL, 'Y' },
191 { "passive-ftp", 0, &which_long_opt, LONG_PASSIVE },
195 * Crack command line.
197 while ((n = getopt_long(argc, argv, "cqO:P:Y:", long_options, &option_index)) != EOF) {
209 /* can't set fname_out to NULL if outputting to stdout, because
210 * this gets interpreted as the auto-gen output filename
211 * case below - tausq@debian.org
216 if (strcmp(optarg, "off") == 0)
220 switch (which_long_opt) {
222 int arglen = strlen(optarg);
223 if(extra_headers_left - arglen - 2 <= 0)
224 error_msg_and_die("extra_headers buffer too small(need %i)", extra_headers_left - arglen);
225 strcpy(extra_headers_ptr, optarg);
226 extra_headers_ptr += arglen;
227 extra_headers_left -= ( arglen + 2 );
228 *extra_headers_ptr++ = '\r';
229 *extra_headers_ptr++ = '\n';
230 *(extra_headers_ptr + 1) = 0;
234 // ignore -- we always use passive mode
243 if (argc - optind != 1)
246 parse_url(argv[optind], &target);
247 server.host = target.host;
248 server.port = target.port;
251 * Use the proxy if necessary.
254 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
256 parse_url(xstrdup(proxy), &server);
259 /* Guess an output filename */
262 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
265 get_last_path_component(target.path);
266 if (fname_out==NULL || strlen(fname_out)<1) {
268 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
273 if (dir_prefix != NULL)
274 fname_out = concat_path_file(dir_prefix, fname_out);
275 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
277 curfile = get_last_path_component(fname_out);
280 if (do_continue && !fname_out)
281 error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
285 * Open the output file stream.
287 if (strcmp(fname_out, "-") == 0) {
291 output = xfopen(fname_out, (do_continue ? "a" : "w"));
295 * Determine where to start transfer.
298 if (fstat(fileno(output), &sbuf) < 0)
299 perror_msg_and_die("fstat()");
300 if (sbuf.st_size > 0)
301 beg_range = sbuf.st_size;
306 if (proxy || !target.is_ftp) {
312 close_delete_and_die("too many redirections");
315 * Open socket to http server
317 if (sfp) fclose(sfp);
318 sfp = open_socket(server.host, server.port);
324 fprintf(sfp, "GET %stp://%s:%d/%s HTTP/1.1\r\n",
325 target.is_ftp ? "f" : "ht", target.host,
326 target.port, target.path);
328 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
331 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
333 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
335 fprintf(sfp, "Authorization: Basic %s\r\n",
336 base64enc(target.user, buf, sizeof(buf)));
338 if (proxy && server.user) {
339 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
340 base64enc(server.user, buf, sizeof(buf)));
345 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
346 if(extra_headers_left < sizeof(extra_headers))
347 fputs(extra_headers,sfp);
348 fprintf(sfp,"Connection: close\r\n\r\n");
351 * Retrieve HTTP response line and check for "200" status code.
353 read_response: if (fgets(buf, sizeof(buf), sfp) == NULL)
354 close_delete_and_die("no response from server");
356 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
358 for ( ; isspace(*s) ; ++s)
360 switch (status = atoi(s)) {
363 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
366 if (do_continue && output != stdout)
367 output = freopen(fname_out, "w", output);
370 case 300: /* redirection */
381 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
385 * Retrieve HTTP headers.
387 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
388 if (strcasecmp(buf, "content-length") == 0) {
393 if (strcasecmp(buf, "transfer-encoding") == 0) {
394 if (strcasecmp(s, "chunked") == 0) {
395 chunked = got_clen = 1;
397 close_delete_and_die("server wants to do %s transfer encoding", s);
400 if (strcasecmp(buf, "location") == 0) {
402 target.path = xstrdup(s+1);
404 parse_url(xstrdup(s), &target);
406 server.host = target.host;
407 server.port = target.port;
412 } while(status >= 300);
422 target.user = xstrdup("anonymous:busybox@");
424 sfp = open_socket(server.host, server.port);
425 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
426 close_delete_and_die("%s", buf+4);
429 * Splitting username:password pair,
432 s = strchr(target.user, ':');
435 switch(ftpcmd("USER ", target.user, sfp, buf)) {
439 if (ftpcmd("PASS ", s, sfp, buf) == 230)
441 /* FALLTHRU (failed login) */
443 close_delete_and_die("ftp login: %s", buf+4);
446 ftpcmd("CDUP", NULL, sfp, buf);
447 ftpcmd("TYPE I", NULL, sfp, buf);
452 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
453 filesize = atol(buf+4);
458 * Entering passive mode
460 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
461 close_delete_and_die("PASV: %s", buf+4);
462 s = strrchr(buf, ',');
465 s = strrchr(buf, ',');
466 port += atoi(s+1) * 256;
467 dfp = open_socket(server.host, port);
470 sprintf(buf, "REST %ld", beg_range);
471 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
472 if (output != stdout)
473 output = freopen(fname_out, "w", output);
476 filesize -= beg_range;
479 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
480 close_delete_and_die("RETR: %s", buf+4);
489 fgets(buf, sizeof(buf), dfp);
490 filesize = strtol(buf, (char **) NULL, 16);
492 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
493 if (quiet_flag==FALSE)
497 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, chunked ? (filesize > sizeof(buf) ? sizeof(buf) : filesize) : sizeof(buf), dfp)) > 0) {
498 if (safe_fwrite(buf, 1, n, output) != n)
499 perror_msg_and_die("write error");
500 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
508 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
509 safe_fgets(buf, sizeof(buf), dfp);
510 filesize = strtol(buf, (char **) NULL, 16);
511 if (filesize==0) chunked = 0; /* all done! */
514 if (n == 0 && ferror(dfp))
515 perror_msg_and_die("network read error");
517 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
518 if (quiet_flag==FALSE)
521 if (!proxy && target.is_ftp) {
523 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
524 error_msg_and_die("ftp error: %s", buf+4);
525 ftpcmd("QUIT", NULL, sfp, buf);
531 void parse_url(char *url, struct host_info *h)
535 if (strncmp(url, "http://", 7) == 0) {
539 } else if (strncmp(url, "ftp://", 6) == 0) {
544 error_msg_and_die("not an http or ftp url: %s", url);
546 sp = strchr(h->host, '/');
551 h->path = xstrdup("");
553 up = strrchr(h->host, '@');
561 cp = strchr(h->host, ':');
570 FILE *open_socket(char *host, int port)
576 snprintf(port_str, sizeof(port_str), "%d", port);
577 fd=xconnect(host, port_str);
580 * Get the server onto a stdio stream.
582 if ((fp = fdopen(fd, "r+")) == NULL)
583 perror_msg_and_die("fdopen()");
589 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
596 /* retrieve header line */
597 if (fgets(buf, bufsiz, fp) == NULL)
600 /* see if we are at the end of the headers */
601 for (s = buf ; *s == '\r' ; ++s)
606 /* convert the header name to lower case */
607 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
610 /* verify we are at the end of the header name */
612 error_msg_and_die("bad header line: %s", buf);
614 /* locate the start of the header value */
615 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
619 /* locate the end of header */
620 while (*s != '\0' && *s != '\r' && *s != '\n')
623 /* end of header found */
629 /* Rats! The buffer isn't big enough to hold the entire header value. */
630 while (c = getc(fp), c != EOF && c != '\n')
636 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
642 fprintf(fp, "%s%s\n", s1, s2);
647 p = fgets(buf, 510, fp);
649 perror_msg_and_die("fgets()");
650 } while (! isdigit(buf[0]) || buf[3] != ' ');
655 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
656 /* Stuff below is from BSD rcp util.c, as added to openshh.
657 * Original copyright notice is retained at the end of this file.
665 struct winsize winsize;
667 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
668 return (winsize.ws_col ? winsize.ws_col : 80);
674 updateprogressmeter(int ignore)
676 int save_errno = errno;
685 struct itimerval itv;
687 itv.it_value.tv_sec = wait;
688 itv.it_value.tv_usec = 0;
689 itv.it_interval = itv.it_value;
690 setitimer(ITIMER_REAL, &itv, NULL);
695 progressmeter(int flag)
697 static const char prefixes[] = " KMGTP";
698 static struct timeval lastupdate;
699 static off_t lastsize, totalsize;
700 struct timeval now, td, wait;
701 off_t cursize, abbrevsize;
703 int ratio, barlength, i, remaining;
707 (void) gettimeofday(&start, (struct timezone *) 0);
710 totalsize = filesize; /* as filesize changes.. */
713 (void) gettimeofday(&now, (struct timezone *) 0);
715 if (totalsize != 0 && !chunked) {
716 ratio = 100.0 * cursize / totalsize;
717 ratio = MAX(ratio, 0);
718 ratio = MIN(ratio, 100);
722 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
724 barlength = getttywidth() - 51;
726 i = barlength * ratio / 100;
727 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
729 "*****************************************************************************"
730 "*****************************************************************************",
734 abbrevsize = cursize;
735 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
739 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
740 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
743 timersub(&now, &lastupdate, &wait);
744 if (cursize > lastsize) {
747 if (wait.tv_sec >= STALLTIME) {
748 start.tv_sec += wait.tv_sec;
749 start.tv_usec += wait.tv_usec;
753 timersub(&now, &start, &td);
754 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
756 if (wait.tv_sec >= STALLTIME) {
757 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
759 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
760 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
763 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
764 i = remaining / 3600;
766 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
769 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
771 i = remaining % 3600;
772 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
773 "%02d:%02d ETA", i / 60, i % 60);
775 write(fileno(stderr), buf, strlen(buf));
779 sa.sa_handler = updateprogressmeter;
780 sigemptyset(&sa.sa_mask);
781 sa.sa_flags = SA_RESTART;
782 sigaction(SIGALRM, &sa, NULL);
784 } else if (flag == 1) {
792 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
793 * much of which was blatently stolen from openssh. */
796 * Copyright (c) 1992, 1993
797 * The Regents of the University of California. All rights reserved.
799 * Redistribution and use in source and binary forms, with or without
800 * modification, are permitted provided that the following conditions
802 * 1. Redistributions of source code must retain the above copyright
803 * notice, this list of conditions and the following disclaimer.
804 * 2. Redistributions in binary form must reproduce the above copyright
805 * notice, this list of conditions and the following disclaimer in the
806 * documentation and/or other materials provided with the distribution.
808 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
809 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
811 * 4. Neither the name of the University nor the names of its contributors
812 * may be used to endorse or promote products derived from this software
813 * without specific prior written permission.
815 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
816 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
817 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
818 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
819 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
820 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
821 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
822 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
823 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
824 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
827 * $Id: wget.c,v 1.51 2002/07/12 00:04:46 sandman Exp $
834 c-file-style: "linux"