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>
42 static void parse_url(char *url, struct host_info *h);
43 static FILE *open_socket(struct sockaddr_in *s_in, int port);
44 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
45 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf);
47 /* Globals (can be accessed from signal handlers */
48 static off_t filesize = 0; /* content-length of the file */
49 static int chunked = 0; /* chunked transfer encoding */
50 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
51 static void progressmeter(int flag);
52 static char *curfile; /* Name of current file being transferred. */
53 static struct timeval start; /* Time a transfer started. */
54 static volatile unsigned long statbytes = 0; /* Number of bytes transferred so far. */
55 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
56 static const int STALLTIME = 5;
59 static void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
61 if (output != stdout && do_continue==0) {
67 /* Read NMEMB elements of SIZE bytes into PTR from STREAM. Returns the
68 * number of elements read, and a short count if an eof or non-interrupt
69 * error is encountered. */
70 static size_t safe_fread(void *ptr, size_t size, size_t nmemb, FILE *stream)
76 ret += fread((char *)ptr + (ret * size), size, nmemb - ret, stream);
77 } while (ret < nmemb && ferror(stream) && errno == EINTR);
82 /* Write NMEMB elements of SIZE bytes from PTR to STREAM. Returns the
83 * number of elements written, and a short count if an eof or non-interrupt
84 * error is encountered. */
85 static size_t safe_fwrite(void *ptr, size_t size, size_t nmemb, FILE *stream)
91 ret += fwrite((char *)ptr + (ret * size), size, nmemb - ret, stream);
92 } while (ret < nmemb && ferror(stream) && errno == EINTR);
97 /* Read a line or SIZE - 1 bytes into S, whichever is less, from STREAM.
98 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
99 static char *safe_fgets(char *s, int size, FILE *stream)
105 ret = fgets(s, size, stream);
106 } while (ret == NULL && ferror(stream) && errno == EINTR);
111 #define close_delete_and_die(s...) { \
112 close_and_delete_outfile(output, fname_out, do_continue); \
113 bb_error_msg_and_die(s); }
116 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
118 * Base64-encode character string
119 * oops... isn't something similar in uuencode.c?
120 * It would be better to use already existing code
122 char *base64enc(char *p, char *buf, int len) {
124 char al[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
130 bb_error_msg_and_die("buffer overflow");
131 *(s++) = al[(*p >> 2) & 0x3F];
132 *(s++) = al[((*p << 4) & 0x30) | ((*(p+1) >> 4) & 0x0F)];
136 *(s++) = al[((*p << 2) & 0x3C) | ((*(p+1) >> 6) & 0x03)];
138 *(s++) = al[*(p++) & 0x3F];
145 int wget_main(int argc, char **argv)
147 int n, try=5, status;
150 char *dir_prefix=NULL;
153 char extra_headers[1024];
154 char *extra_headers_ptr = extra_headers;
155 int extra_headers_left = sizeof(extra_headers);
156 int which_long_opt = 0, option_index = -1;
157 struct host_info server, target;
158 struct sockaddr_in s_in;
160 FILE *sfp = NULL; /* socket to web/ftp server */
161 FILE *dfp = NULL; /* socket to ftp server (data) */
162 char *fname_out = NULL; /* where to direct output (-O) */
163 int do_continue = 0; /* continue a prev transfer (-c) */
164 long beg_range = 0L; /* range at which continue begins */
165 int got_clen = 0; /* got content-length: from server */
166 FILE *output; /* socket to web server */
167 int quiet_flag = FALSE; /* Be verry, verry quiet... */
168 int noproxy = 0; /* Use proxies if env vars are set */
170 #define LONG_HEADER 1
171 #define LONG_PASSIVE 2
173 struct option long_options[] = {
174 { "continue", 0, NULL, 'c' },
175 { "quiet", 0, NULL, 'q' },
176 { "output-document", 1, NULL, 'O' },
177 { "header", 1, &which_long_opt, LONG_HEADER },
178 { "proxy", 1, NULL, 'Y' },
179 { "passive-ftp", 0, &which_long_opt, LONG_PASSIVE },
183 * Crack command line.
185 while ((n = getopt_long(argc, argv, "cqO:P:Y:", long_options, &option_index)) != EOF) {
197 /* can't set fname_out to NULL if outputting to stdout, because
198 * this gets interpreted as the auto-gen output filename
199 * case below - tausq@debian.org
204 if (strcmp(optarg, "off") == 0)
208 switch (which_long_opt) {
210 int arglen = strlen(optarg);
211 if(extra_headers_left - arglen - 2 <= 0)
212 bb_error_msg_and_die("extra_headers buffer too small(need %i)", extra_headers_left - arglen);
213 strcpy(extra_headers_ptr, optarg);
214 extra_headers_ptr += arglen;
215 extra_headers_left -= ( arglen + 2 );
216 *extra_headers_ptr++ = '\r';
217 *extra_headers_ptr++ = '\n';
218 *(extra_headers_ptr + 1) = 0;
222 // ignore -- we always use passive mode
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.
242 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
244 parse_url(bb_xstrdup(proxy), &server);
247 /* Guess an output filename */
250 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
253 bb_get_last_path_component(target.path);
254 if (fname_out==NULL || strlen(fname_out)<1) {
256 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
261 if (dir_prefix != NULL)
262 fname_out = concat_path_file(dir_prefix, fname_out);
263 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
265 curfile = bb_get_last_path_component(fname_out);
268 if (do_continue && !fname_out)
269 bb_error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
273 * Open the output file stream.
275 if (strcmp(fname_out, "-") == 0) {
279 output = bb_xfopen(fname_out, (do_continue ? "a" : "w"));
283 * Determine where to start transfer.
286 if (fstat(fileno(output), &sbuf) < 0)
287 bb_perror_msg_and_die("fstat()");
288 if (sbuf.st_size > 0)
289 beg_range = sbuf.st_size;
294 /* We want to do exactly _one_ DNS lookup, since some
295 * sites (i.e. ftp.us.debian.org) use round-robin DNS
296 * and we want to connect to only one IP... */
297 bb_lookup_host(&s_in, server.host, NULL);
298 if (quiet_flag==FALSE) {
299 fprintf(stdout, "Connecting to %s[%s]:%d\n",
300 server.host, inet_ntoa(s_in.sin_addr), server.port);
303 if (proxy || !target.is_ftp) {
308 got_clen = chunked = 0;
311 close_delete_and_die("too many redirections");
314 * Open socket to http server
316 if (sfp) fclose(sfp);
317 sfp = open_socket(&s_in, server.port);
323 const char *format = "GET %stp://%s:%d/%s HTTP/1.1\r\n";
324 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
325 if (strchr (target.host, ':'))
326 format = "GET %stp://[%s]:%d/%s HTTP/1.1\r\n";
329 target.is_ftp ? "f" : "ht", target.host,
330 target.port, target.path);
332 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
335 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
337 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
339 fprintf(sfp, "Authorization: Basic %s\r\n",
340 base64enc(target.user, buf, sizeof(buf)));
342 if (proxy && server.user) {
343 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
344 base64enc(server.user, buf, sizeof(buf)));
349 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
350 if(extra_headers_left < sizeof(extra_headers))
351 fputs(extra_headers,sfp);
352 fprintf(sfp,"Connection: close\r\n\r\n");
355 * Retrieve HTTP response line and check for "200" status code.
358 if (fgets(buf, sizeof(buf), sfp) == NULL)
359 close_delete_and_die("no response from server");
361 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
363 for ( ; isspace(*s) ; ++s)
365 switch (status = atoi(s)) {
368 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
371 if (do_continue && output != stdout)
372 output = freopen(fname_out, "w", output);
375 case 300: /* redirection */
386 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
390 * Retrieve HTTP headers.
392 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
393 if (strcasecmp(buf, "content-length") == 0) {
398 if (strcasecmp(buf, "transfer-encoding") == 0) {
399 if (strcasecmp(s, "chunked") == 0) {
400 chunked = got_clen = 1;
402 close_delete_and_die("server wants to do %s transfer encoding", s);
405 if (strcasecmp(buf, "location") == 0) {
407 target.path = bb_xstrdup(s+1);
409 parse_url(bb_xstrdup(s), &target);
411 server.host = target.host;
412 server.port = target.port;
417 } while(status >= 300);
427 target.user = bb_xstrdup("anonymous:busybox@");
429 sfp = open_socket(&s_in, server.port);
430 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
431 close_delete_and_die("%s", buf+4);
434 * Splitting username:password pair,
437 s = strchr(target.user, ':');
440 switch(ftpcmd("USER ", target.user, sfp, buf)) {
444 if (ftpcmd("PASS ", s, sfp, buf) == 230)
446 /* FALLTHRU (failed login) */
448 close_delete_and_die("ftp login: %s", buf+4);
451 ftpcmd("CDUP", NULL, sfp, buf);
452 ftpcmd("TYPE I", NULL, sfp, buf);
457 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
458 filesize = atol(buf+4);
463 * Entering passive mode
465 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
466 close_delete_and_die("PASV: %s", buf+4);
467 s = strrchr(buf, ',');
470 s = strrchr(buf, ',');
471 port += atoi(s+1) * 256;
472 dfp = open_socket(&s_in, port);
475 sprintf(buf, "REST %ld", beg_range);
476 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
477 if (output != stdout)
478 output = freopen(fname_out, "w", output);
481 filesize -= beg_range;
484 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
485 close_delete_and_die("RETR: %s", buf+4);
494 fgets(buf, sizeof(buf), dfp);
495 filesize = strtol(buf, (char **) NULL, 16);
497 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
498 if (quiet_flag==FALSE)
502 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, ((chunked || got_clen) && (filesize < sizeof(buf)) ? filesize : sizeof(buf)), dfp)) > 0) {
503 if (safe_fwrite(buf, 1, n, output) != n) {
504 bb_perror_msg_and_die("write error");
506 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
515 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
516 safe_fgets(buf, sizeof(buf), dfp);
517 filesize = strtol(buf, (char **) NULL, 16);
519 chunked = 0; /* all done! */
523 if (n == 0 && ferror(dfp)) {
524 bb_perror_msg_and_die("network read error");
527 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
528 if (quiet_flag==FALSE)
531 if (!proxy && target.is_ftp) {
533 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
534 bb_error_msg_and_die("ftp error: %s", buf+4);
535 ftpcmd("QUIT", NULL, sfp, buf);
541 void parse_url(char *url, struct host_info *h)
543 char *cp, *sp, *up, *pp;
545 if (strncmp(url, "http://", 7) == 0) {
549 } else if (strncmp(url, "ftp://", 6) == 0) {
554 bb_error_msg_and_die("not an http or ftp url: %s", url);
556 sp = strchr(h->host, '/');
561 h->path = bb_xstrdup("");
563 up = strrchr(h->host, '@');
573 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
574 if (h->host[0] == '[') {
578 while (*ep == ':' || isxdigit (*ep))
588 cp = strchr(pp, ':');
597 FILE *open_socket(struct sockaddr_in *s_in, int port)
602 if (port>0 && port < 65536) {
603 s_in->sin_port=htons(port);
609 * Get the server onto a stdio stream.
611 if ((fp = fdopen(fd, "r+")) == NULL)
612 bb_perror_msg_and_die("fdopen()");
618 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
625 /* retrieve header line */
626 if (fgets(buf, bufsiz, fp) == NULL)
629 /* see if we are at the end of the headers */
630 for (s = buf ; *s == '\r' ; ++s)
635 /* convert the header name to lower case */
636 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
639 /* verify we are at the end of the header name */
641 bb_error_msg_and_die("bad header line: %s", buf);
643 /* locate the start of the header value */
644 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
648 /* locate the end of header */
649 while (*s != '\0' && *s != '\r' && *s != '\n')
652 /* end of header found */
658 /* Rats! The buffer isn't big enough to hold the entire header value. */
659 while (c = getc(fp), c != EOF && c != '\n')
665 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
671 fprintf(fp, "%s%s\r\n", s1, s2);
676 p = fgets(buf, 510, fp);
678 bb_perror_msg_and_die("fgets()");
679 } while (! isdigit(buf[0]) || buf[3] != ' ');
684 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
685 /* Stuff below is from BSD rcp util.c, as added to openshh.
686 * Original copyright notice is retained at the end of this file.
695 get_terminal_width_height(0, &width, NULL);
700 updateprogressmeter(int ignore)
702 int save_errno = errno;
711 struct itimerval itv;
713 itv.it_value.tv_sec = wait;
714 itv.it_value.tv_usec = 0;
715 itv.it_interval = itv.it_value;
716 setitimer(ITIMER_REAL, &itv, NULL);
721 progressmeter(int flag)
723 static const char prefixes[] = " KMGTP";
724 static struct timeval lastupdate;
725 static off_t lastsize, totalsize;
726 struct timeval now, td, wait;
727 off_t cursize, abbrevsize;
729 int ratio, barlength, i, remaining;
733 (void) gettimeofday(&start, (struct timezone *) 0);
736 totalsize = filesize; /* as filesize changes.. */
739 (void) gettimeofday(&now, (struct timezone *) 0);
741 if (totalsize != 0 && !chunked) {
742 ratio = 100.0 * cursize / totalsize;
743 ratio = MAX(ratio, 0);
744 ratio = MIN(ratio, 100);
748 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
750 barlength = getttywidth() - 51;
752 i = barlength * ratio / 100;
753 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
755 "*****************************************************************************"
756 "*****************************************************************************",
760 abbrevsize = cursize;
761 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
765 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
766 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
769 timersub(&now, &lastupdate, &wait);
770 if (cursize > lastsize) {
773 if (wait.tv_sec >= STALLTIME) {
774 start.tv_sec += wait.tv_sec;
775 start.tv_usec += wait.tv_usec;
779 timersub(&now, &start, &td);
780 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
782 if (wait.tv_sec >= STALLTIME) {
783 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
785 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
786 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
789 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
790 i = remaining / 3600;
792 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
795 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
797 i = remaining % 3600;
798 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
799 "%02d:%02d ETA", i / 60, i % 60);
801 write(fileno(stderr), buf, strlen(buf));
805 sa.sa_handler = updateprogressmeter;
806 sigemptyset(&sa.sa_mask);
807 sa.sa_flags = SA_RESTART;
808 sigaction(SIGALRM, &sa, NULL);
810 } else if (flag == 1) {
818 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
819 * much of which was blatently stolen from openssh. */
822 * Copyright (c) 1992, 1993
823 * The Regents of the University of California. All rights reserved.
825 * Redistribution and use in source and binary forms, with or without
826 * modification, are permitted provided that the following conditions
828 * 1. Redistributions of source code must retain the above copyright
829 * notice, this list of conditions and the following disclaimer.
830 * 2. Redistributions in binary form must reproduce the above copyright
831 * notice, this list of conditions and the following disclaimer in the
832 * documentation and/or other materials provided with the distribution.
834 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
835 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
837 * 4. Neither the name of the University nor the names of its contributors
838 * may be used to endorse or promote products derived from this software
839 * without specific prior written permission.
841 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
842 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
843 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
844 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
845 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
846 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
847 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
848 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
849 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
850 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
853 * $Id: wget.c,v 1.61 2003/10/31 09:31:43 andersen Exp $
860 c-file-style: "linux"