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(char *host, 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;
159 FILE *sfp = NULL; /* socket to web/ftp server */
160 FILE *dfp = NULL; /* socket to ftp server (data) */
161 char *fname_out = NULL; /* where to direct output (-O) */
162 int do_continue = 0; /* continue a prev transfer (-c) */
163 long beg_range = 0L; /* range at which continue begins */
164 int got_clen = 0; /* got content-length: from server */
165 FILE *output; /* socket to web server */
166 int quiet_flag = FALSE; /* Be verry, verry quiet... */
167 int noproxy = 0; /* Use proxies if env vars are set */
169 #define LONG_HEADER 1
170 #define LONG_PASSIVE 2
172 struct option long_options[] = {
173 { "continue", 0, NULL, 'c' },
174 { "quiet", 0, NULL, 'q' },
175 { "output-document", 1, NULL, 'O' },
176 { "header", 1, &which_long_opt, LONG_HEADER },
177 { "proxy", 1, NULL, 'Y' },
178 { "passive-ftp", 0, &which_long_opt, LONG_PASSIVE },
182 * Crack command line.
184 while ((n = getopt_long(argc, argv, "cqO:P:Y:", long_options, &option_index)) != EOF) {
196 /* can't set fname_out to NULL if outputting to stdout, because
197 * this gets interpreted as the auto-gen output filename
198 * case below - tausq@debian.org
203 if (strcmp(optarg, "off") == 0)
207 switch (which_long_opt) {
209 int arglen = strlen(optarg);
210 if(extra_headers_left - arglen - 2 <= 0)
211 bb_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;
221 // ignore -- we always use passive mode
230 if (argc - optind != 1)
233 parse_url(argv[optind], &target);
234 server.host = target.host;
235 server.port = target.port;
238 * Use the proxy if necessary.
241 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
243 parse_url(bb_xstrdup(proxy), &server);
246 /* Guess an output filename */
249 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
252 bb_get_last_path_component(target.path);
253 if (fname_out==NULL || strlen(fname_out)<1) {
255 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
260 if (dir_prefix != NULL)
261 fname_out = concat_path_file(dir_prefix, fname_out);
262 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
264 curfile = bb_get_last_path_component(fname_out);
267 if (do_continue && !fname_out)
268 bb_error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
272 * Open the output file stream.
274 if (strcmp(fname_out, "-") == 0) {
278 output = bb_xfopen(fname_out, (do_continue ? "a" : "w"));
282 * Determine where to start transfer.
285 if (fstat(fileno(output), &sbuf) < 0)
286 bb_perror_msg_and_die("fstat()");
287 if (sbuf.st_size > 0)
288 beg_range = sbuf.st_size;
293 if (proxy || !target.is_ftp) {
298 got_clen = chunked = 0;
301 close_delete_and_die("too many redirections");
304 * Open socket to http server
306 if (sfp) fclose(sfp);
307 sfp = open_socket(server.host, server.port);
313 const char *format = "GET %stp://%s:%d/%s HTTP/1.1\r\n";
314 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
315 if (strchr (target.host, ':'))
316 format = "GET %stp://[%s]:%d/%s HTTP/1.1\r\n";
319 target.is_ftp ? "f" : "ht", target.host,
320 target.port, target.path);
322 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
325 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
327 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
329 fprintf(sfp, "Authorization: Basic %s\r\n",
330 base64enc(target.user, buf, sizeof(buf)));
332 if (proxy && server.user) {
333 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
334 base64enc(server.user, buf, sizeof(buf)));
339 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
340 if(extra_headers_left < sizeof(extra_headers))
341 fputs(extra_headers,sfp);
342 fprintf(sfp,"Connection: close\r\n\r\n");
345 * Retrieve HTTP response line and check for "200" status code.
348 if (fgets(buf, sizeof(buf), sfp) == NULL)
349 close_delete_and_die("no response from server");
351 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
353 for ( ; isspace(*s) ; ++s)
355 switch (status = atoi(s)) {
358 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
361 if (do_continue && output != stdout)
362 output = freopen(fname_out, "w", output);
365 case 300: /* redirection */
376 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
380 * Retrieve HTTP headers.
382 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
383 if (strcasecmp(buf, "content-length") == 0) {
388 if (strcasecmp(buf, "transfer-encoding") == 0) {
389 if (strcasecmp(s, "chunked") == 0) {
390 chunked = got_clen = 1;
392 close_delete_and_die("server wants to do %s transfer encoding", s);
395 if (strcasecmp(buf, "location") == 0) {
397 target.path = bb_xstrdup(s+1);
399 parse_url(bb_xstrdup(s), &target);
401 server.host = target.host;
402 server.port = target.port;
407 } while(status >= 300);
417 target.user = bb_xstrdup("anonymous:busybox@");
419 sfp = open_socket(server.host, server.port);
420 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
421 close_delete_and_die("%s", buf+4);
424 * Splitting username:password pair,
427 s = strchr(target.user, ':');
430 switch(ftpcmd("USER ", target.user, sfp, buf)) {
434 if (ftpcmd("PASS ", s, sfp, buf) == 230)
436 /* FALLTHRU (failed login) */
438 close_delete_and_die("ftp login: %s", buf+4);
441 ftpcmd("CDUP", NULL, sfp, buf);
442 ftpcmd("TYPE I", NULL, sfp, buf);
447 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
448 filesize = atol(buf+4);
453 * Entering passive mode
455 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
456 close_delete_and_die("PASV: %s", buf+4);
457 s = strrchr(buf, ',');
460 s = strrchr(buf, ',');
461 port += atoi(s+1) * 256;
462 dfp = open_socket(server.host, port);
465 sprintf(buf, "REST %ld", beg_range);
466 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
467 if (output != stdout)
468 output = freopen(fname_out, "w", output);
471 filesize -= beg_range;
474 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
475 close_delete_and_die("RETR: %s", buf+4);
484 fgets(buf, sizeof(buf), dfp);
485 filesize = strtol(buf, (char **) NULL, 16);
487 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
488 if (quiet_flag==FALSE)
492 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, ((chunked || got_clen) && (filesize < sizeof(buf)) ? filesize : sizeof(buf)), dfp)) > 0) {
493 if (safe_fwrite(buf, 1, n, output) != n) {
494 bb_perror_msg_and_die("write error");
496 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
505 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
506 safe_fgets(buf, sizeof(buf), dfp);
507 filesize = strtol(buf, (char **) NULL, 16);
509 chunked = 0; /* all done! */
513 if (n == 0 && ferror(dfp)) {
514 bb_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 bb_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)
533 char *cp, *sp, *up, *pp;
535 if (strncmp(url, "http://", 7) == 0) {
539 } else if (strncmp(url, "ftp://", 6) == 0) {
544 bb_error_msg_and_die("not an http or ftp url: %s", url);
546 sp = strchr(h->host, '/');
551 h->path = bb_xstrdup("");
553 up = strrchr(h->host, '@');
563 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
564 if (h->host[0] == '[') {
568 while (*ep == ':' || isxdigit (*ep))
578 cp = strchr(pp, ':');
587 FILE *open_socket(char *host, int port)
593 snprintf(port_str, sizeof(port_str), "%d", port);
594 fd=xconnect(host, port_str);
597 * Get the server onto a stdio stream.
599 if ((fp = fdopen(fd, "r+")) == NULL)
600 bb_perror_msg_and_die("fdopen()");
606 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
613 /* retrieve header line */
614 if (fgets(buf, bufsiz, fp) == NULL)
617 /* see if we are at the end of the headers */
618 for (s = buf ; *s == '\r' ; ++s)
623 /* convert the header name to lower case */
624 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
627 /* verify we are at the end of the header name */
629 bb_error_msg_and_die("bad header line: %s", buf);
631 /* locate the start of the header value */
632 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
636 /* locate the end of header */
637 while (*s != '\0' && *s != '\r' && *s != '\n')
640 /* end of header found */
646 /* Rats! The buffer isn't big enough to hold the entire header value. */
647 while (c = getc(fp), c != EOF && c != '\n')
653 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
659 fprintf(fp, "%s%s\r\n", s1, s2);
664 p = fgets(buf, 510, fp);
666 bb_perror_msg_and_die("fgets()");
667 } while (! isdigit(buf[0]) || buf[3] != ' ');
672 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
673 /* Stuff below is from BSD rcp util.c, as added to openshh.
674 * Original copyright notice is retained at the end of this file.
683 get_terminal_width_height(0, &width, NULL);
688 updateprogressmeter(int ignore)
690 int save_errno = errno;
699 struct itimerval itv;
701 itv.it_value.tv_sec = wait;
702 itv.it_value.tv_usec = 0;
703 itv.it_interval = itv.it_value;
704 setitimer(ITIMER_REAL, &itv, NULL);
709 progressmeter(int flag)
711 static const char prefixes[] = " KMGTP";
712 static struct timeval lastupdate;
713 static off_t lastsize, totalsize;
714 struct timeval now, td, wait;
715 off_t cursize, abbrevsize;
717 int ratio, barlength, i, remaining;
721 (void) gettimeofday(&start, (struct timezone *) 0);
724 totalsize = filesize; /* as filesize changes.. */
727 (void) gettimeofday(&now, (struct timezone *) 0);
729 if (totalsize != 0 && !chunked) {
730 ratio = 100.0 * cursize / totalsize;
731 ratio = MAX(ratio, 0);
732 ratio = MIN(ratio, 100);
736 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
738 barlength = getttywidth() - 51;
740 i = barlength * ratio / 100;
741 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
743 "*****************************************************************************"
744 "*****************************************************************************",
748 abbrevsize = cursize;
749 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
753 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
754 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
757 timersub(&now, &lastupdate, &wait);
758 if (cursize > lastsize) {
761 if (wait.tv_sec >= STALLTIME) {
762 start.tv_sec += wait.tv_sec;
763 start.tv_usec += wait.tv_usec;
767 timersub(&now, &start, &td);
768 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
770 if (wait.tv_sec >= STALLTIME) {
771 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
773 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
774 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
777 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
778 i = remaining / 3600;
780 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
783 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
785 i = remaining % 3600;
786 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
787 "%02d:%02d ETA", i / 60, i % 60);
789 write(fileno(stderr), buf, strlen(buf));
793 sa.sa_handler = updateprogressmeter;
794 sigemptyset(&sa.sa_mask);
795 sa.sa_flags = SA_RESTART;
796 sigaction(SIGALRM, &sa, NULL);
798 } else if (flag == 1) {
806 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
807 * much of which was blatently stolen from openssh. */
810 * Copyright (c) 1992, 1993
811 * The Regents of the University of California. All rights reserved.
813 * Redistribution and use in source and binary forms, with or without
814 * modification, are permitted provided that the following conditions
816 * 1. Redistributions of source code must retain the above copyright
817 * notice, this list of conditions and the following disclaimer.
818 * 2. Redistributions in binary form must reproduce the above copyright
819 * notice, this list of conditions and the following disclaimer in the
820 * documentation and/or other materials provided with the distribution.
822 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
823 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
825 * 4. Neither the name of the University nor the names of its contributors
826 * may be used to endorse or promote products derived from this software
827 * without specific prior written permission.
829 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
830 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
831 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
832 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
833 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
834 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
835 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
836 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
837 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
838 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
841 * $Id: wget.c,v 1.60 2003/09/15 08:33:37 andersen Exp $
848 c-file-style: "linux"