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>
39 static void parse_url(char *url, struct host_info *h);
40 static FILE *open_socket(struct sockaddr_in *s_in);
41 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
42 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf);
44 /* Globals (can be accessed from signal handlers */
45 static off_t filesize = 0; /* content-length of the file */
46 static int chunked = 0; /* chunked transfer encoding */
47 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
48 static void progressmeter(int flag);
49 static char *curfile; /* Name of current file being transferred. */
50 static struct timeval start; /* Time a transfer started. */
51 static volatile unsigned long statbytes = 0; /* Number of bytes transferred so far. */
52 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
53 static const int STALLTIME = 5;
56 static void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
58 if (output != stdout && do_continue==0) {
64 /* Read NMEMB elements of SIZE bytes into PTR from STREAM. Returns the
65 * number of elements read, and a short count if an eof or non-interrupt
66 * error is encountered. */
67 static size_t safe_fread(void *ptr, size_t size, size_t nmemb, FILE *stream)
73 ret += fread((char *)ptr + (ret * size), size, nmemb - ret, stream);
74 } while (ret < nmemb && ferror(stream) && errno == EINTR);
79 /* Write NMEMB elements of SIZE bytes from PTR to STREAM. Returns the
80 * number of elements written, and a short count if an eof or non-interrupt
81 * error is encountered. */
82 static size_t safe_fwrite(void *ptr, size_t size, size_t nmemb, FILE *stream)
88 ret += fwrite((char *)ptr + (ret * size), size, nmemb - ret, stream);
89 } while (ret < nmemb && ferror(stream) && errno == EINTR);
94 /* Read a line or SIZE - 1 bytes into S, whichever is less, from STREAM.
95 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
96 static char *safe_fgets(char *s, int size, FILE *stream)
102 ret = fgets(s, size, stream);
103 } while (ret == NULL && ferror(stream) && errno == EINTR);
108 #define close_delete_and_die(s...) { \
109 close_and_delete_outfile(output, fname_out, do_continue); \
110 bb_error_msg_and_die(s); }
113 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
115 * Base64-encode character string
116 * oops... isn't something similar in uuencode.c?
117 * XXX: It would be better to use already existing code
119 static char *base64enc(unsigned char *p, char *buf, int len) {
121 char al[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
127 bb_error_msg_and_die("buffer overflow");
128 *(s++) = al[(*p >> 2) & 0x3F];
129 *(s++) = al[((*p << 4) & 0x30) | ((*(p+1) >> 4) & 0x0F)];
133 *(s++) = al[((*p << 2) & 0x3C) | ((*(p+1) >> 6) & 0x03)];
135 *(s++) = al[*(p++) & 0x3F];
142 #define WGET_OPT_CONTINUE 1
143 #define WGET_OPT_QUIET 2
144 #define WGET_OPT_PASSIVE 4
145 #define WGET_OPT_OUTNAME 8
146 #define WGET_OPT_HEADER 16
147 #define WGET_OPT_PREFIX 32
148 #define WGET_OPT_PROXY 64
150 static const struct option wget_long_options[] = {
151 { "continue", 0, NULL, 'c' },
152 { "quiet", 0, NULL, 'q' },
153 { "passive-ftp", 0, NULL, 139 },
154 { "output-document", 1, NULL, 'O' },
155 { "header", 1, NULL, 131 },
156 { "directory-prefix",1, NULL, 'P' },
157 { "proxy", 1, NULL, 'Y' },
161 int wget_main(int argc, char **argv)
163 int n, try=5, status;
167 char *dir_prefix=NULL;
170 char extra_headers[1024];
171 char *extra_headers_ptr = extra_headers;
172 int extra_headers_left = sizeof(extra_headers);
173 struct host_info server, target;
174 struct sockaddr_in s_in;
175 llist_t *headers_llist = NULL;
177 FILE *sfp = NULL; /* socket to web/ftp server */
178 FILE *dfp = NULL; /* socket to ftp server (data) */
179 char *fname_out = NULL; /* where to direct output (-O) */
180 int do_continue = 0; /* continue a prev transfer (-c) */
181 long beg_range = 0L; /* range at which continue begins */
182 int got_clen = 0; /* got content-length: from server */
183 FILE *output; /* socket to web server */
184 int quiet_flag = FALSE; /* Be verry, verry quiet... */
185 int use_proxy = 1; /* Use proxies if env vars are set */
186 char *proxy_flag = "on"; /* Use proxies if env vars are set */
189 * Crack command line.
191 bb_opt_complementally = "-1:\203::";
192 bb_applet_long_options = wget_long_options;
193 opt = bb_getopt_ulflags(argc, argv, "cq\213O:\203:P:Y:",
194 &fname_out, &headers_llist,
195 &dir_prefix, &proxy_flag);
196 if (opt & WGET_OPT_CONTINUE) {
199 if (opt & WGET_OPT_QUIET) {
202 if (strcmp(proxy_flag, "off") == 0) {
203 /* Use the proxy if necessary. */
206 if (opt & WGET_OPT_HEADER) {
207 while (headers_llist) {
208 int arglen = strlen(headers_llist->data);
209 if (extra_headers_left - arglen - 2 <= 0)
210 bb_error_msg_and_die("extra_headers buffer too small(need %i)", extra_headers_left - arglen);
211 strcpy(extra_headers_ptr, headers_llist->data);
212 extra_headers_ptr += arglen;
213 extra_headers_left -= ( arglen + 2 );
214 *extra_headers_ptr++ = '\r';
215 *extra_headers_ptr++ = '\n';
216 *(extra_headers_ptr + 1) = 0;
217 headers_llist = headers_llist->link;
221 parse_url(argv[optind], &target);
222 server.host = target.host;
223 server.port = target.port;
226 * Use the proxy if necessary.
229 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
230 if (proxy && *proxy) {
231 parse_url(bb_xstrdup(proxy), &server);
237 /* Guess an output filename */
239 // Dirty hack. Needed because bb_get_last_path_component
240 // will destroy trailing / by storing '\0' in last byte!
241 if(*target.path && target.path[strlen(target.path)-1]!='/') {
243 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
246 bb_get_last_path_component(target.path);
248 if (fname_out==NULL || strlen(fname_out)<1) {
250 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
255 if (dir_prefix != NULL)
256 fname_out = concat_path_file(dir_prefix, fname_out);
257 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
259 curfile = bb_get_last_path_component(fname_out);
262 if (do_continue && !fname_out)
263 bb_error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
267 * Open the output file stream.
269 if (strcmp(fname_out, "-") == 0) {
273 output = bb_xfopen(fname_out, (do_continue ? "a" : "w"));
277 * Determine where to start transfer.
280 if (fstat(fileno(output), &sbuf) < 0)
281 bb_perror_msg_and_die("fstat()");
282 if (sbuf.st_size > 0)
283 beg_range = sbuf.st_size;
288 /* We want to do exactly _one_ DNS lookup, since some
289 * sites (i.e. ftp.us.debian.org) use round-robin DNS
290 * and we want to connect to only one IP... */
291 bb_lookup_host(&s_in, server.host);
292 s_in.sin_port = server.port;
293 if (quiet_flag==FALSE) {
294 fprintf(stdout, "Connecting to %s[%s]:%d\n",
295 server.host, inet_ntoa(s_in.sin_addr), ntohs(server.port));
298 if (use_proxy || !target.is_ftp) {
303 got_clen = chunked = 0;
306 close_delete_and_die("too many redirections");
309 * Open socket to http server
311 if (sfp) fclose(sfp);
312 sfp = open_socket(&s_in);
318 const char *format = "GET %stp://%s:%d/%s HTTP/1.1\r\n";
319 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
320 if (strchr (target.host, ':'))
321 format = "GET %stp://[%s]:%d/%s HTTP/1.1\r\n";
324 target.is_ftp ? "f" : "ht", target.host,
325 ntohs(target.port), target.path);
327 fprintf(sfp, "GET /%s HTTP/1.1\r\n", target.path);
330 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", target.host);
332 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
334 fprintf(sfp, "Authorization: Basic %s\r\n",
335 base64enc((unsigned char*)target.user, buf, sizeof(buf)));
337 if (use_proxy && server.user) {
338 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
339 base64enc((unsigned char*)server.user, buf, sizeof(buf)));
344 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
345 if(extra_headers_left < sizeof(extra_headers))
346 fputs(extra_headers,sfp);
347 fprintf(sfp,"Connection: close\r\n\r\n");
350 * Retrieve HTTP response line and check for "200" status code.
353 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) {
390 if (safe_strtoul(s, &value)) {
391 close_delete_and_die("content-length %s is garbage", s);
397 if (strcasecmp(buf, "transfer-encoding") == 0) {
398 if (strcasecmp(s, "chunked") == 0) {
399 chunked = got_clen = 1;
401 close_delete_and_die("server wants to do %s transfer encoding", s);
404 if (strcasecmp(buf, "location") == 0) {
406 target.path = bb_xstrdup(s+1);
408 parse_url(bb_xstrdup(s), &target);
409 if (use_proxy == 0) {
410 server.host = target.host;
411 server.port = target.port;
413 bb_lookup_host(&s_in, server.host);
414 s_in.sin_port = server.port;
419 } while(status >= 300);
429 target.user = bb_xstrdup("anonymous:busybox@");
431 sfp = open_socket(&s_in);
432 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
433 close_delete_and_die("%s", buf+4);
436 * Splitting username:password pair,
439 s = strchr(target.user, ':');
442 switch(ftpcmd("USER ", target.user, sfp, buf)) {
446 if (ftpcmd("PASS ", s, sfp, buf) == 230)
448 /* FALLTHRU (failed login) */
450 close_delete_and_die("ftp login: %s", buf+4);
453 ftpcmd("CDUP", NULL, sfp, buf);
454 ftpcmd("TYPE I", NULL, sfp, buf);
459 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
461 if (safe_strtoul(buf+4, &value)) {
462 close_delete_and_die("SIZE value is garbage");
469 * Entering passive mode
471 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
472 close_delete_and_die("PASV: %s", buf+4);
473 s = strrchr(buf, ',');
476 s = strrchr(buf, ',');
477 port += atoi(s+1) * 256;
478 s_in.sin_port = htons(port);
479 dfp = open_socket(&s_in);
482 sprintf(buf, "REST %ld", beg_range);
483 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
484 if (output != stdout)
485 output = freopen(fname_out, "w", output);
488 filesize -= beg_range;
491 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
492 close_delete_and_die("RETR: %s", buf+4);
501 fgets(buf, sizeof(buf), dfp);
502 filesize = strtol(buf, (char **) NULL, 16);
504 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
505 if (quiet_flag==FALSE)
509 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, ((chunked || got_clen) && (filesize < sizeof(buf)) ? filesize : sizeof(buf)), dfp)) > 0) {
510 if (safe_fwrite(buf, 1, n, output) != n) {
511 bb_perror_msg_and_die("write error");
513 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
522 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
523 safe_fgets(buf, sizeof(buf), dfp);
524 filesize = strtol(buf, (char **) NULL, 16);
526 chunked = 0; /* all done! */
530 if (n == 0 && ferror(dfp)) {
531 bb_perror_msg_and_die("network read error");
534 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
535 if (quiet_flag==FALSE)
538 if ((use_proxy == 0) && target.is_ftp) {
540 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
541 bb_error_msg_and_die("ftp error: %s", buf+4);
542 ftpcmd("QUIT", NULL, sfp, buf);
548 void parse_url(char *url, struct host_info *h)
550 char *cp, *sp, *up, *pp;
552 if (strncmp(url, "http://", 7) == 0) {
553 h->port = bb_lookup_port("http", "tcp", 80);
556 } else if (strncmp(url, "ftp://", 6) == 0) {
557 h->port = bb_lookup_port("ftp", "tfp", 21);
561 bb_error_msg_and_die("not an http or ftp url: %s", url);
563 sp = strchr(h->host, '/');
568 h->path = bb_xstrdup("");
570 up = strrchr(h->host, '@');
580 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
581 if (h->host[0] == '[') {
585 while (*ep == ':' || isxdigit (*ep))
595 cp = strchr(pp, ':');
598 h->port = htons(atoi(cp));
603 FILE *open_socket(struct sockaddr_in *s_in)
607 fp = fdopen(xconnect(s_in), "r+");
609 bb_perror_msg_and_die("fdopen()");
615 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
622 /* retrieve header line */
623 if (fgets(buf, bufsiz, fp) == NULL)
626 /* see if we are at the end of the headers */
627 for (s = buf ; *s == '\r' ; ++s)
632 /* convert the header name to lower case */
633 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
636 /* verify we are at the end of the header name */
638 bb_error_msg_and_die("bad header line: %s", buf);
640 /* locate the start of the header value */
641 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
645 /* locate the end of header */
646 while (*s != '\0' && *s != '\r' && *s != '\n')
649 /* end of header found */
655 /* Rats! The buffer isn't big enough to hold the entire header value. */
656 while (c = getc(fp), c != EOF && c != '\n')
662 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
666 fprintf(fp, "%s%s\r\n", s1, s2);
673 if (fgets(buf, 510, fp) == NULL) {
674 bb_perror_msg_and_die("fgets()");
676 buf_ptr = strstr(buf, "\r\n");
680 } while (! isdigit(buf[0]) || buf[3] != ' ');
685 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
686 /* Stuff below is from BSD rcp util.c, as added to openshh.
687 * Original copyright notice is retained at the end of this file.
696 get_terminal_width_height(0, &width, NULL);
701 updateprogressmeter(int ignore)
703 int save_errno = errno;
712 struct itimerval itv;
714 itv.it_value.tv_sec = wait;
715 itv.it_value.tv_usec = 0;
716 itv.it_interval = itv.it_value;
717 setitimer(ITIMER_REAL, &itv, NULL);
722 progressmeter(int flag)
724 static const char prefixes[] = " KMGTP";
725 static struct timeval lastupdate;
726 static off_t lastsize, totalsize;
727 struct timeval now, td, wait;
728 off_t cursize, abbrevsize;
730 int ratio, barlength, i, remaining;
734 (void) gettimeofday(&start, (struct timezone *) 0);
737 totalsize = filesize; /* as filesize changes.. */
740 (void) gettimeofday(&now, (struct timezone *) 0);
742 if (totalsize != 0 && !chunked) {
743 ratio = 100.0 * cursize / totalsize;
744 ratio = MAX(ratio, 0);
745 ratio = MIN(ratio, 100);
749 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
751 barlength = getttywidth() - 51;
753 i = barlength * ratio / 100;
754 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
756 "*****************************************************************************"
757 "*****************************************************************************",
761 abbrevsize = cursize;
762 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
766 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
767 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
770 timersub(&now, &lastupdate, &wait);
771 if (cursize > lastsize) {
774 if (wait.tv_sec >= STALLTIME) {
775 start.tv_sec += wait.tv_sec;
776 start.tv_usec += wait.tv_usec;
780 timersub(&now, &start, &td);
781 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
783 if (wait.tv_sec >= STALLTIME) {
784 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
786 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
787 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
790 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
791 i = remaining / 3600;
793 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
796 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
798 i = remaining % 3600;
799 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
800 "%02d:%02d ETA", i / 60, i % 60);
802 write(STDERR_FILENO, buf, strlen(buf));
806 sa.sa_handler = updateprogressmeter;
807 sigemptyset(&sa.sa_mask);
808 sa.sa_flags = SA_RESTART;
809 sigaction(SIGALRM, &sa, NULL);
811 } else if (flag == 1) {
819 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
820 * much of which was blatantly stolen from openssh. */
823 * Copyright (c) 1992, 1993
824 * The Regents of the University of California. All rights reserved.
826 * Redistribution and use in source and binary forms, with or without
827 * modification, are permitted provided that the following conditions
829 * 1. Redistributions of source code must retain the above copyright
830 * notice, this list of conditions and the following disclaimer.
831 * 2. Redistributions in binary form must reproduce the above copyright
832 * notice, this list of conditions and the following disclaimer in the
833 * documentation and/or other materials provided with the distribution.
835 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
836 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
838 * 4. Neither the name of the University nor the names of its contributors
839 * may be used to endorse or promote products derived from this software
840 * without specific prior written permission.
842 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
843 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
844 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
845 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
846 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
847 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
848 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
849 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
850 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
851 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
854 * $Id: wget.c,v 1.75 2004/10/08 08:27:40 andersen Exp $
861 c-file-style: "linux"