1 /* vi: set sw=4 ts=4: */
3 * wget - retrieve a file using HTTP or FTP
5 * Chip Rosenthal Covad Communications <chip@laserlink.net>
18 #include <sys/ioctl.h>
21 #include <sys/types.h>
23 #include <sys/socket.h>
24 #include <netinet/in.h>
25 #include <arpa/inet.h>
40 static void parse_url(char *url, struct host_info *h);
41 static FILE *open_socket(struct sockaddr_in *s_in);
42 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
43 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf);
45 /* Globals (can be accessed from signal handlers */
46 static off_t filesize = 0; /* content-length of the file */
47 static int chunked = 0; /* chunked transfer encoding */
48 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
49 static void progressmeter(int flag);
50 static char *curfile; /* Name of current file being transferred. */
51 static struct timeval start; /* Time a transfer started. */
52 static volatile unsigned long statbytes = 0; /* Number of bytes transferred so far. */
53 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
54 static const int STALLTIME = 5;
57 static void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
59 if (output != stdout && do_continue==0) {
65 /* Read NMEMB elements of SIZE bytes into PTR from STREAM. Returns the
66 * number of elements read, and a short count if an eof or non-interrupt
67 * error is encountered. */
68 static size_t safe_fread(void *ptr, size_t size, size_t nmemb, FILE *stream)
74 ret += fread((char *)ptr + (ret * size), size, nmemb - ret, stream);
75 } while (ret < nmemb && ferror(stream) && errno == EINTR);
80 /* Write NMEMB elements of SIZE bytes from PTR to STREAM. Returns the
81 * number of elements written, and a short count if an eof or non-interrupt
82 * error is encountered. */
83 static size_t safe_fwrite(void *ptr, size_t size, size_t nmemb, FILE *stream)
89 ret += fwrite((char *)ptr + (ret * size), size, nmemb - ret, stream);
90 } while (ret < nmemb && ferror(stream) && errno == EINTR);
95 /* Read a line or SIZE - 1 bytes into S, whichever is less, from STREAM.
96 * Returns S, or NULL if an eof or non-interrupt error is encountered. */
97 static char *safe_fgets(char *s, int size, FILE *stream)
103 ret = fgets(s, size, stream);
104 } while (ret == NULL && ferror(stream) && errno == EINTR);
109 #define close_delete_and_die(s...) { \
110 close_and_delete_outfile(output, fname_out, do_continue); \
111 bb_error_msg_and_die(s); }
114 #ifdef CONFIG_FEATURE_WGET_AUTHENTICATION
116 * Base64-encode character string
117 * oops... isn't something similar in uuencode.c?
118 * XXX: It would be better to use already existing code
120 static char *base64enc(unsigned char *p, char *buf, int len) {
122 char al[] = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
128 bb_error_msg_and_die("buffer overflow");
129 *(s++) = al[(*p >> 2) & 0x3F];
130 *(s++) = al[((*p << 4) & 0x30) | ((*(p+1) >> 4) & 0x0F)];
134 *(s++) = al[((*p << 2) & 0x3C) | ((*(p+1) >> 6) & 0x03)];
136 *(s++) = al[*(p++) & 0x3F];
143 #define WGET_OPT_CONTINUE 1
144 #define WGET_OPT_QUIET 2
145 #define WGET_OPT_PASSIVE 4
146 #define WGET_OPT_OUTNAME 8
147 #define WGET_OPT_HEADER 16
148 #define WGET_OPT_PREFIX 32
149 #define WGET_OPT_PROXY 64
151 static const struct option wget_long_options[] = {
152 { "continue", 0, NULL, 'c' },
153 { "quiet", 0, NULL, 'q' },
154 { "passive-ftp", 0, NULL, 139 },
155 { "output-document", 1, NULL, 'O' },
156 { "header", 1, NULL, 131 },
157 { "directory-prefix",1, NULL, 'P' },
158 { "proxy", 1, NULL, 'Y' },
162 int wget_main(int argc, char **argv)
164 int n, try=5, status;
168 char *dir_prefix=NULL;
171 char extra_headers[1024];
172 char *extra_headers_ptr = extra_headers;
173 int extra_headers_left = sizeof(extra_headers);
174 struct host_info server, target;
175 struct sockaddr_in s_in;
176 llist_t *headers_llist = NULL;
178 FILE *sfp = NULL; /* socket to web/ftp server */
179 FILE *dfp = NULL; /* socket to ftp server (data) */
180 char *fname_out = NULL; /* where to direct output (-O) */
181 int do_continue = 0; /* continue a prev transfer (-c) */
182 long beg_range = 0L; /* range at which continue begins */
183 int got_clen = 0; /* got content-length: from server */
184 FILE *output; /* socket to web server */
185 int quiet_flag = FALSE; /* Be verry, verry quiet... */
186 int use_proxy = 1; /* Use proxies if env vars are set */
187 char *proxy_flag = "on"; /* Use proxies if env vars are set */
190 * Crack command line.
192 bb_opt_complementally = "-1:\203::";
193 bb_applet_long_options = wget_long_options;
194 opt = bb_getopt_ulflags(argc, argv, "cq\213O:\203:P:Y:",
195 &fname_out, &headers_llist,
196 &dir_prefix, &proxy_flag);
197 if (opt & WGET_OPT_CONTINUE) {
200 if (opt & WGET_OPT_QUIET) {
203 if (strcmp(proxy_flag, "off") == 0) {
204 /* Use the proxy if necessary. */
207 if (opt & WGET_OPT_HEADER) {
208 while (headers_llist) {
209 int arglen = strlen(headers_llist->data);
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, headers_llist->data);
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;
218 headers_llist = headers_llist->link;
222 parse_url(argv[optind], &target);
223 server.host = target.host;
224 server.port = target.port;
227 * Use the proxy if necessary.
230 proxy = getenv(target.is_ftp ? "ftp_proxy" : "http_proxy");
231 if (proxy && *proxy) {
232 parse_url(bb_xstrdup(proxy), &server);
238 /* Guess an output filename */
240 // Dirty hack. Needed because bb_get_last_path_component
241 // will destroy trailing / by storing '\0' in last byte!
242 if(*target.path && target.path[strlen(target.path)-1]!='/') {
244 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
247 bb_get_last_path_component(target.path);
249 if (fname_out==NULL || strlen(fname_out)<1) {
251 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
256 if (dir_prefix != NULL)
257 fname_out = concat_path_file(dir_prefix, fname_out);
258 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
260 curfile = bb_get_last_path_component(fname_out);
263 if (do_continue && !fname_out)
264 bb_error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
268 * Open the output file stream.
270 if (strcmp(fname_out, "-") == 0) {
274 output = bb_xfopen(fname_out, (do_continue ? "a" : "w"));
278 * Determine where to start transfer.
281 if (fstat(fileno(output), &sbuf) < 0)
282 bb_perror_msg_and_die("fstat()");
283 if (sbuf.st_size > 0)
284 beg_range = sbuf.st_size;
289 /* We want to do exactly _one_ DNS lookup, since some
290 * sites (i.e. ftp.us.debian.org) use round-robin DNS
291 * and we want to connect to only one IP... */
292 bb_lookup_host(&s_in, server.host);
293 s_in.sin_port = server.port;
294 if (quiet_flag==FALSE) {
295 fprintf(stdout, "Connecting to %s[%s]:%d\n",
296 server.host, inet_ntoa(s_in.sin_addr), ntohs(server.port));
299 if (use_proxy || !target.is_ftp) {
304 got_clen = chunked = 0;
307 close_delete_and_die("too many redirections");
310 * Open socket to http server
312 if (sfp) fclose(sfp);
313 sfp = open_socket(&s_in);
319 const char *format = "GET %stp://%s:%d/%s HTTP/1.1\r\n";
320 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
321 if (strchr (target.host, ':'))
322 format = "GET %stp://[%s]:%d/%s HTTP/1.1\r\n";
325 target.is_ftp ? "f" : "ht", target.host,
326 ntohs(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((unsigned char*)target.user, buf, sizeof(buf)));
338 if (use_proxy && server.user) {
339 fprintf(sfp, "Proxy-Authorization: Basic %s\r\n",
340 base64enc((unsigned char*)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.
354 if (fgets(buf, sizeof(buf), sfp) == NULL)
355 close_delete_and_die("no response from server");
357 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
359 for ( ; isspace(*s) ; ++s)
361 switch (status = atoi(s)) {
364 while (gethdr(buf, sizeof(buf), sfp, &n) != NULL);
367 if (do_continue && output != stdout)
368 output = freopen(fname_out, "w", output);
371 case 300: /* redirection */
382 close_delete_and_die("server returned error %d: %s", atoi(s), buf);
386 * Retrieve HTTP headers.
388 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
389 if (strcasecmp(buf, "content-length") == 0) {
391 if (safe_strtoul(s, &value)) {
392 close_delete_and_die("content-length %s is garbage", s);
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);
410 if (use_proxy == 0) {
411 server.host = target.host;
412 server.port = target.port;
414 bb_lookup_host(&s_in, server.host);
415 s_in.sin_port = server.port;
420 } while(status >= 300);
430 target.user = bb_xstrdup("anonymous:busybox@");
432 sfp = open_socket(&s_in);
433 if (ftpcmd(NULL, NULL, sfp, buf) != 220)
434 close_delete_and_die("%s", buf+4);
437 * Splitting username:password pair,
440 s = strchr(target.user, ':');
443 switch(ftpcmd("USER ", target.user, sfp, buf)) {
447 if (ftpcmd("PASS ", s, sfp, buf) == 230)
449 /* FALLTHRU (failed login) */
451 close_delete_and_die("ftp login: %s", buf+4);
454 ftpcmd("CDUP", NULL, sfp, buf);
455 ftpcmd("TYPE I", NULL, sfp, buf);
460 if (ftpcmd("SIZE /", target.path, sfp, buf) == 213) {
462 if (safe_strtoul(buf+4, &value)) {
463 close_delete_and_die("SIZE value is garbage");
470 * Entering passive mode
472 if (ftpcmd("PASV", NULL, sfp, buf) != 227)
473 close_delete_and_die("PASV: %s", buf+4);
474 s = strrchr(buf, ',');
477 s = strrchr(buf, ',');
478 port += atoi(s+1) * 256;
479 s_in.sin_port = htons(port);
480 dfp = open_socket(&s_in);
483 sprintf(buf, "REST %ld", beg_range);
484 if (ftpcmd(buf, NULL, sfp, buf) != 350) {
485 if (output != stdout)
486 output = freopen(fname_out, "w", output);
489 filesize -= beg_range;
492 if (ftpcmd("RETR /", target.path, sfp, buf) > 150)
493 close_delete_and_die("RETR: %s", buf+4);
502 fgets(buf, sizeof(buf), dfp);
503 filesize = strtol(buf, (char **) NULL, 16);
505 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
506 if (quiet_flag==FALSE)
510 while ((filesize > 0 || !got_clen) && (n = safe_fread(buf, 1, ((chunked || got_clen) && (filesize < sizeof(buf)) ? filesize : sizeof(buf)), dfp)) > 0) {
511 if (safe_fwrite(buf, 1, n, output) != n) {
512 bb_perror_msg_and_die("write error");
514 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
523 safe_fgets(buf, sizeof(buf), dfp); /* This is a newline */
524 safe_fgets(buf, sizeof(buf), dfp);
525 filesize = strtol(buf, (char **) NULL, 16);
527 chunked = 0; /* all done! */
531 if (n == 0 && ferror(dfp)) {
532 bb_perror_msg_and_die("network read error");
535 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
536 if (quiet_flag==FALSE)
539 if ((use_proxy == 0) && target.is_ftp) {
541 if (ftpcmd(NULL, NULL, sfp, buf) != 226)
542 bb_error_msg_and_die("ftp error: %s", buf+4);
543 ftpcmd("QUIT", NULL, sfp, buf);
549 void parse_url(char *url, struct host_info *h)
551 char *cp, *sp, *up, *pp;
553 if (strncmp(url, "http://", 7) == 0) {
554 h->port = bb_lookup_port("http", "tcp", 80);
557 } else if (strncmp(url, "ftp://", 6) == 0) {
558 h->port = bb_lookup_port("ftp", "tfp", 21);
562 bb_error_msg_and_die("not an http or ftp url: %s", url);
564 sp = strchr(h->host, '/');
569 h->path = bb_xstrdup("");
571 up = strrchr(h->host, '@');
581 #ifdef CONFIG_FEATURE_WGET_IP6_LITERAL
582 if (h->host[0] == '[') {
586 while (*ep == ':' || isxdigit (*ep))
596 cp = strchr(pp, ':');
599 h->port = htons(atoi(cp));
604 FILE *open_socket(struct sockaddr_in *s_in)
608 fp = fdopen(xconnect(s_in), "r+");
610 bb_perror_msg_and_die("fdopen()");
616 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
623 /* retrieve header line */
624 if (fgets(buf, bufsiz, fp) == NULL)
627 /* see if we are at the end of the headers */
628 for (s = buf ; *s == '\r' ; ++s)
633 /* convert the header name to lower case */
634 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
637 /* verify we are at the end of the header name */
639 bb_error_msg_and_die("bad header line: %s", buf);
641 /* locate the start of the header value */
642 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
646 /* locate the end of header */
647 while (*s != '\0' && *s != '\r' && *s != '\n')
650 /* end of header found */
656 /* Rats! The buffer isn't big enough to hold the entire header value. */
657 while (c = getc(fp), c != EOF && c != '\n')
663 static int ftpcmd(char *s1, char *s2, FILE *fp, char *buf)
667 fprintf(fp, "%s%s\r\n", s1, s2);
674 if (fgets(buf, 510, fp) == NULL) {
675 bb_perror_msg_and_die("fgets()");
677 buf_ptr = strstr(buf, "\r\n");
681 } while (! isdigit(buf[0]) || buf[3] != ' ');
686 #ifdef CONFIG_FEATURE_WGET_STATUSBAR
687 /* Stuff below is from BSD rcp util.c, as added to openshh.
688 * Original copyright notice is retained at the end of this file.
697 get_terminal_width_height(0, &width, NULL);
702 updateprogressmeter(int ignore)
704 int save_errno = errno;
713 struct itimerval itv;
715 itv.it_value.tv_sec = wait;
716 itv.it_value.tv_usec = 0;
717 itv.it_interval = itv.it_value;
718 setitimer(ITIMER_REAL, &itv, NULL);
723 progressmeter(int flag)
725 static const char prefixes[] = " KMGTP";
726 static struct timeval lastupdate;
727 static off_t lastsize, totalsize;
728 struct timeval now, td, wait;
729 off_t cursize, abbrevsize;
731 int ratio, barlength, i, remaining;
735 (void) gettimeofday(&start, (struct timezone *) 0);
738 totalsize = filesize; /* as filesize changes.. */
741 (void) gettimeofday(&now, (struct timezone *) 0);
743 if (totalsize != 0 && !chunked) {
744 ratio = 100.0 * cursize / totalsize;
745 ratio = MAX(ratio, 0);
746 ratio = MIN(ratio, 100);
750 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
752 barlength = getttywidth() - 51;
754 i = barlength * ratio / 100;
755 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
757 "*****************************************************************************"
758 "*****************************************************************************",
762 abbrevsize = cursize;
763 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
767 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
768 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
771 timersub(&now, &lastupdate, &wait);
772 if (cursize > lastsize) {
775 if (wait.tv_sec >= STALLTIME) {
776 start.tv_sec += wait.tv_sec;
777 start.tv_usec += wait.tv_usec;
781 timersub(&now, &start, &td);
782 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
784 if (wait.tv_sec >= STALLTIME) {
785 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
787 } else if (statbytes <= 0 || elapsed <= 0.0 || cursize > totalsize || chunked) {
788 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
791 remaining = (int) (totalsize / (statbytes / elapsed) - elapsed);
792 i = remaining / 3600;
794 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
797 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
799 i = remaining % 3600;
800 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
801 "%02d:%02d ETA", i / 60, i % 60);
803 write(STDERR_FILENO, buf, strlen(buf));
807 sa.sa_handler = updateprogressmeter;
808 sigemptyset(&sa.sa_mask);
809 sa.sa_flags = SA_RESTART;
810 sigaction(SIGALRM, &sa, NULL);
812 } else if (flag == 1) {
820 /* Original copyright notice which applies to the CONFIG_FEATURE_WGET_STATUSBAR stuff,
821 * much of which was blatantly stolen from openssh. */
824 * Copyright (c) 1992, 1993
825 * The Regents of the University of California. All rights reserved.
827 * Redistribution and use in source and binary forms, with or without
828 * modification, are permitted provided that the following conditions
830 * 1. Redistributions of source code must retain the above copyright
831 * notice, this list of conditions and the following disclaimer.
832 * 2. Redistributions in binary form must reproduce the above copyright
833 * notice, this list of conditions and the following disclaimer in the
834 * documentation and/or other materials provided with the distribution.
836 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
837 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
839 * 4. Neither the name of the University nor the names of its contributors
840 * may be used to endorse or promote products derived from this software
841 * without specific prior written permission.
843 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
844 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
845 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
846 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
847 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
848 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
849 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
850 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
851 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
852 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
855 * $Id: wget.c,v 1.75 2004/10/08 08:27:40 andersen Exp $