1 /* vi: set sw=4 ts=4: */
3 * wget - retrieve a file using HTTP
5 * Chip Rosenthal Covad Communications <chip@laserlink.net>
7 * Note: According to RFC2616 section 3.6.1, "All HTTP/1.1 applications MUST be
8 * able to receive and decode the "chunked" transfer-coding, and MUST ignore
9 * chunk-extension extensions they do not understand."
11 * This prevents this particular wget app from completely RFC compliant, and as
12 * such, prevents it from being used as a general purpose web browser... This
13 * is a design decision, since it makes the code smaller.
25 #include <sys/ioctl.h>
28 #include <sys/types.h>
30 #include <sys/socket.h>
31 #include <netinet/in.h>
32 #include <arpa/inet.h>
36 void parse_url(char *url, char **uri_host, int *uri_port, char **uri_path);
37 FILE *open_socket(char *host, int port);
38 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
39 void progressmeter(int flag);
41 /* Globals (can be accessed from signal handlers */
42 static off_t filesize = 0; /* content-length of the file */
43 #ifdef BB_FEATURE_WGET_STATUSBAR
44 static char *curfile; /* Name of current file being transferred. */
45 static struct timeval start; /* Time a transfer started. */
46 volatile unsigned long statbytes; /* Number of bytes transferred so far. */
47 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
48 static const int STALLTIME = 5;
51 void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
53 if (output != stdout && do_continue==0) {
59 int wget_main(int argc, char **argv)
62 char *proxy, *proxy_host;
63 int uri_port, proxy_port;
67 FILE *sfp; /* socket to web server */
68 char *uri_host, *uri_path; /* parsed from command line url */
69 char *fname_out = NULL; /* where to direct output (-O) */
70 int do_continue = 0; /* continue a prev transfer (-c) */
71 long beg_range = 0L; /* range at which continue begins */
72 int got_clen = 0; /* got content-length: from server */
73 FILE *output; /* socket to web server */
74 int quiet_flag = FALSE; /* Be verry, verry quiet... */
79 while ((n = getopt(argc, argv, "cqO:")) != EOF) {
88 /* can't set fname_out to NULL if outputting to stdout, because
89 * this gets interpreted as the auto-gen output filename
90 * case below - tausq@debian.org
92 fname_out = (strcmp(optarg, "-") == 0 ? (char *)1 : optarg);
99 if (argc - optind != 1)
103 * Use the proxy if necessary.
105 if ((proxy = getenv("http_proxy")) != NULL) {
106 proxy = xstrdup(proxy);
107 parse_url(proxy, &proxy_host, &proxy_port, &uri_path);
108 parse_url(argv[optind], &uri_host, &uri_port, &uri_path);
111 * Parse url into components.
113 parse_url(argv[optind], &uri_host, &uri_port, &uri_path);
118 /* Guess an output filename */
121 #ifdef BB_FEATURE_WGET_STATUSBAR
124 get_last_path_component(uri_path);
125 if (fname_out==NULL || strlen(fname_out)<1) {
127 #ifdef BB_FEATURE_WGET_STATUSBAR
132 #ifdef BB_FEATURE_WGET_STATUSBAR
134 curfile=argv[optind];
137 if (do_continue && !fname_out)
138 error_msg_and_die("cannot specify continue (-c) without a filename (-O)\n");
142 * Open socket to server.
144 sfp = open_socket(proxy_host, proxy_port);
146 /* Make the assumption that if the file already exists
147 * on disk that the intention is to continue downloading
148 * a previously aborted download -Erik */
149 if (stat(fname_out, &sbuf) == 0) {
154 * Open the output file stream.
156 if (fname_out != (char *)1) {
157 if ( (output=fopen(fname_out, (do_continue ? "a" : "w"))) == NULL)
158 perror_msg_and_die("fopen(%s)", fname_out);
164 * Determine where to start transfer.
167 if (fstat(fileno(output), &sbuf) < 0)
168 perror_msg_and_die("fstat()");
169 if (sbuf.st_size > 0)
170 beg_range = sbuf.st_size;
178 fprintf(sfp, "GET http://%s:%d/%s HTTP/1.1\r\n",
179 uri_host, uri_port, uri_path);
180 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", uri_host);
183 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
184 fprintf(sfp,"Connection: close\r\n\r\n");
187 * Retrieve HTTP response line and check for "200" status code.
189 if (fgets(buf, sizeof(buf), sfp) == NULL) {
190 close_and_delete_outfile(output, fname_out, do_continue);
191 error_msg_and_die("no response from server\n");
193 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
195 for ( ; isspace(*s) ; ++s)
206 close_and_delete_outfile(output, fname_out, do_continue);
207 error_msg_and_die("server returned error %d: %s", atoi(s), buf);
211 * Retrieve HTTP headers.
213 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
214 if (strcasecmp(buf, "content-length") == 0) {
219 if (strcasecmp(buf, "transfer-encoding") == 0) {
220 close_and_delete_outfile(output, fname_out, do_continue);
221 error_msg_and_die("server wants to do %s transfer encoding\n", s);
227 * Retrieve HTTP body.
229 #ifdef BB_FEATURE_WGET_STATUSBAR
231 if (quiet_flag==FALSE)
234 while (filesize > 0 && (n = fread(buf, 1, sizeof(buf), sfp)) > 0) {
235 fwrite(buf, 1, n, output);
236 #ifdef BB_FEATURE_WGET_STATUSBAR
238 if (quiet_flag==FALSE)
244 if (n == 0 && ferror(sfp))
245 perror_msg_and_die("network read error");
251 void parse_url(char *url, char **uri_host, int *uri_port, char **uri_path)
257 if (strncmp(url, "http://", 7) != 0)
258 error_msg_and_die("not an http url: %s\n", url);
262 cp = strchr(*uri_host, ':');
263 sp = strchr(*uri_host, '/');
265 if (cp != NULL && (sp == NULL || cp < sp)) {
267 *uri_port = atoi(cp);
278 FILE *open_socket(char *host, int port)
280 struct sockaddr_in sin;
285 memzero(&sin, sizeof(sin));
286 sin.sin_family = AF_INET;
287 if ((hp = (struct hostent *) gethostbyname(host)) == NULL)
288 error_msg_and_die("cannot resolve %s\n", host);
289 memcpy(&sin.sin_addr, hp->h_addr_list[0], hp->h_length);
290 sin.sin_port = htons(port);
293 * Get the server onto a stdio stream.
295 if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0)
296 perror_msg_and_die("socket()");
297 if (connect(fd, (struct sockaddr *) &sin, sizeof(sin)) < 0)
298 perror_msg_and_die("connect(%s)", host);
299 if ((fp = fdopen(fd, "r+")) == NULL)
300 perror_msg_and_die("fdopen()");
306 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
313 /* retrieve header line */
314 if (fgets(buf, bufsiz, fp) == NULL)
317 /* see if we are at the end of the headers */
318 for (s = buf ; *s == '\r' ; ++s)
323 /* convert the header name to lower case */
324 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
327 /* verify we are at the end of the header name */
329 error_msg_and_die("bad header line: %s\n", buf);
331 /* locate the start of the header value */
332 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
336 /* locate the end of header */
337 while (*s != '\0' && *s != '\r' && *s != '\n')
340 /* end of header found */
346 /* Rats! The buffer isn't big enough to hold the entire header value. */
347 while (c = getc(fp), c != EOF && c != '\n')
353 #ifdef BB_FEATURE_WGET_STATUSBAR
354 /* Stuff below is from BSD rcp util.c, as added to openshh.
355 * Original copyright notice is retained at the end of this file.
363 struct winsize winsize;
365 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
366 return (winsize.ws_col ? winsize.ws_col : 80);
372 updateprogressmeter(int ignore)
374 int save_errno = errno;
383 struct itimerval itv;
385 itv.it_value.tv_sec = wait;
386 itv.it_value.tv_usec = 0;
387 itv.it_interval = itv.it_value;
388 setitimer(ITIMER_REAL, &itv, NULL);
393 progressmeter(int flag)
395 static const char prefixes[] = " KMGTP";
396 static struct timeval lastupdate;
397 static off_t lastsize;
398 struct timeval now, td, wait;
399 off_t cursize, abbrevsize;
401 int ratio, barlength, i, remaining;
405 (void) gettimeofday(&start, (struct timezone *) 0);
410 (void) gettimeofday(&now, (struct timezone *) 0);
413 ratio = 100.0 * cursize / filesize;
414 ratio = MAX(ratio, 0);
415 ratio = MIN(ratio, 100);
419 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
421 barlength = getttywidth() - 51;
423 i = barlength * ratio / 100;
424 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
426 "*****************************************************************************"
427 "*****************************************************************************",
431 abbrevsize = cursize;
432 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
436 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
437 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
440 timersub(&now, &lastupdate, &wait);
441 if (cursize > lastsize) {
444 if (wait.tv_sec >= STALLTIME) {
445 start.tv_sec += wait.tv_sec;
446 start.tv_usec += wait.tv_usec;
450 timersub(&now, &start, &td);
451 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
453 if (statbytes <= 0 || elapsed <= 0.0 || cursize > filesize) {
454 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
456 } else if (wait.tv_sec >= STALLTIME) {
457 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
460 remaining = (int) (filesize / (statbytes / elapsed) - elapsed);
461 i = remaining / 3600;
463 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
466 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
468 i = remaining % 3600;
469 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
470 "%02d:%02d ETA", i / 60, i % 60);
472 write(fileno(stderr), buf, strlen(buf));
476 sa.sa_handler = updateprogressmeter;
477 sigemptyset(&sa.sa_mask);
478 sa.sa_flags = SA_RESTART;
479 sigaction(SIGALRM, &sa, NULL);
481 } else if (flag == 1) {
488 /* Original copyright notice which applies to the BB_FEATURE_WGET_STATUSBAR stuff,
489 * much of which was blatently stolen from openssh. */
492 * Copyright (c) 1992, 1993
493 * The Regents of the University of California. All rights reserved.
495 * Redistribution and use in source and binary forms, with or without
496 * modification, are permitted provided that the following conditions
498 * 1. Redistributions of source code must retain the above copyright
499 * notice, this list of conditions and the following disclaimer.
500 * 2. Redistributions in binary form must reproduce the above copyright
501 * notice, this list of conditions and the following disclaimer in the
502 * documentation and/or other materials provided with the distribution.
504 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
505 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
507 * 4. Neither the name of the University nor the names of its contributors
508 * may be used to endorse or promote products derived from this software
509 * without specific prior written permission.
511 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
512 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
513 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
514 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
515 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
516 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
517 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
518 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
519 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
520 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
523 * $Id: wget.c,v 1.21 2001/01/24 20:28:35 andersen Exp $
530 c-file-style: "linux"