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>
37 /* Stupid libc5 doesn't define this... */
39 #define timersub(a, b, result) \
41 (result)->tv_sec = (a)->tv_sec - (b)->tv_sec; \
42 (result)->tv_usec = (a)->tv_usec - (b)->tv_usec; \
43 if ((result)->tv_usec < 0) { \
45 (result)->tv_usec += 1000000; \
50 static void parse_url(char *url, char **uri_host, int *uri_port, char **uri_path);
51 static FILE *open_socket(char *host, int port);
52 static char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
53 static void progressmeter(int flag);
55 /* Globals (can be accessed from signal handlers */
56 static off_t filesize = 0; /* content-length of the file */
57 #ifdef BB_FEATURE_WGET_STATUSBAR
58 static char *curfile; /* Name of current file being transferred. */
59 static struct timeval start; /* Time a transfer started. */
60 static volatile unsigned long statbytes; /* Number of bytes transferred so far. */
61 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
62 static const int STALLTIME = 5;
65 static void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
67 if (output != stdout && do_continue==0) {
73 int wget_main(int argc, char **argv)
76 char *proxy, *proxy_host;
77 int uri_port, proxy_port;
81 FILE *sfp; /* socket to web server */
82 char *uri_host, *uri_path; /* parsed from command line url */
83 char *fname_out = NULL; /* where to direct output (-O) */
84 int do_continue = 0; /* continue a prev transfer (-c) */
85 long beg_range = 0L; /* range at which continue begins */
86 int got_clen = 0; /* got content-length: from server */
87 FILE *output; /* socket to web server */
88 int quiet_flag = FALSE; /* Be verry, verry quiet... */
93 while ((n = getopt(argc, argv, "cqO:")) != EOF) {
102 /* can't set fname_out to NULL if outputting to stdout, because
103 * this gets interpreted as the auto-gen output filename
104 * case below - tausq@debian.org
106 fname_out = (strcmp(optarg, "-") == 0 ? (char *)1 : optarg);
113 if (argc - optind != 1)
117 * Use the proxy if necessary.
119 if ((proxy = getenv("http_proxy")) != NULL) {
120 proxy = xstrdup(proxy);
121 parse_url(proxy, &proxy_host, &proxy_port, &uri_path);
122 parse_url(argv[optind], &uri_host, &uri_port, &uri_path);
125 * Parse url into components.
127 parse_url(argv[optind], &uri_host, &uri_port, &uri_path);
132 /* Guess an output filename */
135 #ifdef BB_FEATURE_WGET_STATUSBAR
138 get_last_path_component(uri_path);
139 if (fname_out==NULL || strlen(fname_out)<1) {
141 #ifdef BB_FEATURE_WGET_STATUSBAR
146 #ifdef BB_FEATURE_WGET_STATUSBAR
148 curfile=argv[optind];
151 if (do_continue && !fname_out)
152 error_msg_and_die("cannot specify continue (-c) without a filename (-O)");
156 * Open socket to server.
158 sfp = open_socket(proxy_host, proxy_port);
160 /* Make the assumption that if the file already exists
161 * on disk that the intention is to continue downloading
162 * a previously aborted download -Erik */
163 if (stat(fname_out, &sbuf) == 0) {
168 * Open the output file stream.
170 if (fname_out != (char *)1) {
171 output = xfopen( fname_out, (do_continue ? "a" : "w") );
177 * Determine where to start transfer.
180 if (fstat(fileno(output), &sbuf) < 0)
181 perror_msg_and_die("fstat()");
182 if (sbuf.st_size > 0)
183 beg_range = sbuf.st_size;
191 fprintf(sfp, "GET http://%s:%d/%s HTTP/1.1\r\n",
192 uri_host, uri_port, uri_path);
193 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", uri_host);
196 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
197 fprintf(sfp,"Connection: close\r\n\r\n");
200 * Retrieve HTTP response line and check for "200" status code.
202 if (fgets(buf, sizeof(buf), sfp) == NULL) {
203 close_and_delete_outfile(output, fname_out, do_continue);
204 error_msg_and_die("no response from server");
206 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
208 for ( ; isspace(*s) ; ++s)
219 close_and_delete_outfile(output, fname_out, do_continue);
221 error_msg_and_die("server returned error %d: %s", atoi(s), buf);
225 * Retrieve HTTP headers.
227 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
228 if (strcasecmp(buf, "content-length") == 0) {
233 if (strcasecmp(buf, "transfer-encoding") == 0) {
234 close_and_delete_outfile(output, fname_out, do_continue);
235 error_msg_and_die("server wants to do %s transfer encoding", s);
241 * Retrieve HTTP body.
243 #ifdef BB_FEATURE_WGET_STATUSBAR
245 if (quiet_flag==FALSE)
248 while (filesize > 0 && (n = fread(buf, 1, sizeof(buf), sfp)) > 0) {
249 fwrite(buf, 1, n, output);
250 #ifdef BB_FEATURE_WGET_STATUSBAR
252 if (quiet_flag==FALSE)
258 if (n == 0 && ferror(sfp))
259 perror_msg_and_die("network read error");
265 void parse_url(char *url, char **uri_host, int *uri_port, char **uri_path)
271 if (strncmp(url, "http://", 7) != 0)
272 error_msg_and_die("not an http url: %s", url);
276 cp = strchr(*uri_host, ':');
277 sp = strchr(*uri_host, '/');
279 if (cp != NULL && (sp == NULL || cp < sp)) {
281 *uri_port = atoi(cp);
292 FILE *open_socket(char *host, int port)
294 struct sockaddr_in sin;
299 memset(&sin, 0, sizeof(sin));
300 sin.sin_family = AF_INET;
301 if ((hp = (struct hostent *) gethostbyname(host)) == NULL)
302 error_msg_and_die("cannot resolve %s", host);
303 memcpy(&sin.sin_addr, hp->h_addr_list[0], hp->h_length);
304 sin.sin_port = htons(port);
307 * Get the server onto a stdio stream.
309 if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0)
310 perror_msg_and_die("socket()");
311 if (connect(fd, (struct sockaddr *) &sin, sizeof(sin)) < 0)
312 perror_msg_and_die("connect(%s)", host);
313 if ((fp = fdopen(fd, "r+")) == NULL)
314 perror_msg_and_die("fdopen()");
320 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
327 /* retrieve header line */
328 if (fgets(buf, bufsiz, fp) == NULL)
331 /* see if we are at the end of the headers */
332 for (s = buf ; *s == '\r' ; ++s)
337 /* convert the header name to lower case */
338 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
341 /* verify we are at the end of the header name */
343 error_msg_and_die("bad header line: %s", buf);
345 /* locate the start of the header value */
346 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
350 /* locate the end of header */
351 while (*s != '\0' && *s != '\r' && *s != '\n')
354 /* end of header found */
360 /* Rats! The buffer isn't big enough to hold the entire header value. */
361 while (c = getc(fp), c != EOF && c != '\n')
367 #ifdef BB_FEATURE_WGET_STATUSBAR
368 /* Stuff below is from BSD rcp util.c, as added to openshh.
369 * Original copyright notice is retained at the end of this file.
377 struct winsize winsize;
379 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
380 return (winsize.ws_col ? winsize.ws_col : 80);
386 updateprogressmeter(int ignore)
388 int save_errno = errno;
397 struct itimerval itv;
399 itv.it_value.tv_sec = wait;
400 itv.it_value.tv_usec = 0;
401 itv.it_interval = itv.it_value;
402 setitimer(ITIMER_REAL, &itv, NULL);
407 progressmeter(int flag)
409 static const char prefixes[] = " KMGTP";
410 static struct timeval lastupdate;
411 static off_t lastsize;
412 struct timeval now, td, wait;
413 off_t cursize, abbrevsize;
415 int ratio, barlength, i, remaining;
419 (void) gettimeofday(&start, (struct timezone *) 0);
424 (void) gettimeofday(&now, (struct timezone *) 0);
427 ratio = 100.0 * cursize / filesize;
428 ratio = MAX(ratio, 0);
429 ratio = MIN(ratio, 100);
433 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
435 barlength = getttywidth() - 51;
437 i = barlength * ratio / 100;
438 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
440 "*****************************************************************************"
441 "*****************************************************************************",
445 abbrevsize = cursize;
446 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
450 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
451 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
454 timersub(&now, &lastupdate, &wait);
455 if (cursize > lastsize) {
458 if (wait.tv_sec >= STALLTIME) {
459 start.tv_sec += wait.tv_sec;
460 start.tv_usec += wait.tv_usec;
464 timersub(&now, &start, &td);
465 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
467 if (statbytes <= 0 || elapsed <= 0.0 || cursize > filesize) {
468 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
470 } else if (wait.tv_sec >= STALLTIME) {
471 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
474 remaining = (int) (filesize / (statbytes / elapsed) - elapsed);
475 i = remaining / 3600;
477 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
480 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
482 i = remaining % 3600;
483 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
484 "%02d:%02d ETA", i / 60, i % 60);
486 write(fileno(stderr), buf, strlen(buf));
490 sa.sa_handler = updateprogressmeter;
491 sigemptyset(&sa.sa_mask);
492 sa.sa_flags = SA_RESTART;
493 sigaction(SIGALRM, &sa, NULL);
495 } else if (flag == 1) {
502 /* Original copyright notice which applies to the BB_FEATURE_WGET_STATUSBAR stuff,
503 * much of which was blatently stolen from openssh. */
506 * Copyright (c) 1992, 1993
507 * The Regents of the University of California. All rights reserved.
509 * Redistribution and use in source and binary forms, with or without
510 * modification, are permitted provided that the following conditions
512 * 1. Redistributions of source code must retain the above copyright
513 * notice, this list of conditions and the following disclaimer.
514 * 2. Redistributions in binary form must reproduce the above copyright
515 * notice, this list of conditions and the following disclaimer in the
516 * documentation and/or other materials provided with the distribution.
518 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
519 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
521 * 4. Neither the name of the University nor the names of its contributors
522 * may be used to endorse or promote products derived from this software
523 * without specific prior written permission.
525 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
526 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
527 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
528 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
529 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
530 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
531 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
532 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
533 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
534 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
537 * $Id: wget.c,v 1.29 2001/03/09 21:24:12 andersen Exp $
544 c-file-style: "linux"