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.
26 #include <sys/ioctl.h>
29 #include <sys/types.h>
31 #include <sys/socket.h>
32 #include <netinet/in.h>
33 #include <arpa/inet.h>
37 void parse_url(char *url, char **uri_host, int *uri_port, char **uri_path);
38 FILE *open_socket(char *host, int port);
39 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc);
40 void progressmeter(int flag);
42 /* Globals (can be accessed from signal handlers */
43 static off_t filesize = 0; /* content-length of the file */
44 #ifdef BB_FEATURE_WGET_STATUSBAR
45 static char *curfile; /* Name of current file being transferred. */
46 static struct timeval start; /* Time a transfer started. */
47 volatile unsigned long statbytes; /* Number of bytes transferred so far. */
48 /* For progressmeter() -- number of seconds before xfer considered "stalled" */
49 static const int STALLTIME = 5;
52 void close_and_delete_outfile(FILE* output, char *fname_out, int do_continue)
54 if (output != stdout && do_continue==0) {
60 int wget_main(int argc, char **argv)
63 char *proxy, *proxy_host;
64 int uri_port, proxy_port;
68 FILE *sfp; /* socket to web server */
69 char *uri_host, *uri_path; /* parsed from command line url */
70 char *fname_out = NULL; /* where to direct output (-O) */
71 int do_continue = 0; /* continue a prev transfer (-c) */
72 long beg_range = 0L; /* range at which continue begins */
73 int got_clen = 0; /* got content-length: from server */
74 FILE *output; /* socket to web server */
75 int quiet_flag = FALSE; /* Be verry, verry quiet... */
80 while ((n = getopt(argc, argv, "cqO:")) != EOF) {
89 /* can't set fname_out to NULL if outputting to stdout, because
90 * this gets interpreted as the auto-gen output filename
91 * case below - tausq@debian.org
93 fname_out = (strcmp(optarg, "-") == 0 ? (char *)1 : optarg);
100 if (argc - optind != 1)
104 * Use the proxy if necessary.
106 if ((proxy = getenv("http_proxy")) != NULL) {
107 proxy = xstrdup(proxy);
108 parse_url(proxy, &proxy_host, &proxy_port, &uri_path);
109 parse_url(argv[optind], &uri_host, &uri_port, &uri_path);
112 * Parse url into components.
114 parse_url(argv[optind], &uri_host, &uri_port, &uri_path);
119 /* Guess an output filename */
122 #ifdef BB_FEATURE_WGET_STATUSBAR
125 get_last_path_component(uri_path);
126 if (fname_out==NULL || strlen(fname_out)<1) {
128 #ifdef BB_FEATURE_WGET_STATUSBAR
133 #ifdef BB_FEATURE_WGET_STATUSBAR
135 curfile=argv[optind];
138 if (do_continue && !fname_out)
139 error_msg_and_die("cannot specify continue (-c) without a filename (-O)\n");
143 * Open socket to server.
145 sfp = open_socket(proxy_host, proxy_port);
147 /* Make the assumption that if the file already exists
148 * on disk that the intention is to continue downloading
149 * a previously aborted download -Erik */
150 if (stat(fname_out, &sbuf) == 0) {
155 * Open the output file stream.
157 if (fname_out != (char *)1) {
158 if ( (output=fopen(fname_out, (do_continue ? "a" : "w"))) == NULL)
159 perror_msg_and_die("fopen(%s)", fname_out);
165 * Determine where to start transfer.
168 if (fstat(fileno(output), &sbuf) < 0)
169 perror_msg_and_die("fstat()");
170 if (sbuf.st_size > 0)
171 beg_range = sbuf.st_size;
179 fprintf(sfp, "GET http://%s:%d/%s HTTP/1.1\r\n",
180 uri_host, uri_port, uri_path);
181 fprintf(sfp, "Host: %s\r\nUser-Agent: Wget\r\n", uri_host);
184 fprintf(sfp, "Range: bytes=%ld-\r\n", beg_range);
185 fprintf(sfp,"Connection: close\r\n\r\n");
188 * Retrieve HTTP response line and check for "200" status code.
190 if (fgets(buf, sizeof(buf), sfp) == NULL) {
191 close_and_delete_outfile(output, fname_out, do_continue);
192 error_msg_and_die("no response from server\n");
194 for (s = buf ; *s != '\0' && !isspace(*s) ; ++s)
196 for ( ; isspace(*s) ; ++s)
207 close_and_delete_outfile(output, fname_out, do_continue);
208 error_msg_and_die("server returned error %d: %s", atoi(s), buf);
212 * Retrieve HTTP headers.
214 while ((s = gethdr(buf, sizeof(buf), sfp, &n)) != NULL) {
215 if (strcasecmp(buf, "content-length") == 0) {
220 if (strcasecmp(buf, "transfer-encoding") == 0) {
221 close_and_delete_outfile(output, fname_out, do_continue);
222 error_msg_and_die("server wants to do %s transfer encoding\n", s);
228 * Retrieve HTTP body.
230 #ifdef BB_FEATURE_WGET_STATUSBAR
232 if (quiet_flag==FALSE)
235 while (filesize > 0 && (n = fread(buf, 1, sizeof(buf), sfp)) > 0) {
236 fwrite(buf, 1, n, output);
237 #ifdef BB_FEATURE_WGET_STATUSBAR
239 if (quiet_flag==FALSE)
245 if (n == 0 && ferror(sfp))
246 perror_msg_and_die("network read error");
252 void parse_url(char *url, char **uri_host, int *uri_port, char **uri_path)
258 if (strncmp(url, "http://", 7) != 0)
259 error_msg_and_die("not an http url: %s\n", url);
263 cp = strchr(*uri_host, ':');
264 sp = strchr(*uri_host, '/');
266 if (cp != NULL && (sp == NULL || cp < sp)) {
268 *uri_port = atoi(cp);
279 FILE *open_socket(char *host, int port)
281 struct sockaddr_in sin;
286 memzero(&sin, sizeof(sin));
287 sin.sin_family = AF_INET;
288 if ((hp = (struct hostent *) gethostbyname(host)) == NULL)
289 error_msg_and_die("cannot resolve %s\n", host);
290 memcpy(&sin.sin_addr, hp->h_addr_list[0], hp->h_length);
291 sin.sin_port = htons(port);
294 * Get the server onto a stdio stream.
296 if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0)
297 perror_msg_and_die("socket()");
298 if (connect(fd, (struct sockaddr *) &sin, sizeof(sin)) < 0)
299 perror_msg_and_die("connect(%s)", host);
300 if ((fp = fdopen(fd, "r+")) == NULL)
301 perror_msg_and_die("fdopen()");
307 char *gethdr(char *buf, size_t bufsiz, FILE *fp, int *istrunc)
314 /* retrieve header line */
315 if (fgets(buf, bufsiz, fp) == NULL)
318 /* see if we are at the end of the headers */
319 for (s = buf ; *s == '\r' ; ++s)
324 /* convert the header name to lower case */
325 for (s = buf ; isalnum(*s) || *s == '-' ; ++s)
328 /* verify we are at the end of the header name */
330 error_msg_and_die("bad header line: %s\n", buf);
332 /* locate the start of the header value */
333 for (*s++ = '\0' ; *s == ' ' || *s == '\t' ; ++s)
337 /* locate the end of header */
338 while (*s != '\0' && *s != '\r' && *s != '\n')
341 /* end of header found */
347 /* Rats! The buffer isn't big enough to hold the entire header value. */
348 while (c = getc(fp), c != EOF && c != '\n')
354 #ifdef BB_FEATURE_WGET_STATUSBAR
355 /* Stuff below is from BSD rcp util.c, as added to openshh.
356 * Original copyright notice is retained at the end of this file.
364 struct winsize winsize;
366 if (ioctl(fileno(stdout), TIOCGWINSZ, &winsize) != -1)
367 return (winsize.ws_col ? winsize.ws_col : 80);
373 updateprogressmeter(int ignore)
375 int save_errno = errno;
384 struct itimerval itv;
386 itv.it_value.tv_sec = wait;
387 itv.it_value.tv_usec = 0;
388 itv.it_interval = itv.it_value;
389 setitimer(ITIMER_REAL, &itv, NULL);
394 progressmeter(int flag)
396 static const char prefixes[] = " KMGTP";
397 static struct timeval lastupdate;
398 static off_t lastsize;
399 struct timeval now, td, wait;
400 off_t cursize, abbrevsize;
402 int ratio, barlength, i, remaining;
406 (void) gettimeofday(&start, (struct timezone *) 0);
411 (void) gettimeofday(&now, (struct timezone *) 0);
414 ratio = 100.0 * cursize / filesize;
415 ratio = MAX(ratio, 0);
416 ratio = MIN(ratio, 100);
420 snprintf(buf, sizeof(buf), "\r%-20.20s %3d%% ", curfile, ratio);
422 barlength = getttywidth() - 51;
424 i = barlength * ratio / 100;
425 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
427 "*****************************************************************************"
428 "*****************************************************************************",
432 abbrevsize = cursize;
433 while (abbrevsize >= 100000 && i < sizeof(prefixes)) {
437 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf), " %5d %c%c ",
438 (int) abbrevsize, prefixes[i], prefixes[i] == ' ' ? ' ' :
441 timersub(&now, &lastupdate, &wait);
442 if (cursize > lastsize) {
445 if (wait.tv_sec >= STALLTIME) {
446 start.tv_sec += wait.tv_sec;
447 start.tv_usec += wait.tv_usec;
451 timersub(&now, &start, &td);
452 elapsed = td.tv_sec + (td.tv_usec / 1000000.0);
454 if (statbytes <= 0 || elapsed <= 0.0 || cursize > filesize) {
455 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
457 } else if (wait.tv_sec >= STALLTIME) {
458 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
461 remaining = (int) (filesize / (statbytes / elapsed) - elapsed);
462 i = remaining / 3600;
464 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
467 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
469 i = remaining % 3600;
470 snprintf(buf + strlen(buf), sizeof(buf) - strlen(buf),
471 "%02d:%02d ETA", i / 60, i % 60);
473 write(fileno(stderr), buf, strlen(buf));
477 sa.sa_handler = updateprogressmeter;
478 sigemptyset(&sa.sa_mask);
479 sa.sa_flags = SA_RESTART;
480 sigaction(SIGALRM, &sa, NULL);
482 } else if (flag == 1) {
489 /* Original copyright notice which applies to the BB_FEATURE_WGET_STATUSBAR stuff,
490 * much of which was blatently stolen from openssh. */
493 * Copyright (c) 1992, 1993
494 * The Regents of the University of California. All rights reserved.
496 * Redistribution and use in source and binary forms, with or without
497 * modification, are permitted provided that the following conditions
499 * 1. Redistributions of source code must retain the above copyright
500 * notice, this list of conditions and the following disclaimer.
501 * 2. Redistributions in binary form must reproduce the above copyright
502 * notice, this list of conditions and the following disclaimer in the
503 * documentation and/or other materials provided with the distribution.
505 * 3. <BSD Advertising Clause omitted per the July 22, 1999 licensing change
506 * ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change>
508 * 4. Neither the name of the University nor the names of its contributors
509 * may be used to endorse or promote products derived from this software
510 * without specific prior written permission.
512 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
513 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
514 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
515 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
516 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
517 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
518 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
519 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
520 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
521 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
524 * $Id: wget.c,v 1.22 2001/01/26 02:04:49 andersen Exp $
531 c-file-style: "linux"