fetch.c revision 63067 
1/*-
2 * Copyright (c) 2000 Dag-Erling Co�dan Sm�rgrav
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer
10 *    in this position and unchanged.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 *    notice, this list of conditions and the following disclaimer in the
13 *    documentation and/or other materials provided with the distribution.
14 * 3. The name of the author may not be used to endorse or promote products
15 *    derived from this software without specific prior written permission
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
18 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
19 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
20 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
21 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
22 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
23 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
24 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
26 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
27 *
28 *	$FreeBSD: head/usr.bin/fetch/fetch.c 63067 2000-07-13 08:37:39Z des $
29 */
30
31#include <sys/param.h>
32#include <sys/stat.h>
33#include <sys/socket.h>
34
35#include <ctype.h>
36#include <err.h>
37#include <errno.h>
38#include <stdio.h>
39#include <stdlib.h>
40#include <string.h>
41#include <sysexits.h>
42#include <unistd.h>
43
44#include <fetch.h>
45
46#define MINBUFSIZE	4096
47
48/* Option flags */
49int	 A_flag;	/*    -A: do not follow 302 redirects */
50int	 a_flag;	/*    -a: auto retry */
51size_t	 B_size;	/*    -B: buffer size */
52int	 b_flag;	/*!   -b: workaround TCP bug */
53char    *c_dirname;	/*    -c: remote directory */
54int	 d_flag;	/*    -d: direct connection */
55int	 F_flag;	/*    -F: restart without checking mtime  */
56char	*f_filename;	/*    -f: file to fetch */
57int	 H_flag;	/*    -H: use high port */
58char	*h_hostname;	/*    -h: host to fetch from */
59int	 l_flag;	/*    -l: link rather than copy file: URLs */
60int	 m_flag;	/* -[Mm]: mirror mode */
61int	 n_flag;	/*    -n: do not preserve modification time */
62int	 o_flag;	/*    -o: specify output file */
63int	 o_directory;	/*        output file is a directory */
64char	*o_filename;	/*        name of output file */
65int	 o_stdout;	/*        output file is stdout */
66int	 once_flag;	/*    -1: stop at first successful file */
67int	 p_flag = 1;	/* -[Pp]: use passive FTP */
68int	 R_flag;	/*    -R: don't delete partially transferred files */
69int	 r_flag;	/*    -r: restart previously interrupted transfer */
70u_int	 T_secs = 0;	/*    -T: transfer timeout in seconds */
71int	 s_flag;        /*    -s: show size, don't fetch */
72off_t	 S_size;        /*    -S: require size to match */
73int	 t_flag;	/*!   -t: workaround TCP bug */
74int	 v_level = 1;	/*    -v: verbosity level */
75int	 v_tty;		/*        stdout is a tty */
76u_int	 w_secs;	/*    -w: retry delay */
77int	 family = PF_UNSPEC;	/* -[46]: address family to use */
78
79int	 sigalrm;	/* SIGALRM received */
80int	 sigint;	/* SIGINT received */
81
82u_int	 ftp_timeout;	/* default timeout for FTP transfers */
83u_int	 http_timeout;	/* default timeout for HTTP transfers */
84u_char	*buf;		/* transfer buffer */
85
86
87void
88sig_handler(int sig)
89{
90    switch (sig) {
91    case SIGALRM:
92	sigalrm = 1;
93	break;
94    case SIGINT:
95	sigint = 1;
96	break;
97    }
98}
99
100struct xferstat {
101    char		 name[40];
102    struct timeval	 start;
103    struct timeval	 end;
104    struct timeval	 last;
105    off_t		 size;
106    off_t		 offset;
107    off_t		 rcvd;
108};
109
110void
111stat_display(struct xferstat *xs, int force)
112{
113    struct timeval now;
114
115    if (!v_tty)
116	return;
117
118    gettimeofday(&now, NULL);
119    if (!force && now.tv_sec <= xs->last.tv_sec)
120	return;
121    xs->last = now;
122
123    fprintf(stderr, "\rReceiving %s", xs->name);
124    if (xs->size == -1)
125	fprintf(stderr, ": %lld bytes", xs->rcvd);
126    else
127	fprintf(stderr, " (%lld bytes): %d%%", xs->size,
128		(int)((100.0 * xs->rcvd) / xs->size));
129}
130
131void
132stat_start(struct xferstat *xs, char *name, off_t size, off_t offset)
133{
134    snprintf(xs->name, sizeof xs->name, "%s", name);
135    gettimeofday(&xs->start, NULL);
136    xs->last.tv_sec = xs->last.tv_usec = 0;
137    xs->end = xs->last;
138    xs->size = size;
139    xs->offset = offset;
140    xs->rcvd = offset;
141    stat_display(xs, 1);
142}
143
144void
145stat_update(struct xferstat *xs, off_t rcvd, int force)
146{
147    xs->rcvd = rcvd;
148    stat_display(xs, 0);
149}
150
151void
152stat_end(struct xferstat *xs)
153{
154    double delta;
155    double bps;
156
157    gettimeofday(&xs->end, NULL);
158
159    stat_display(xs, 1);
160    fputc('\n', stderr);
161    delta = (xs->end.tv_sec + (xs->end.tv_usec / 1.e6))
162	- (xs->start.tv_sec + (xs->start.tv_usec / 1.e6));
163    fprintf(stderr, "%lld bytes transferred in %.1f seconds ",
164	    xs->rcvd - xs->offset, delta);
165    bps = (xs->rcvd - xs->offset) / delta;
166    if (bps > 1024*1024)
167	fprintf(stderr, "(%.2f MBps)\n", bps / (1024*1024));
168    else if (bps > 1024)
169	fprintf(stderr, "(%.2f kBps)\n", bps / 1024);
170    else
171	fprintf(stderr, "(%.2f Bps)\n", bps);
172}
173
174int
175fetch(char *URL, char *path)
176{
177    struct url *url;
178    struct url_stat us;
179    struct stat sb;
180    struct xferstat xs;
181    FILE *f, *of;
182    size_t size;
183    off_t count;
184    char flags[8];
185    int n, r;
186    u_int timeout;
187
188    f = of = NULL;
189
190    /* parse URL */
191    if ((url = fetchParseURL(URL)) == NULL) {
192	warnx("%s: parse error", URL);
193	goto failure;
194    }
195
196    timeout = 0;
197    *flags = 0;
198
199    /* common flags */
200    if (v_level > 2)
201	strcat(flags, "v");
202    switch (family) {
203    case PF_INET:
204	strcat(flags, "4");
205	break;
206    case PF_INET6:
207	strcat(flags, "6");
208	break;
209    }
210
211    /* FTP specific flags */
212    if (strcmp(url->scheme, "ftp") == 0) {
213	if (p_flag)
214	    strcat(flags, "p");
215	if (d_flag)
216	    strcat(flags, "d");
217	if (H_flag)
218	    strcat(flags, "h");
219	timeout = T_secs ? T_secs : ftp_timeout;
220    }
221
222    /* HTTP specific flags */
223    if (strcmp(url->scheme, "http") == 0) {
224	if (d_flag)
225	    strcat(flags, "d");
226	if (A_flag)
227	    strcat(flags, "A");
228	timeout = T_secs ? T_secs : http_timeout;
229    }
230
231    /* set the protocol timeout. */
232    fetchTimeout = timeout;
233
234    /* stat remote file */
235    if (fetchStat(url, &us, flags) == -1)
236	goto failure;
237
238    /* just print size */
239    if (s_flag) {
240	if (us.size == -1)
241	    printf("Unknown\n");
242	else
243	    printf("%lld\n", us.size);
244	goto success;
245    }
246
247    /* check that size is as expected */
248    if (S_size && us.size != -1 && us.size != S_size) {
249	warnx("%s: size mismatch: expected %lld, actual %lld",
250	      path, S_size, us.size);
251	goto failure;
252    }
253
254    /* symlink instead of copy */
255    if (l_flag && strcmp(url->scheme, "file") == 0 && !o_stdout) {
256	if (symlink(url->doc, path) == -1) {
257	    warn("%s: symlink()", path);
258	    goto failure;
259	}
260	goto success;
261    }
262
263    if (o_stdout) {
264	/* output to stdout */
265	of = stdout;
266    } else if (r_flag && us.size != -1 && stat(path, &sb) != -1
267	       && (F_flag || (us.mtime && sb.st_mtime == us.mtime))) {
268	/* output to file, restart aborted transfer */
269	if (us.size == sb.st_size)
270	    goto success;
271	else if (sb.st_size > us.size && truncate(path, us.size) == -1) {
272	    warn("%s: truncate()", path);
273	    goto failure;
274	}
275	if ((of = fopen(path, "a")) == NULL) {
276	    warn("%s: open()", path);
277	    goto failure;
278	}
279	url->offset = sb.st_size;
280    } else if (m_flag && us.size != -1 && stat(path, &sb) != -1) {
281	/* output to file, mirror mode */
282	if (sb.st_size == us.size && sb.st_mtime == us.mtime)
283	    return 0;
284	if ((of = fopen(path, "w")) == NULL) {
285	    warn("%s: open()", path);
286	    goto failure;
287	}
288    } else {
289	/* output to file, all other cases */
290	if ((of = fopen(path, "w")) == NULL) {
291	    warn("%s: open()", path);
292	    goto failure;
293	}
294    }
295    count = url->offset;
296
297    /* start the transfer */
298    if ((f = fetchGet(url, flags)) == NULL) {
299	warnx("%s", fetchLastErrString);
300	if (!R_flag && !r_flag && !o_stdout)
301	    unlink(path);
302	goto failure;
303    }
304
305    /* start the counter */
306    stat_start(&xs, path, us.size, count);
307
308    sigint = sigalrm = 0;
309
310    /* suck in the data */
311    for (n = 0; !sigint && !sigalrm; ++n) {
312	if (us.size != -1 && us.size - count < B_size)
313	    size = us.size - count;
314	else
315	    size = B_size;
316	if (timeout)
317	    alarm(timeout);
318	if ((size = fread(buf, 1, size, f)) <= 0)
319	    break;
320	stat_update(&xs, count += size, 0);
321	if (fwrite(buf, size, 1, of) != 1)
322	    break;
323    }
324
325    if (timeout)
326	alarm(0);
327
328    stat_end(&xs);
329
330    /* Set mtime of local file */
331    if (!n_flag && us.mtime && !o_stdout) {
332	struct timeval tv[2];
333
334	fflush(of);
335	tv[0].tv_sec = (long)(us.atime ? us.atime : us.mtime);
336	tv[1].tv_sec = (long)us.mtime;
337	tv[0].tv_usec = tv[1].tv_usec = 0;
338	if (utimes(path, tv))
339	    warn("%s: utimes()", path);
340    }
341
342    /* timed out or interrupted? */
343    if (sigalrm)
344	warnx("transfer timed out");
345    if (sigint)
346	warnx("transfer interrupted");
347
348    /* check the status of our files */
349    if (ferror(f))
350	warn("%s", URL);
351    if (ferror(of))
352	warn("%s", path);
353    if (ferror(f) || ferror(of))
354	goto failure;
355
356    /* did the transfer complete normally? */
357    if (us.size != -1 && count < us.size) {
358	warnx("%s appears to be truncated: %lld/%lld bytes",
359	      path, count, us.size);
360	goto failure_keep;
361    }
362
363 success:
364    r = 0;
365    goto done;
366 failure:
367    if (of && of != stdout && !R_flag && !r_flag)
368	unlink(path);
369 failure_keep:
370    r = -1;
371    goto done;
372 done:
373    if (f)
374	fclose(f);
375    if (of && of != stdout)
376	fclose(of);
377    if (url)
378	fetchFreeURL(url);
379    return r;
380}
381
382void
383usage(void)
384{
385    /* XXX badly out of synch */
386    fprintf(stderr,
387	    "Usage: fetch [-1AFHMPRabdlmnpqrstv] [-o outputfile] [-S bytes]\n"
388	    "             [-B bytes] [-T seconds] [-w seconds]\n"
389	    "             [-f file -h host [-c dir] | URL ...]\n"
390	);
391}
392
393
394#define PARSENUM(NAME, TYPE)		\
395int					\
396NAME(char *s, TYPE *v)			\
397{					\
398    *v = 0;				\
399    for (*v = 0; *s; s++)		\
400	if (isdigit(*s))		\
401	    *v = *v * 10 + *s - '0';	\
402	else				\
403	    return -1;			\
404    return 0;				\
405}
406
407PARSENUM(parseint, u_int)
408PARSENUM(parsesize, size_t)
409PARSENUM(parseoff, off_t)
410
411int
412main(int argc, char *argv[])
413{
414    struct stat sb;
415    char *p, *q, *s;
416    int c, e, r;
417
418    while ((c = getopt(argc, argv,
419		       "146AaB:bc:dFf:h:lHMmnPpo:qRrS:sT:tvw:")) != EOF)
420	switch (c) {
421	case '1':
422	    once_flag = 1;
423	    break;
424	case '4':
425	    family = PF_INET;
426	    break;
427	case '6':
428	    family = PF_INET6;
429	    break;
430	case 'A':
431	    A_flag = 1;
432	    break;
433	case 'a':
434	    a_flag = 1;
435	    break;
436	case 'B':
437	    if (parsesize(optarg, &B_size) == -1)
438		errx(1, "invalid buffer size");
439	    break;
440	case 'b':
441	    warnx("warning: the -b option is deprecated");
442	    b_flag = 1;
443	    break;
444	case 'c':
445	    c_dirname = optarg;
446	    break;
447	case 'd':
448	    d_flag = 1;
449	    break;
450	case 'F':
451	    F_flag = 1;
452	    break;
453	case 'f':
454	    f_filename = optarg;
455	    break;
456	case 'H':
457	    H_flag = 1;
458	    break;
459	case 'h':
460	    h_hostname = optarg;
461	    break;
462	case 'l':
463	    l_flag = 1;
464	    break;
465	case 'o':
466	    o_flag = 1;
467	    o_filename = optarg;
468	    break;
469	case 'M':
470	case 'm':
471	    m_flag = 1;
472	    break;
473	case 'n':
474	    n_flag = 1;
475	    break;
476	case 'P':
477	case 'p':
478	    p_flag = 1;
479	    break;
480	case 'q':
481	    v_level = 0;
482	    break;
483	case 'R':
484	    R_flag = 1;
485	    break;
486	case 'r':
487	    r_flag = 1;
488	    break;
489	case 'S':
490	    if (parseoff(optarg, &S_size) == -1)
491		errx(1, "invalid size");
492	    break;
493	case 's':
494	    s_flag = 1;
495	    break;
496	case 'T':
497	    if (parseint(optarg, &T_secs) == -1)
498		errx(1, "invalid timeout");
499	    break;
500	case 't':
501	    t_flag = 1;
502	    warnx("warning: the -t option is deprecated");
503	    break;
504	case 'v':
505	    v_level++;
506	    break;
507	case 'w':
508	    a_flag = 1;
509	    if (parseint(optarg, &w_secs) == -1)
510		errx(1, "invalid delay");
511	    break;
512	default:
513	    usage();
514	    exit(EX_USAGE);
515	}
516
517    argc -= optind;
518    argv += optind;
519
520    if (h_hostname || f_filename || c_dirname) {
521	if (!h_hostname || !f_filename || argc) {
522	    usage();
523	    exit(EX_USAGE);
524	}
525	/* XXX this is a hack. */
526	if (strcspn(h_hostname, "@:/") != strlen(h_hostname))
527	    errx(1, "invalid hostname");
528	if (asprintf(argv, "ftp://%s/%s/%s", h_hostname,
529		     c_dirname ? c_dirname : "", f_filename) == -1)
530	    errx(1, strerror(ENOMEM));
531	argc++;
532    }
533
534    if (!argc) {
535	usage();
536	exit(EX_USAGE);
537    }
538
539    /* allocate buffer */
540    if (B_size < MINBUFSIZE)
541	B_size = MINBUFSIZE;
542    if ((buf = malloc(B_size)) == NULL)
543	errx(1, strerror(ENOMEM));
544
545    /* timeout handling */
546    signal(SIGALRM, sig_handler);
547    if ((s = getenv("FTP_TIMEOUT")) != NULL) {
548	if (parseint(s, &ftp_timeout) == -1) {
549	    warnx("FTP_TIMEOUT is not a positive integer");
550	    ftp_timeout = 0;
551	}
552    }
553    if ((s = getenv("HTTP_TIMEOUT")) != NULL) {
554	if (parseint(s, &http_timeout) == -1) {
555	    warnx("HTTP_TIMEOUT is not a positive integer");
556	    http_timeout = 0;
557	}
558    }
559
560    /* interrupt handling */
561    signal(SIGINT, sig_handler);
562
563    /* output file */
564    if (o_flag) {
565	if (strcmp(o_filename, "-") == 0) {
566	    o_stdout = 1;
567	} else if (stat(o_filename, &sb) == -1) {
568	    if (errno == ENOENT) {
569		if (argc > 1)
570		    errx(EX_USAGE, "%s is not a directory", o_filename);
571	    } else {
572		err(EX_IOERR, "%s", o_filename);
573	    }
574	} else {
575	    if (sb.st_mode & S_IFDIR)
576		o_directory = 1;
577	}
578    }
579
580    /* check if output is to a tty (for progress report) */
581    v_tty = isatty(STDERR_FILENO);
582    r = 0;
583
584    while (argc) {
585	if ((p = strrchr(*argv, '/')) == NULL)
586	    p = *argv;
587	else
588	    p++;
589
590	if (!*p)
591	    p = "fetch.out";
592
593	fetchLastErrCode = 0;
594
595	if (o_flag) {
596	    if (o_stdout) {
597		e = fetch(*argv, "-");
598	    } else if (o_directory) {
599		asprintf(&q, "%s/%s", o_filename, p);
600		e = fetch(*argv, q);
601		free(q);
602	    } else {
603		e = fetch(*argv, o_filename);
604	    }
605	} else {
606	    e = fetch(*argv, p);
607	}
608
609	if (sigint)
610	    exit(1);
611
612	if (e == 0 && once_flag)
613	    exit(0);
614
615	if (e) {
616	    r = 1;
617	    if ((fetchLastErrCode
618		 && fetchLastErrCode != FETCH_UNAVAIL
619		 && fetchLastErrCode != FETCH_MOVED
620		 && fetchLastErrCode != FETCH_URL
621		 && fetchLastErrCode != FETCH_RESOLV
622		 && fetchLastErrCode != FETCH_UNKNOWN)) {
623		if (w_secs) {
624		    if (v_level)
625			fprintf(stderr, "Waiting %d seconds before retrying\n", w_secs);
626		    sleep(w_secs);
627		}
628		if (a_flag)
629		    continue;
630		fprintf(stderr, "Skipping %s\n", *argv);
631	    }
632	}
633
634	argc--, argv++;
635    }
636
637    exit(r);
638}
639