fetch.c revision 63345 
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 63345 2000-07-17 21:49:01Z 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 <signal.h>
39#include <stdio.h>
40#include <stdlib.h>
41#include <string.h>
42#include <sysexits.h>
43#include <unistd.h>
44
45#include <fetch.h>
46
47#define MINBUFSIZE	4096
48
49/* Option flags */
50int	 A_flag;	/*    -A: do not follow 302 redirects */
51int	 a_flag;	/*    -a: auto retry */
52size_t	 B_size;	/*    -B: buffer size */
53int	 b_flag;	/*!   -b: workaround TCP bug */
54char    *c_dirname;	/*    -c: remote directory */
55int	 d_flag;	/*    -d: direct connection */
56int	 F_flag;	/*    -F: restart without checking mtime  */
57char	*f_filename;	/*    -f: file to fetch */
58int	 H_flag;	/*    -H: use high port */
59char	*h_hostname;	/*    -h: host to fetch from */
60int	 l_flag;	/*    -l: link rather than copy file: URLs */
61int	 m_flag;	/* -[Mm]: mirror mode */
62int	 n_flag;	/*    -n: do not preserve modification time */
63int	 o_flag;	/*    -o: specify output file */
64int	 o_directory;	/*        output file is a directory */
65char	*o_filename;	/*        name of output file */
66int	 o_stdout;	/*        output file is stdout */
67int	 once_flag;	/*    -1: stop at first successful file */
68int	 p_flag = 1;	/* -[Pp]: use passive FTP */
69int	 R_flag;	/*    -R: don't delete partially transferred files */
70int	 r_flag;	/*    -r: restart previously interrupted transfer */
71u_int	 T_secs = 0;	/*    -T: transfer timeout in seconds */
72int	 s_flag;        /*    -s: show size, don't fetch */
73off_t	 S_size;        /*    -S: require size to match */
74int	 t_flag;	/*!   -t: workaround TCP bug */
75int	 v_level = 1;	/*    -v: verbosity level */
76int	 v_tty;		/*        stdout is a tty */
77u_int	 w_secs;	/*    -w: retry delay */
78int	 family = PF_UNSPEC;	/* -[46]: address family to use */
79
80int	 sigalrm;	/* SIGALRM received */
81int	 sigint;	/* SIGINT received */
82
83u_int	 ftp_timeout;	/* default timeout for FTP transfers */
84u_int	 http_timeout;	/* default timeout for HTTP transfers */
85u_char	*buf;		/* transfer buffer */
86
87
88void
89sig_handler(int sig)
90{
91    switch (sig) {
92    case SIGALRM:
93	sigalrm = 1;
94	break;
95    case SIGINT:
96	sigint = 1;
97	break;
98    }
99}
100
101struct xferstat {
102    char		 name[40];
103    struct timeval	 start;
104    struct timeval	 end;
105    struct timeval	 last;
106    off_t		 size;
107    off_t		 offset;
108    off_t		 rcvd;
109};
110
111void
112stat_display(struct xferstat *xs, int force)
113{
114    struct timeval now;
115
116    if (!v_tty)
117	return;
118
119    gettimeofday(&now, NULL);
120    if (!force && now.tv_sec <= xs->last.tv_sec)
121	return;
122    xs->last = now;
123
124    fprintf(stderr, "\rReceiving %s", xs->name);
125    if (xs->size == -1)
126	fprintf(stderr, ": %lld bytes", xs->rcvd);
127    else
128	fprintf(stderr, " (%lld bytes): %d%%", xs->size,
129		(int)((100.0 * xs->rcvd) / xs->size));
130}
131
132void
133stat_start(struct xferstat *xs, char *name, off_t size, off_t offset)
134{
135    snprintf(xs->name, sizeof xs->name, "%s", name);
136    gettimeofday(&xs->start, NULL);
137    xs->last.tv_sec = xs->last.tv_usec = 0;
138    xs->end = xs->last;
139    xs->size = size;
140    xs->offset = offset;
141    xs->rcvd = offset;
142    stat_display(xs, 1);
143}
144
145void
146stat_update(struct xferstat *xs, off_t rcvd, int force)
147{
148    xs->rcvd = rcvd;
149    stat_display(xs, 0);
150}
151
152void
153stat_end(struct xferstat *xs)
154{
155    double delta;
156    double bps;
157
158    gettimeofday(&xs->end, NULL);
159
160    stat_display(xs, 1);
161    fputc('\n', stderr);
162    delta = (xs->end.tv_sec + (xs->end.tv_usec / 1.e6))
163	- (xs->start.tv_sec + (xs->start.tv_usec / 1.e6));
164    fprintf(stderr, "%lld bytes transferred in %.1f seconds ",
165	    xs->rcvd - xs->offset, delta);
166    bps = (xs->rcvd - xs->offset) / delta;
167    if (bps > 1024*1024)
168	fprintf(stderr, "(%.2f MBps)\n", bps / (1024*1024));
169    else if (bps > 1024)
170	fprintf(stderr, "(%.2f kBps)\n", bps / 1024);
171    else
172	fprintf(stderr, "(%.2f Bps)\n", bps);
173}
174
175int
176fetch(char *URL, char *path)
177{
178    struct url *url;
179    struct url_stat us;
180    struct stat sb;
181    struct xferstat xs;
182    FILE *f, *of;
183    size_t size;
184    off_t count;
185    char flags[8];
186    int n, r;
187    u_int timeout;
188
189    f = of = NULL;
190
191    /* parse URL */
192    if ((url = fetchParseURL(URL)) == NULL) {
193	warnx("%s: parse error", URL);
194	goto failure;
195    }
196
197    timeout = 0;
198    *flags = 0;
199    count = 0;
200
201    /* common flags */
202    if (v_level > 2)
203	strcat(flags, "v");
204    switch (family) {
205    case PF_INET:
206	strcat(flags, "4");
207	break;
208    case PF_INET6:
209	strcat(flags, "6");
210	break;
211    }
212
213    /* FTP specific flags */
214    if (strcmp(url->scheme, "ftp") == 0) {
215	if (p_flag)
216	    strcat(flags, "p");
217	if (d_flag)
218	    strcat(flags, "d");
219	if (H_flag)
220	    strcat(flags, "h");
221	timeout = T_secs ? T_secs : ftp_timeout;
222    }
223
224    /* HTTP specific flags */
225    if (strcmp(url->scheme, "http") == 0) {
226	if (d_flag)
227	    strcat(flags, "d");
228	if (A_flag)
229	    strcat(flags, "A");
230	timeout = T_secs ? T_secs : http_timeout;
231    }
232
233    /* set the protocol timeout. */
234    fetchTimeout = timeout;
235
236    /* just print size */
237    if (s_flag) {
238	if (fetchStat(url, &us, flags) == -1)
239	    goto failure;
240	if (us.size == -1)
241	    printf("Unknown\n");
242	else
243	    printf("%lld\n", us.size);
244	goto success;
245    }
246
247    /*
248     * If the -r flag was specified, we have to compare the local and
249     * remote files, so we should really do a fetchStat() first, but I
250     * know of at least one HTTP server that only sends the content
251     * size in response to GET requests, and leaves it out of replies
252     * to HEAD requests. Also, in the (frequent) case that the local
253     * and remote files match but the local file is truncated, we have
254     * sufficient information *before* the compare to issue a correct
255     * request. Therefore, we always issue a GET request as if we were
256     * sure the local file was a truncated copy of the remote file; we
257     * can drop the connection later if we change our minds.
258     */
259    if (r_flag && !o_stdout && stat(path, &sb) != -1)
260	url->offset = sb.st_size;
261
262    /* start the transfer */
263    if ((f = fetchXGet(url, &us, flags)) == NULL) {
264	warnx("%s: %s", path, fetchLastErrString);
265	goto failure;
266    }
267    if (sigint)
268	goto signal;
269
270    /* check that size is as expected */
271    if (S_size) {
272	if (us.size == -1) {
273	    warnx("%s: size unknown", path);
274	    goto failure;
275	} else if (us.size != S_size) {
276	    warnx("%s: size mismatch: expected %lld, actual %lld",
277		  path, S_size, us.size);
278	    goto failure;
279	}
280    }
281
282    /* symlink instead of copy */
283    if (l_flag && strcmp(url->scheme, "file") == 0 && !o_stdout) {
284	if (symlink(url->doc, path) == -1) {
285	    warn("%s: symlink()", path);
286	    goto failure;
287	}
288	goto success;
289    }
290
291    /* open output file */
292    if (o_stdout) {
293	/* output to stdout */
294	of = stdout;
295    } else if (url->offset) {
296	/* resume mode, local file exists */
297	if (!F_flag && us.mtime && sb.st_mtime != us.mtime) {
298	    /* no match! have to refetch */
299	    fclose(f);
300	    url->offset = 0;
301	    if ((f = fetchXGet(url, &us, flags)) == NULL) {
302		warnx("%s: %s", path, fetchLastErrString);
303		goto failure;
304	    }
305	    if (sigint)
306		goto signal;
307	} else {
308	    us.size += url->offset;
309	}
310	if (us.size == sb.st_size)
311	    /* nothing to do */
312	    goto success;
313	if (sb.st_size > us.size) {
314	    /* local file too long! */
315	    warnx("%s: local file (%lld bytes) is longer "
316		  "than remote file (%lld bytes)",
317		  path, sb.st_size, us.size);
318	    goto failure;
319	}
320	/* we got through, open local file in append mode */
321	/*
322	 * XXX there's a race condition here - the file we open is not
323         * necessarily the same as the one we stat()'ed earlier...
324	 */
325	if ((of = fopen(path, "a")) == NULL) {
326	    warn("%s: open()", path);
327	    goto failure;
328	}
329    }
330    if (m_flag && stat(path, &sb) != -1) {
331	/* mirror mode, local file exists */
332	if (sb.st_size == us.size && sb.st_mtime == us.mtime)
333	    goto success;
334    }
335    if (!of) {
336	/*
337	 * We don't yet have an output file; either this is a vanilla
338	 * run with no special flags, or the local and remote files
339	 * didn't match.
340	 */
341	if ((of = fopen(path, "w")) == NULL) {
342	    warn("%s: open()", path);
343	    goto failure;
344	}
345    }
346    count = url->offset;
347
348    /* start the counter */
349    stat_start(&xs, path, us.size, count);
350
351    sigint = sigalrm = 0;
352
353    /* suck in the data */
354    for (n = 0; !sigint && !sigalrm; ++n) {
355	if (us.size != -1 && us.size - count < B_size)
356	    size = us.size - count;
357	else
358	    size = B_size;
359	if (timeout)
360	    alarm(timeout);
361	if ((size = fread(buf, 1, size, f)) <= 0)
362	    break;
363	stat_update(&xs, count += size, 0);
364	if (fwrite(buf, size, 1, of) != 1)
365	    break;
366    }
367
368    if (timeout)
369	alarm(0);
370
371    stat_end(&xs);
372
373    /* Set mtime of local file */
374    if (!n_flag && us.mtime && !o_stdout) {
375	struct timeval tv[2];
376
377	fflush(of);
378	tv[0].tv_sec = (long)(us.atime ? us.atime : us.mtime);
379	tv[1].tv_sec = (long)us.mtime;
380	tv[0].tv_usec = tv[1].tv_usec = 0;
381	if (utimes(path, tv))
382	    warn("%s: utimes()", path);
383    }
384
385    /* timed out or interrupted? */
386 signal:
387    if (sigalrm)
388	warnx("transfer timed out");
389    if (sigint)
390	warnx("transfer interrupted");
391
392    if (!sigalrm && !sigint) {
393	/* check the status of our files */
394	if (ferror(f))
395	    warn("%s", URL);
396	if (ferror(of))
397	    warn("%s", path);
398	if (ferror(f) || ferror(of))
399	    goto failure;
400    }
401
402    /* did the transfer complete normally? */
403    if (us.size != -1 && count < us.size) {
404	warnx("%s appears to be truncated: %lld/%lld bytes",
405	      path, count, us.size);
406	goto failure_keep;
407    }
408
409 success:
410    r = 0;
411    goto done;
412 failure:
413    if (of && of != stdout && !R_flag && !r_flag)
414	unlink(path);
415 failure_keep:
416    r = -1;
417    goto done;
418 done:
419    if (f)
420	fclose(f);
421    if (of && of != stdout)
422	fclose(of);
423    if (url)
424	fetchFreeURL(url);
425    return r;
426}
427
428void
429usage(void)
430{
431    /* XXX badly out of synch */
432    fprintf(stderr,
433	    "Usage: fetch [-1AFHMPRabdlmnpqrstv] [-o outputfile] [-S bytes]\n"
434	    "             [-B bytes] [-T seconds] [-w seconds]\n"
435	    "             [-f file -h host [-c dir] | URL ...]\n"
436	);
437}
438
439
440#define PARSENUM(NAME, TYPE)		\
441int					\
442NAME(char *s, TYPE *v)			\
443{					\
444    *v = 0;				\
445    for (*v = 0; *s; s++)		\
446	if (isdigit(*s))		\
447	    *v = *v * 10 + *s - '0';	\
448	else				\
449	    return -1;			\
450    return 0;				\
451}
452
453PARSENUM(parseint, u_int)
454PARSENUM(parsesize, size_t)
455PARSENUM(parseoff, off_t)
456
457int
458main(int argc, char *argv[])
459{
460    struct stat sb;
461    struct sigaction sa;
462    char *p, *q, *s;
463    int c, e, r;
464
465    while ((c = getopt(argc, argv,
466		       "146AaB:bc:dFf:h:lHMmnPpo:qRrS:sT:tvw:")) != EOF)
467	switch (c) {
468	case '1':
469	    once_flag = 1;
470	    break;
471	case '4':
472	    family = PF_INET;
473	    break;
474	case '6':
475	    family = PF_INET6;
476	    break;
477	case 'A':
478	    A_flag = 1;
479	    break;
480	case 'a':
481	    a_flag = 1;
482	    break;
483	case 'B':
484	    if (parsesize(optarg, &B_size) == -1)
485		errx(1, "invalid buffer size");
486	    break;
487	case 'b':
488	    warnx("warning: the -b option is deprecated");
489	    b_flag = 1;
490	    break;
491	case 'c':
492	    c_dirname = optarg;
493	    break;
494	case 'd':
495	    d_flag = 1;
496	    break;
497	case 'F':
498	    F_flag = 1;
499	    break;
500	case 'f':
501	    f_filename = optarg;
502	    break;
503	case 'H':
504	    H_flag = 1;
505	    break;
506	case 'h':
507	    h_hostname = optarg;
508	    break;
509	case 'l':
510	    l_flag = 1;
511	    break;
512	case 'o':
513	    o_flag = 1;
514	    o_filename = optarg;
515	    break;
516	case 'M':
517	case 'm':
518	    if (r_flag)
519		errx(1, "the -m and -r flags are mutually exclusive");
520	    m_flag = 1;
521	    break;
522	case 'n':
523	    n_flag = 1;
524	    break;
525	case 'P':
526	case 'p':
527	    p_flag = 1;
528	    break;
529	case 'q':
530	    v_level = 0;
531	    break;
532	case 'R':
533	    R_flag = 1;
534	    break;
535	case 'r':
536	    if (m_flag)
537		errx(1, "the -m and -r flags are mutually exclusive");
538	    r_flag = 1;
539	    break;
540	case 'S':
541	    if (parseoff(optarg, &S_size) == -1)
542		errx(1, "invalid size");
543	    break;
544	case 's':
545	    s_flag = 1;
546	    break;
547	case 'T':
548	    if (parseint(optarg, &T_secs) == -1)
549		errx(1, "invalid timeout");
550	    break;
551	case 't':
552	    t_flag = 1;
553	    warnx("warning: the -t option is deprecated");
554	    break;
555	case 'v':
556	    v_level++;
557	    break;
558	case 'w':
559	    a_flag = 1;
560	    if (parseint(optarg, &w_secs) == -1)
561		errx(1, "invalid delay");
562	    break;
563	default:
564	    usage();
565	    exit(EX_USAGE);
566	}
567
568    argc -= optind;
569    argv += optind;
570
571    if (h_hostname || f_filename || c_dirname) {
572	if (!h_hostname || !f_filename || argc) {
573	    usage();
574	    exit(EX_USAGE);
575	}
576	/* XXX this is a hack. */
577	if (strcspn(h_hostname, "@:/") != strlen(h_hostname))
578	    errx(1, "invalid hostname");
579	if (asprintf(argv, "ftp://%s/%s/%s", h_hostname,
580		     c_dirname ? c_dirname : "", f_filename) == -1)
581	    errx(1, strerror(ENOMEM));
582	argc++;
583    }
584
585    if (!argc) {
586	usage();
587	exit(EX_USAGE);
588    }
589
590    /* allocate buffer */
591    if (B_size < MINBUFSIZE)
592	B_size = MINBUFSIZE;
593    if ((buf = malloc(B_size)) == NULL)
594	errx(1, strerror(ENOMEM));
595
596    /* timeouts */
597    if ((s = getenv("FTP_TIMEOUT")) != NULL) {
598	if (parseint(s, &ftp_timeout) == -1) {
599	    warnx("FTP_TIMEOUT is not a positive integer");
600	    ftp_timeout = 0;
601	}
602    }
603    if ((s = getenv("HTTP_TIMEOUT")) != NULL) {
604	if (parseint(s, &http_timeout) == -1) {
605	    warnx("HTTP_TIMEOUT is not a positive integer");
606	    http_timeout = 0;
607	}
608    }
609
610    /* signal handling */
611    sa.sa_flags = 0;
612    sa.sa_handler = sig_handler;
613    sigemptyset(&sa.sa_mask);
614    sigaction(SIGALRM, &sa, NULL);
615    sa.sa_flags = SA_RESETHAND;
616    sigaction(SIGINT, &sa, NULL);
617    fetchRestartCalls = 0;
618
619    /* output file */
620    if (o_flag) {
621	if (strcmp(o_filename, "-") == 0) {
622	    o_stdout = 1;
623	} else if (stat(o_filename, &sb) == -1) {
624	    if (errno == ENOENT) {
625		if (argc > 1)
626		    errx(EX_USAGE, "%s is not a directory", o_filename);
627	    } else {
628		err(EX_IOERR, "%s", o_filename);
629	    }
630	} else {
631	    if (sb.st_mode & S_IFDIR)
632		o_directory = 1;
633	}
634    }
635
636    /* check if output is to a tty (for progress report) */
637    v_tty = isatty(STDERR_FILENO);
638    r = 0;
639
640    while (argc) {
641	if ((p = strrchr(*argv, '/')) == NULL)
642	    p = *argv;
643	else
644	    p++;
645
646	if (!*p)
647	    p = "fetch.out";
648
649	fetchLastErrCode = 0;
650
651	if (o_flag) {
652	    if (o_stdout) {
653		e = fetch(*argv, "-");
654	    } else if (o_directory) {
655		asprintf(&q, "%s/%s", o_filename, p);
656		e = fetch(*argv, q);
657		free(q);
658	    } else {
659		e = fetch(*argv, o_filename);
660	    }
661	} else {
662	    e = fetch(*argv, p);
663	}
664
665	if (sigint)
666	    kill(getpid(), SIGINT);
667
668	if (e == 0 && once_flag)
669	    exit(0);
670
671	if (e) {
672	    r = 1;
673	    if ((fetchLastErrCode
674		 && fetchLastErrCode != FETCH_UNAVAIL
675		 && fetchLastErrCode != FETCH_MOVED
676		 && fetchLastErrCode != FETCH_URL
677		 && fetchLastErrCode != FETCH_RESOLV
678		 && fetchLastErrCode != FETCH_UNKNOWN)) {
679		if (w_secs) {
680		    if (v_level)
681			fprintf(stderr, "Waiting %d seconds before retrying\n", w_secs);
682		    sleep(w_secs);
683		}
684		if (a_flag)
685		    continue;
686		fprintf(stderr, "Skipping %s\n", *argv);
687	    }
688	}
689
690	argc--, argv++;
691    }
692
693    exit(r);
694}
695