fetch.c revision 69383 
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 69383 2000-11-30 10:08:22Z 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;	/* -[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 || !v_level)
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    if (!v_level)
159	return;
160
161    gettimeofday(&xs->end, NULL);
162
163    stat_display(xs, 1);
164    fputc('\n', stderr);
165    delta = (xs->end.tv_sec + (xs->end.tv_usec / 1.e6))
166	- (xs->start.tv_sec + (xs->start.tv_usec / 1.e6));
167    fprintf(stderr, "%lld bytes transferred in %.1f seconds ",
168	    xs->rcvd - xs->offset, delta);
169    bps = (xs->rcvd - xs->offset) / delta;
170    if (bps > 1024*1024)
171	fprintf(stderr, "(%.2f MBps)\n", bps / (1024*1024));
172    else if (bps > 1024)
173	fprintf(stderr, "(%.2f kBps)\n", bps / 1024);
174    else
175	fprintf(stderr, "(%.2f Bps)\n", bps);
176}
177
178int
179fetch(char *URL, char *path)
180{
181    struct url *url;
182    struct url_stat us;
183    struct stat sb;
184    struct xferstat xs;
185    FILE *f, *of;
186    size_t size;
187    off_t count;
188    char flags[8];
189    int n, r;
190    u_int timeout;
191
192    f = of = NULL;
193
194    /* parse URL */
195    if ((url = fetchParseURL(URL)) == NULL) {
196	warnx("%s: parse error", URL);
197	goto failure;
198    }
199
200    timeout = 0;
201    *flags = 0;
202    count = 0;
203
204    /* common flags */
205    if (v_level > 1)
206	strcat(flags, "v");
207    switch (family) {
208    case PF_INET:
209	strcat(flags, "4");
210	break;
211    case PF_INET6:
212	strcat(flags, "6");
213	break;
214    }
215
216    /* FTP specific flags */
217    if (strcmp(url->scheme, "ftp") == 0) {
218	if (p_flag)
219	    strcat(flags, "p");
220	if (d_flag)
221	    strcat(flags, "d");
222	if (H_flag)
223	    strcat(flags, "h");
224	timeout = T_secs ? T_secs : ftp_timeout;
225    }
226
227    /* HTTP specific flags */
228    if (strcmp(url->scheme, "http") == 0) {
229	if (d_flag)
230	    strcat(flags, "d");
231	if (A_flag)
232	    strcat(flags, "A");
233	timeout = T_secs ? T_secs : http_timeout;
234    }
235
236    /* set the protocol timeout. */
237    fetchTimeout = timeout;
238
239    /* just print size */
240    if (s_flag) {
241	if (fetchStat(url, &us, flags) == -1)
242	    goto failure;
243	if (us.size == -1)
244	    printf("Unknown\n");
245	else
246	    printf("%lld\n", us.size);
247	goto success;
248    }
249
250    /*
251     * If the -r flag was specified, we have to compare the local and
252     * remote files, so we should really do a fetchStat() first, but I
253     * know of at least one HTTP server that only sends the content
254     * size in response to GET requests, and leaves it out of replies
255     * to HEAD requests. Also, in the (frequent) case that the local
256     * and remote files match but the local file is truncated, we have
257     * sufficient information *before* the compare to issue a correct
258     * request. Therefore, we always issue a GET request as if we were
259     * sure the local file was a truncated copy of the remote file; we
260     * can drop the connection later if we change our minds.
261     */
262    if ((r_flag  || m_flag) && !o_stdout && stat(path, &sb) != -1) {
263	if (r_flag)
264	    url->offset = sb.st_size;
265    } else {
266	sb.st_size = -1;
267    }
268
269    /* start the transfer */
270    if ((f = fetchXGet(url, &us, flags)) == NULL) {
271	warnx("%s: %s", path, fetchLastErrString);
272	goto failure;
273    }
274    if (sigint)
275	goto signal;
276
277    /* check that size is as expected */
278    if (S_size) {
279	if (us.size == -1) {
280	    warnx("%s: size unknown", path);
281	    goto failure;
282	} else if (us.size != S_size) {
283	    warnx("%s: size mismatch: expected %lld, actual %lld",
284		  path, S_size, us.size);
285	    goto failure;
286	}
287    }
288
289    /* symlink instead of copy */
290    if (l_flag && strcmp(url->scheme, "file") == 0 && !o_stdout) {
291	if (symlink(url->doc, path) == -1) {
292	    warn("%s: symlink()", path);
293	    goto failure;
294	}
295	goto success;
296    }
297
298    if (v_level > 1) {
299	if (sb.st_size != -1)
300	    fprintf(stderr, "local size / mtime: %lld / %ld\n",
301		    sb.st_size, sb.st_mtime);
302	fprintf(stderr, "remote size / mtime: %lld / %ld\n",
303		us.size, us.mtime);
304    }
305
306    /* open output file */
307    if (o_stdout) {
308	/* output to stdout */
309	of = stdout;
310    } else if (sb.st_size != -1) {
311	/* resume mode, local file exists */
312	if (!F_flag && us.mtime && sb.st_mtime != us.mtime) {
313	    /* no match! have to refetch */
314	    fclose(f);
315	    url->offset = 0;
316	    if ((f = fetchXGet(url, &us, flags)) == NULL) {
317		warnx("%s: %s", path, fetchLastErrString);
318		goto failure;
319	    }
320	    if (sigint)
321		goto signal;
322	} else {
323	    if (us.size == sb.st_size)
324		/* nothing to do */
325		goto success;
326	    if (sb.st_size > us.size) {
327		/* local file too long! */
328		warnx("%s: local file (%lld bytes) is longer "
329		      "than remote file (%lld bytes)",
330		      path, sb.st_size, us.size);
331		goto failure;
332	    }
333	    /* we got through, open local file and seek to offset */
334	    /*
335	     * XXX there's a race condition here - the file we open is not
336	     * necessarily the same as the one we stat()'ed earlier...
337	     */
338	    if ((of = fopen(path, "a")) == NULL) {
339		warn("%s: fopen()", path);
340		goto failure;
341	    }
342	    if (fseek(of, url->offset, SEEK_SET) == -1) {
343		warn("%s: fseek()", path);
344		goto failure;
345	    }
346	}
347    }
348    if (m_flag && sb.st_size != -1) {
349	/* mirror mode, local file exists */
350	if (sb.st_size == us.size && sb.st_mtime == us.mtime)
351	    goto success;
352    }
353    if (!of) {
354	/*
355	 * We don't yet have an output file; either this is a vanilla
356	 * run with no special flags, or the local and remote files
357	 * didn't match.
358	 */
359	if ((of = fopen(path, "w")) == NULL) {
360	    warn("%s: open()", path);
361	    goto failure;
362	}
363    }
364    count = url->offset;
365
366    /* start the counter */
367    stat_start(&xs, path, us.size, count);
368
369    sigint = sigalrm = 0;
370
371    /* suck in the data */
372    for (n = 0; !sigint && !sigalrm; ++n) {
373	if (us.size != -1 && us.size - count < B_size)
374	    size = us.size - count;
375	else
376	    size = B_size;
377	if (timeout)
378	    alarm(timeout);
379	if ((size = fread(buf, 1, size, f)) <= 0)
380	    break;
381	stat_update(&xs, count += size, 0);
382	if (fwrite(buf, size, 1, of) != 1)
383	    break;
384    }
385
386    if (timeout)
387	alarm(0);
388
389    stat_end(&xs);
390
391    /* set mtime of local file */
392    if (!n_flag && us.mtime && !o_stdout
393	&& (stat(path, &sb) != -1) && sb.st_mode & S_IFREG) {
394	struct timeval tv[2];
395
396	fflush(of);
397	tv[0].tv_sec = (long)(us.atime ? us.atime : us.mtime);
398	tv[1].tv_sec = (long)us.mtime;
399	tv[0].tv_usec = tv[1].tv_usec = 0;
400	if (utimes(path, tv))
401	    warn("%s: utimes()", path);
402    }
403
404    /* timed out or interrupted? */
405 signal:
406    if (sigalrm)
407	warnx("transfer timed out");
408    if (sigint) {
409	warnx("transfer interrupted");
410	goto failure;
411    }
412
413    if (!sigalrm) {
414	/* check the status of our files */
415	if (ferror(f))
416	    warn("%s", URL);
417	if (ferror(of))
418	    warn("%s", path);
419	if (ferror(f) || ferror(of))
420	    goto failure;
421    }
422
423    /* did the transfer complete normally? */
424    if (us.size != -1 && count < us.size) {
425	warnx("%s appears to be truncated: %lld/%lld bytes",
426	      path, count, us.size);
427	goto failure_keep;
428    }
429
430    /*
431     * If the transfer timed out and we didn't know how much to
432     * expect, assume the worst (i.e. we didn't get all of it)
433     */
434    if (sigalrm && us.size == -1) {
435	warnx("%s may be truncated", path);
436	goto failure_keep;
437    }
438
439 success:
440    r = 0;
441    goto done;
442 failure:
443    if (of && of != stdout && !R_flag && !r_flag)
444	if (stat(path, &sb) != -1 && (sb.st_mode & S_IFREG))
445	    unlink(path);
446 failure_keep:
447    r = -1;
448    goto done;
449 done:
450    if (f)
451	fclose(f);
452    if (of && of != stdout)
453	fclose(of);
454    if (url)
455	fetchFreeURL(url);
456    return r;
457}
458
459void
460usage(void)
461{
462    /* XXX badly out of synch */
463    fprintf(stderr,
464	    "Usage: fetch [-1AFHMPRabdlmnpqrstv] [-o outputfile] [-S bytes]\n"
465	    "             [-B bytes] [-T seconds] [-w seconds]\n"
466	    "             [-f file -h host [-c dir] | URL ...]\n"
467	);
468}
469
470
471#define PARSENUM(NAME, TYPE)		\
472int					\
473NAME(char *s, TYPE *v)			\
474{					\
475    *v = 0;				\
476    for (*v = 0; *s; s++)		\
477	if (isdigit(*s))		\
478	    *v = *v * 10 + *s - '0';	\
479	else				\
480	    return -1;			\
481    return 0;				\
482}
483
484PARSENUM(parseint, u_int)
485PARSENUM(parsesize, size_t)
486PARSENUM(parseoff, off_t)
487
488int
489main(int argc, char *argv[])
490{
491    struct stat sb;
492    struct sigaction sa;
493    char *p, *q, *s;
494    int c, e, r;
495
496    while ((c = getopt(argc, argv,
497		       "146AaB:bc:dFf:h:lHMmnPpo:qRrS:sT:tvw:")) != EOF)
498	switch (c) {
499	case '1':
500	    once_flag = 1;
501	    break;
502	case '4':
503	    family = PF_INET;
504	    break;
505	case '6':
506	    family = PF_INET6;
507	    break;
508	case 'A':
509	    A_flag = 1;
510	    break;
511	case 'a':
512	    a_flag = 1;
513	    break;
514	case 'B':
515	    if (parsesize(optarg, &B_size) == -1)
516		errx(1, "invalid buffer size");
517	    break;
518	case 'b':
519	    warnx("warning: the -b option is deprecated");
520	    b_flag = 1;
521	    break;
522	case 'c':
523	    c_dirname = optarg;
524	    break;
525	case 'd':
526	    d_flag = 1;
527	    break;
528	case 'F':
529	    F_flag = 1;
530	    break;
531	case 'f':
532	    f_filename = optarg;
533	    break;
534	case 'H':
535	    H_flag = 1;
536	    break;
537	case 'h':
538	    h_hostname = optarg;
539	    break;
540	case 'l':
541	    l_flag = 1;
542	    break;
543	case 'o':
544	    o_flag = 1;
545	    o_filename = optarg;
546	    break;
547	case 'M':
548	case 'm':
549	    if (r_flag)
550		errx(1, "the -m and -r flags are mutually exclusive");
551	    m_flag = 1;
552	    break;
553	case 'n':
554	    n_flag = 1;
555	    break;
556	case 'P':
557	case 'p':
558	    p_flag = 1;
559	    break;
560	case 'q':
561	    v_level = 0;
562	    break;
563	case 'R':
564	    R_flag = 1;
565	    break;
566	case 'r':
567	    if (m_flag)
568		errx(1, "the -m and -r flags are mutually exclusive");
569	    r_flag = 1;
570	    break;
571	case 'S':
572	    if (parseoff(optarg, &S_size) == -1)
573		errx(1, "invalid size");
574	    break;
575	case 's':
576	    s_flag = 1;
577	    break;
578	case 'T':
579	    if (parseint(optarg, &T_secs) == -1)
580		errx(1, "invalid timeout");
581	    break;
582	case 't':
583	    t_flag = 1;
584	    warnx("warning: the -t option is deprecated");
585	    break;
586	case 'v':
587	    v_level++;
588	    break;
589	case 'w':
590	    a_flag = 1;
591	    if (parseint(optarg, &w_secs) == -1)
592		errx(1, "invalid delay");
593	    break;
594	default:
595	    usage();
596	    exit(EX_USAGE);
597	}
598
599    argc -= optind;
600    argv += optind;
601
602    if (h_hostname || f_filename || c_dirname) {
603	if (!h_hostname || !f_filename || argc) {
604	    usage();
605	    exit(EX_USAGE);
606	}
607	/* XXX this is a hack. */
608	if (strcspn(h_hostname, "@:/") != strlen(h_hostname))
609	    errx(1, "invalid hostname");
610	if (asprintf(argv, "ftp://%s/%s/%s", h_hostname,
611		     c_dirname ? c_dirname : "", f_filename) == -1)
612	    errx(1, "%s", strerror(ENOMEM));
613	argc++;
614    }
615
616    if (!argc) {
617	usage();
618	exit(EX_USAGE);
619    }
620
621    /* allocate buffer */
622    if (B_size < MINBUFSIZE)
623	B_size = MINBUFSIZE;
624    if ((buf = malloc(B_size)) == NULL)
625	errx(1, "%s", strerror(ENOMEM));
626
627    /* timeouts */
628    if ((s = getenv("FTP_TIMEOUT")) != NULL) {
629	if (parseint(s, &ftp_timeout) == -1) {
630	    warnx("FTP_TIMEOUT is not a positive integer");
631	    ftp_timeout = 0;
632	}
633    }
634    if ((s = getenv("HTTP_TIMEOUT")) != NULL) {
635	if (parseint(s, &http_timeout) == -1) {
636	    warnx("HTTP_TIMEOUT is not a positive integer");
637	    http_timeout = 0;
638	}
639    }
640
641    /* signal handling */
642    sa.sa_flags = 0;
643    sa.sa_handler = sig_handler;
644    sigemptyset(&sa.sa_mask);
645    sigaction(SIGALRM, &sa, NULL);
646    sa.sa_flags = SA_RESETHAND;
647    sigaction(SIGINT, &sa, NULL);
648    fetchRestartCalls = 0;
649
650    /* output file */
651    if (o_flag) {
652	if (strcmp(o_filename, "-") == 0) {
653	    o_stdout = 1;
654	} else if (stat(o_filename, &sb) == -1) {
655	    if (errno == ENOENT) {
656		if (argc > 1)
657		    errx(EX_USAGE, "%s is not a directory", o_filename);
658	    } else {
659		err(EX_IOERR, "%s", o_filename);
660	    }
661	} else {
662	    if (sb.st_mode & S_IFDIR)
663		o_directory = 1;
664	}
665    }
666
667    /* check if output is to a tty (for progress report) */
668    v_tty = isatty(STDERR_FILENO);
669    r = 0;
670
671    while (argc) {
672	if ((p = strrchr(*argv, '/')) == NULL)
673	    p = *argv;
674	else
675	    p++;
676
677	if (!*p)
678	    p = "fetch.out";
679
680	fetchLastErrCode = 0;
681
682	if (o_flag) {
683	    if (o_stdout) {
684		e = fetch(*argv, "-");
685	    } else if (o_directory) {
686		asprintf(&q, "%s/%s", o_filename, p);
687		e = fetch(*argv, q);
688		free(q);
689	    } else {
690		e = fetch(*argv, o_filename);
691	    }
692	} else {
693	    e = fetch(*argv, p);
694	}
695
696	if (sigint)
697	    kill(getpid(), SIGINT);
698
699	if (e == 0 && once_flag)
700	    exit(0);
701
702	if (e) {
703	    r = 1;
704	    if ((fetchLastErrCode
705		 && fetchLastErrCode != FETCH_UNAVAIL
706		 && fetchLastErrCode != FETCH_MOVED
707		 && fetchLastErrCode != FETCH_URL
708		 && fetchLastErrCode != FETCH_RESOLV
709		 && fetchLastErrCode != FETCH_UNKNOWN)) {
710		if (w_secs) {
711		    if (v_level)
712			fprintf(stderr, "Waiting %d seconds before retrying\n",
713				w_secs);
714		    sleep(w_secs);
715		}
716		if (a_flag)
717		    continue;
718	    }
719	}
720
721	argc--, argv++;
722    }
723
724    exit(r);
725}
726