fetch.c revision 63877 
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 63877 2000-07-26 07:28:33Z 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 && !o_stdout && stat(path, &sb) != -1)
263	url->offset = sb.st_size;
264    else
265	sb.st_size = 0;
266
267    /* start the transfer */
268    if ((f = fetchXGet(url, &us, flags)) == NULL) {
269	warnx("%s: %s", path, fetchLastErrString);
270	goto failure;
271    }
272    if (sigint)
273	goto signal;
274
275    /* check that size is as expected */
276    if (S_size) {
277	if (us.size == -1) {
278	    warnx("%s: size unknown", path);
279	    goto failure;
280	} else if (us.size != S_size) {
281	    warnx("%s: size mismatch: expected %lld, actual %lld",
282		  path, S_size, us.size);
283	    goto failure;
284	}
285    }
286
287    /* symlink instead of copy */
288    if (l_flag && strcmp(url->scheme, "file") == 0 && !o_stdout) {
289	if (symlink(url->doc, path) == -1) {
290	    warn("%s: symlink()", path);
291	    goto failure;
292	}
293	goto success;
294    }
295
296    if (v_level > 1) {
297	if (sb.st_size)
298	    warnx("local: %lld / %ld", sb.st_size, sb.st_mtime);
299	warnx("remote: %lld / %ld", us.size, us.mtime);
300    }
301
302    /* open output file */
303    if (o_stdout) {
304	/* output to stdout */
305	of = stdout;
306    } else if (sb.st_size) {
307	/* resume mode, local file exists */
308	if (!F_flag && us.mtime && sb.st_mtime != us.mtime) {
309	    /* no match! have to refetch */
310	    fclose(f);
311	    url->offset = 0;
312	    if ((f = fetchXGet(url, &us, flags)) == NULL) {
313		warnx("%s: %s", path, fetchLastErrString);
314		goto failure;
315	    }
316	    if (sigint)
317		goto signal;
318	} else {
319	    if (us.size == sb.st_size)
320		/* nothing to do */
321		goto success;
322	    if (sb.st_size > us.size) {
323		/* local file too long! */
324		warnx("%s: local file (%lld bytes) is longer "
325		      "than remote file (%lld bytes)",
326		      path, sb.st_size, us.size);
327		goto failure;
328	    }
329	    /* we got through, open local file and seek to offset */
330	    /*
331	     * XXX there's a race condition here - the file we open is not
332	     * necessarily the same as the one we stat()'ed earlier...
333	     */
334	    if ((of = fopen(path, "a")) == NULL) {
335		warn("%s: fopen()", path);
336		goto failure;
337	    }
338	    if (fseek(of, url->offset, SEEK_SET) == -1) {
339		warn("%s: fseek()", path);
340		goto failure;
341	    }
342	}
343    }
344    if (m_flag && stat(path, &sb) != -1) {
345	/* mirror mode, local file exists */
346	if (sb.st_size == us.size && sb.st_mtime == us.mtime)
347	    goto success;
348    }
349    if (!of) {
350	/*
351	 * We don't yet have an output file; either this is a vanilla
352	 * run with no special flags, or the local and remote files
353	 * didn't match.
354	 */
355	if ((of = fopen(path, "w")) == NULL) {
356	    warn("%s: open()", path);
357	    goto failure;
358	}
359    }
360    count = url->offset;
361
362    /* start the counter */
363    stat_start(&xs, path, us.size, count);
364
365    sigint = sigalrm = 0;
366
367    /* suck in the data */
368    for (n = 0; !sigint && !sigalrm; ++n) {
369	if (us.size != -1 && us.size - count < B_size)
370	    size = us.size - count;
371	else
372	    size = B_size;
373	if (timeout)
374	    alarm(timeout);
375	if ((size = fread(buf, 1, size, f)) <= 0)
376	    break;
377	stat_update(&xs, count += size, 0);
378	if (fwrite(buf, size, 1, of) != 1)
379	    break;
380    }
381
382    if (timeout)
383	alarm(0);
384
385    stat_end(&xs);
386
387    /* Set mtime of local file */
388    if (!n_flag && us.mtime && !o_stdout) {
389	struct timeval tv[2];
390
391	fflush(of);
392	tv[0].tv_sec = (long)(us.atime ? us.atime : us.mtime);
393	tv[1].tv_sec = (long)us.mtime;
394	tv[0].tv_usec = tv[1].tv_usec = 0;
395	if (utimes(path, tv))
396	    warn("%s: utimes()", path);
397    }
398
399    /* timed out or interrupted? */
400 signal:
401    if (sigalrm)
402	warnx("transfer timed out");
403    if (sigint) {
404	warnx("transfer interrupted");
405	goto failure;
406    }
407
408    if (!sigalrm) {
409	/* check the status of our files */
410	if (ferror(f))
411	    warn("%s", URL);
412	if (ferror(of))
413	    warn("%s", path);
414	if (ferror(f) || ferror(of))
415	    goto failure;
416    }
417
418    /* did the transfer complete normally? */
419    if (us.size != -1 && count < us.size) {
420	warnx("%s appears to be truncated: %lld/%lld bytes",
421	      path, count, us.size);
422	goto failure_keep;
423    }
424
425 success:
426    r = 0;
427    goto done;
428 failure:
429    if (of && of != stdout && !R_flag && !r_flag)
430	unlink(path);
431 failure_keep:
432    r = -1;
433    goto done;
434 done:
435    if (f)
436	fclose(f);
437    if (of && of != stdout)
438	fclose(of);
439    if (url)
440	fetchFreeURL(url);
441    return r;
442}
443
444void
445usage(void)
446{
447    /* XXX badly out of synch */
448    fprintf(stderr,
449	    "Usage: fetch [-1AFHMPRabdlmnpqrstv] [-o outputfile] [-S bytes]\n"
450	    "             [-B bytes] [-T seconds] [-w seconds]\n"
451	    "             [-f file -h host [-c dir] | URL ...]\n"
452	);
453}
454
455
456#define PARSENUM(NAME, TYPE)		\
457int					\
458NAME(char *s, TYPE *v)			\
459{					\
460    *v = 0;				\
461    for (*v = 0; *s; s++)		\
462	if (isdigit(*s))		\
463	    *v = *v * 10 + *s - '0';	\
464	else				\
465	    return -1;			\
466    return 0;				\
467}
468
469PARSENUM(parseint, u_int)
470PARSENUM(parsesize, size_t)
471PARSENUM(parseoff, off_t)
472
473int
474main(int argc, char *argv[])
475{
476    struct stat sb;
477    struct sigaction sa;
478    char *p, *q, *s;
479    int c, e, r;
480
481    while ((c = getopt(argc, argv,
482		       "146AaB:bc:dFf:h:lHMmnPpo:qRrS:sT:tvw:")) != EOF)
483	switch (c) {
484	case '1':
485	    once_flag = 1;
486	    break;
487	case '4':
488	    family = PF_INET;
489	    break;
490	case '6':
491	    family = PF_INET6;
492	    break;
493	case 'A':
494	    A_flag = 1;
495	    break;
496	case 'a':
497	    a_flag = 1;
498	    break;
499	case 'B':
500	    if (parsesize(optarg, &B_size) == -1)
501		errx(1, "invalid buffer size");
502	    break;
503	case 'b':
504	    warnx("warning: the -b option is deprecated");
505	    b_flag = 1;
506	    break;
507	case 'c':
508	    c_dirname = optarg;
509	    break;
510	case 'd':
511	    d_flag = 1;
512	    break;
513	case 'F':
514	    F_flag = 1;
515	    break;
516	case 'f':
517	    f_filename = optarg;
518	    break;
519	case 'H':
520	    H_flag = 1;
521	    break;
522	case 'h':
523	    h_hostname = optarg;
524	    break;
525	case 'l':
526	    l_flag = 1;
527	    break;
528	case 'o':
529	    o_flag = 1;
530	    o_filename = optarg;
531	    break;
532	case 'M':
533	case 'm':
534	    if (r_flag)
535		errx(1, "the -m and -r flags are mutually exclusive");
536	    m_flag = 1;
537	    break;
538	case 'n':
539	    n_flag = 1;
540	    break;
541	case 'P':
542	case 'p':
543	    p_flag = 1;
544	    break;
545	case 'q':
546	    v_level = 0;
547	    break;
548	case 'R':
549	    R_flag = 1;
550	    break;
551	case 'r':
552	    if (m_flag)
553		errx(1, "the -m and -r flags are mutually exclusive");
554	    r_flag = 1;
555	    break;
556	case 'S':
557	    if (parseoff(optarg, &S_size) == -1)
558		errx(1, "invalid size");
559	    break;
560	case 's':
561	    s_flag = 1;
562	    break;
563	case 'T':
564	    if (parseint(optarg, &T_secs) == -1)
565		errx(1, "invalid timeout");
566	    break;
567	case 't':
568	    t_flag = 1;
569	    warnx("warning: the -t option is deprecated");
570	    break;
571	case 'v':
572	    v_level++;
573	    break;
574	case 'w':
575	    a_flag = 1;
576	    if (parseint(optarg, &w_secs) == -1)
577		errx(1, "invalid delay");
578	    break;
579	default:
580	    usage();
581	    exit(EX_USAGE);
582	}
583
584    argc -= optind;
585    argv += optind;
586
587    if (h_hostname || f_filename || c_dirname) {
588	if (!h_hostname || !f_filename || argc) {
589	    usage();
590	    exit(EX_USAGE);
591	}
592	/* XXX this is a hack. */
593	if (strcspn(h_hostname, "@:/") != strlen(h_hostname))
594	    errx(1, "invalid hostname");
595	if (asprintf(argv, "ftp://%s/%s/%s", h_hostname,
596		     c_dirname ? c_dirname : "", f_filename) == -1)
597	    errx(1, strerror(ENOMEM));
598	argc++;
599    }
600
601    if (!argc) {
602	usage();
603	exit(EX_USAGE);
604    }
605
606    /* allocate buffer */
607    if (B_size < MINBUFSIZE)
608	B_size = MINBUFSIZE;
609    if ((buf = malloc(B_size)) == NULL)
610	errx(1, strerror(ENOMEM));
611
612    /* timeouts */
613    if ((s = getenv("FTP_TIMEOUT")) != NULL) {
614	if (parseint(s, &ftp_timeout) == -1) {
615	    warnx("FTP_TIMEOUT is not a positive integer");
616	    ftp_timeout = 0;
617	}
618    }
619    if ((s = getenv("HTTP_TIMEOUT")) != NULL) {
620	if (parseint(s, &http_timeout) == -1) {
621	    warnx("HTTP_TIMEOUT is not a positive integer");
622	    http_timeout = 0;
623	}
624    }
625
626    /* signal handling */
627    sa.sa_flags = 0;
628    sa.sa_handler = sig_handler;
629    sigemptyset(&sa.sa_mask);
630    sigaction(SIGALRM, &sa, NULL);
631    sa.sa_flags = SA_RESETHAND;
632    sigaction(SIGINT, &sa, NULL);
633    fetchRestartCalls = 0;
634
635    /* output file */
636    if (o_flag) {
637	if (strcmp(o_filename, "-") == 0) {
638	    o_stdout = 1;
639	} else if (stat(o_filename, &sb) == -1) {
640	    if (errno == ENOENT) {
641		if (argc > 1)
642		    errx(EX_USAGE, "%s is not a directory", o_filename);
643	    } else {
644		err(EX_IOERR, "%s", o_filename);
645	    }
646	} else {
647	    if (sb.st_mode & S_IFDIR)
648		o_directory = 1;
649	}
650    }
651
652    /* check if output is to a tty (for progress report) */
653    v_tty = isatty(STDERR_FILENO);
654    r = 0;
655
656    while (argc) {
657	if ((p = strrchr(*argv, '/')) == NULL)
658	    p = *argv;
659	else
660	    p++;
661
662	if (!*p)
663	    p = "fetch.out";
664
665	fetchLastErrCode = 0;
666
667	if (o_flag) {
668	    if (o_stdout) {
669		e = fetch(*argv, "-");
670	    } else if (o_directory) {
671		asprintf(&q, "%s/%s", o_filename, p);
672		e = fetch(*argv, q);
673		free(q);
674	    } else {
675		e = fetch(*argv, o_filename);
676	    }
677	} else {
678	    e = fetch(*argv, p);
679	}
680
681	if (sigint)
682	    kill(getpid(), SIGINT);
683
684	if (e == 0 && once_flag)
685	    exit(0);
686
687	if (e) {
688	    r = 1;
689	    if ((fetchLastErrCode
690		 && fetchLastErrCode != FETCH_UNAVAIL
691		 && fetchLastErrCode != FETCH_MOVED
692		 && fetchLastErrCode != FETCH_URL
693		 && fetchLastErrCode != FETCH_RESOLV
694		 && fetchLastErrCode != FETCH_UNKNOWN)) {
695		if (w_secs) {
696		    if (v_level)
697			fprintf(stderr, "Waiting %d seconds before retrying\n",
698				w_secs);
699		    sleep(w_secs);
700		}
701		if (a_flag)
702		    continue;
703	    }
704	}
705
706	argc--, argv++;
707    }
708
709    exit(r);
710}
711