9e052883 |
1 | /* |
2 | * untgz.c -- Display contents and extract files from a gzip'd TAR file |
3 | * |
4 | * written by Pedro A. Aranda Gutierrez <paag@tid.es> |
5 | * adaptation to Unix by Jean-loup Gailly <jloup@gzip.org> |
6 | * various fixes by Cosmin Truta <cosmint@cs.ubbcluj.ro> |
648db22b |
7 | * |
8 | * This software is provided 'as-is', without any express or implied |
9 | * warranty. In no event will the authors be held liable for any damages |
10 | * arising from the use of this software. |
11 | * |
12 | * Permission is granted to anyone to use this software for any purpose, |
13 | * including commercial applications, and to alter it and redistribute it |
14 | * freely, subject to the following restrictions: |
15 | * |
16 | * 1. The origin of this software must not be misrepresented; you must not |
17 | * claim that you wrote the original software. If you use this software |
18 | * in a product, an acknowledgment in the product documentation would be |
19 | * appreciated but is not required. |
20 | * 2. Altered source versions must be plainly marked as such, and must not be |
21 | * misrepresented as being the original software. |
22 | * 3. This notice may not be removed or altered from any source distribution. |
9e052883 |
23 | */ |
24 | |
25 | #include <stdio.h> |
26 | #include <stdlib.h> |
27 | #include <string.h> |
28 | #include <time.h> |
29 | #include <errno.h> |
30 | |
31 | #include "zlib.h" |
32 | |
648db22b |
33 | #ifdef _WIN32 |
9e052883 |
34 | # include <direct.h> |
35 | # include <io.h> |
648db22b |
36 | # include <windows.h> |
9e052883 |
37 | # ifndef F_OK |
38 | # define F_OK 0 |
39 | # endif |
40 | # define mkdir(dirname,mode) _mkdir(dirname) |
41 | # ifdef _MSC_VER |
42 | # define access(path,mode) _access(path,mode) |
43 | # define chmod(path,mode) _chmod(path,mode) |
44 | # define strdup(str) _strdup(str) |
45 | # endif |
46 | #else |
648db22b |
47 | # include <sys/stat.h> |
48 | # include <unistd.h> |
9e052883 |
49 | # include <utime.h> |
50 | #endif |
51 | |
52 | |
53 | /* values used in typeflag field */ |
54 | |
55 | #define REGTYPE '0' /* regular file */ |
56 | #define AREGTYPE '\0' /* regular file */ |
57 | #define LNKTYPE '1' /* link */ |
58 | #define SYMTYPE '2' /* reserved */ |
59 | #define CHRTYPE '3' /* character special */ |
60 | #define BLKTYPE '4' /* block special */ |
61 | #define DIRTYPE '5' /* directory */ |
62 | #define FIFOTYPE '6' /* FIFO special */ |
63 | #define CONTTYPE '7' /* reserved */ |
64 | |
65 | /* GNU tar extensions */ |
66 | |
67 | #define GNUTYPE_DUMPDIR 'D' /* file names from dumped directory */ |
68 | #define GNUTYPE_LONGLINK 'K' /* long link name */ |
69 | #define GNUTYPE_LONGNAME 'L' /* long file name */ |
70 | #define GNUTYPE_MULTIVOL 'M' /* continuation of file from another volume */ |
71 | #define GNUTYPE_NAMES 'N' /* file name that does not fit into main hdr */ |
72 | #define GNUTYPE_SPARSE 'S' /* sparse file */ |
73 | #define GNUTYPE_VOLHDR 'V' /* tape/volume header */ |
74 | |
75 | |
76 | /* tar header */ |
77 | |
78 | #define BLOCKSIZE 512 |
79 | #define SHORTNAMESIZE 100 |
80 | |
81 | struct tar_header |
82 | { /* byte offset */ |
83 | char name[100]; /* 0 */ |
84 | char mode[8]; /* 100 */ |
85 | char uid[8]; /* 108 */ |
86 | char gid[8]; /* 116 */ |
87 | char size[12]; /* 124 */ |
88 | char mtime[12]; /* 136 */ |
89 | char chksum[8]; /* 148 */ |
90 | char typeflag; /* 156 */ |
91 | char linkname[100]; /* 157 */ |
92 | char magic[6]; /* 257 */ |
93 | char version[2]; /* 263 */ |
94 | char uname[32]; /* 265 */ |
95 | char gname[32]; /* 297 */ |
96 | char devmajor[8]; /* 329 */ |
97 | char devminor[8]; /* 337 */ |
98 | char prefix[155]; /* 345 */ |
99 | /* 500 */ |
100 | }; |
101 | |
102 | union tar_buffer |
103 | { |
104 | char buffer[BLOCKSIZE]; |
105 | struct tar_header header; |
106 | }; |
107 | |
108 | struct attr_item |
109 | { |
110 | struct attr_item *next; |
111 | char *fname; |
112 | int mode; |
113 | time_t time; |
114 | }; |
115 | |
116 | enum { TGZ_EXTRACT, TGZ_LIST, TGZ_INVALID }; |
117 | |
9e052883 |
118 | char *prog; |
119 | |
648db22b |
120 | void error(const char *msg) |
121 | { |
122 | fprintf(stderr, "%s: %s\n", prog, msg); |
123 | exit(1); |
124 | } |
125 | |
9e052883 |
126 | const char *TGZsuffix[] = { "\0", ".tar", ".tar.gz", ".taz", ".tgz", NULL }; |
127 | |
128 | /* return the file name of the TGZ archive */ |
129 | /* or NULL if it does not exist */ |
130 | |
131 | char *TGZfname (const char *arcname) |
132 | { |
133 | static char buffer[1024]; |
134 | int origlen,i; |
135 | |
136 | strcpy(buffer,arcname); |
137 | origlen = strlen(buffer); |
138 | |
139 | for (i=0; TGZsuffix[i]; i++) |
140 | { |
141 | strcpy(buffer+origlen,TGZsuffix[i]); |
142 | if (access(buffer,F_OK) == 0) |
143 | return buffer; |
144 | } |
145 | return NULL; |
146 | } |
147 | |
148 | |
149 | /* error message for the filename */ |
150 | |
151 | void TGZnotfound (const char *arcname) |
152 | { |
153 | int i; |
154 | |
155 | fprintf(stderr,"%s: Couldn't find ",prog); |
156 | for (i=0;TGZsuffix[i];i++) |
157 | fprintf(stderr,(TGZsuffix[i+1]) ? "%s%s, " : "or %s%s\n", |
158 | arcname, |
159 | TGZsuffix[i]); |
160 | exit(1); |
161 | } |
162 | |
163 | |
164 | /* convert octal digits to int */ |
165 | /* on error return -1 */ |
166 | |
167 | int getoct (char *p,int width) |
168 | { |
169 | int result = 0; |
170 | char c; |
171 | |
172 | while (width--) |
173 | { |
174 | c = *p++; |
175 | if (c == 0) |
176 | break; |
177 | if (c == ' ') |
178 | continue; |
179 | if (c < '0' || c > '7') |
180 | return -1; |
181 | result = result * 8 + (c - '0'); |
182 | } |
183 | return result; |
184 | } |
185 | |
186 | |
187 | /* convert time_t to string */ |
188 | /* use the "YYYY/MM/DD hh:mm:ss" format */ |
189 | |
190 | char *strtime (time_t *t) |
191 | { |
192 | struct tm *local; |
193 | static char result[32]; |
194 | |
195 | local = localtime(t); |
196 | sprintf(result,"%4d/%02d/%02d %02d:%02d:%02d", |
197 | local->tm_year+1900, local->tm_mon+1, local->tm_mday, |
198 | local->tm_hour, local->tm_min, local->tm_sec); |
199 | return result; |
200 | } |
201 | |
202 | |
203 | /* set file time */ |
204 | |
205 | int setfiletime (char *fname,time_t ftime) |
206 | { |
648db22b |
207 | #ifdef _WIN32 |
9e052883 |
208 | static int isWinNT = -1; |
209 | SYSTEMTIME st; |
210 | FILETIME locft, modft; |
211 | struct tm *loctm; |
212 | HANDLE hFile; |
213 | int result; |
214 | |
215 | loctm = localtime(&ftime); |
216 | if (loctm == NULL) |
217 | return -1; |
218 | |
219 | st.wYear = (WORD)loctm->tm_year + 1900; |
220 | st.wMonth = (WORD)loctm->tm_mon + 1; |
221 | st.wDayOfWeek = (WORD)loctm->tm_wday; |
222 | st.wDay = (WORD)loctm->tm_mday; |
223 | st.wHour = (WORD)loctm->tm_hour; |
224 | st.wMinute = (WORD)loctm->tm_min; |
225 | st.wSecond = (WORD)loctm->tm_sec; |
226 | st.wMilliseconds = 0; |
227 | if (!SystemTimeToFileTime(&st, &locft) || |
228 | !LocalFileTimeToFileTime(&locft, &modft)) |
229 | return -1; |
230 | |
231 | if (isWinNT < 0) |
232 | isWinNT = (GetVersion() < 0x80000000) ? 1 : 0; |
233 | hFile = CreateFile(fname, GENERIC_WRITE, 0, NULL, OPEN_EXISTING, |
234 | (isWinNT ? FILE_FLAG_BACKUP_SEMANTICS : 0), |
235 | NULL); |
236 | if (hFile == INVALID_HANDLE_VALUE) |
237 | return -1; |
238 | result = SetFileTime(hFile, NULL, NULL, &modft) ? 0 : -1; |
239 | CloseHandle(hFile); |
240 | return result; |
241 | #else |
242 | struct utimbuf settime; |
243 | |
244 | settime.actime = settime.modtime = ftime; |
245 | return utime(fname,&settime); |
246 | #endif |
247 | } |
248 | |
249 | |
250 | /* push file attributes */ |
251 | |
252 | void push_attr(struct attr_item **list,char *fname,int mode,time_t time) |
253 | { |
254 | struct attr_item *item; |
255 | |
256 | item = (struct attr_item *)malloc(sizeof(struct attr_item)); |
257 | if (item == NULL) |
258 | error("Out of memory"); |
259 | item->fname = strdup(fname); |
260 | item->mode = mode; |
261 | item->time = time; |
262 | item->next = *list; |
263 | *list = item; |
264 | } |
265 | |
266 | |
267 | /* restore file attributes */ |
268 | |
269 | void restore_attr(struct attr_item **list) |
270 | { |
271 | struct attr_item *item, *prev; |
272 | |
273 | for (item = *list; item != NULL; ) |
274 | { |
275 | setfiletime(item->fname,item->time); |
276 | chmod(item->fname,item->mode); |
277 | prev = item; |
278 | item = item->next; |
279 | free(prev); |
280 | } |
281 | *list = NULL; |
282 | } |
283 | |
284 | |
285 | /* match regular expression */ |
286 | |
287 | #define ISSPECIAL(c) (((c) == '*') || ((c) == '/')) |
288 | |
289 | int ExprMatch (char *string,char *expr) |
290 | { |
291 | while (1) |
292 | { |
293 | if (ISSPECIAL(*expr)) |
294 | { |
295 | if (*expr == '/') |
296 | { |
297 | if (*string != '\\' && *string != '/') |
298 | return 0; |
299 | string ++; expr++; |
300 | } |
301 | else if (*expr == '*') |
302 | { |
303 | if (*expr ++ == 0) |
304 | return 1; |
305 | while (*++string != *expr) |
306 | if (*string == 0) |
307 | return 0; |
308 | } |
309 | } |
310 | else |
311 | { |
312 | if (*string != *expr) |
313 | return 0; |
314 | if (*expr++ == 0) |
315 | return 1; |
316 | string++; |
317 | } |
318 | } |
319 | } |
320 | |
321 | |
322 | /* recursive mkdir */ |
323 | /* abort on ENOENT; ignore other errors like "directory already exists" */ |
324 | /* return 1 if OK */ |
325 | /* 0 on error */ |
326 | |
327 | int makedir (char *newdir) |
328 | { |
329 | char *buffer = strdup(newdir); |
330 | char *p; |
331 | int len = strlen(buffer); |
332 | |
333 | if (len <= 0) { |
334 | free(buffer); |
335 | return 0; |
336 | } |
337 | if (buffer[len-1] == '/') { |
338 | buffer[len-1] = '\0'; |
339 | } |
340 | if (mkdir(buffer, 0755) == 0) |
341 | { |
342 | free(buffer); |
343 | return 1; |
344 | } |
345 | |
346 | p = buffer+1; |
347 | while (1) |
348 | { |
349 | char hold; |
350 | |
351 | while(*p && *p != '\\' && *p != '/') |
352 | p++; |
353 | hold = *p; |
354 | *p = 0; |
355 | if ((mkdir(buffer, 0755) == -1) && (errno == ENOENT)) |
356 | { |
357 | fprintf(stderr,"%s: Couldn't create directory %s\n",prog,buffer); |
358 | free(buffer); |
359 | return 0; |
360 | } |
361 | if (hold == 0) |
362 | break; |
363 | *p++ = hold; |
364 | } |
365 | free(buffer); |
366 | return 1; |
367 | } |
368 | |
369 | |
370 | int matchname (int arg,int argc,char **argv,char *fname) |
371 | { |
372 | if (arg == argc) /* no arguments given (untgz tgzarchive) */ |
373 | return 1; |
374 | |
375 | while (arg < argc) |
376 | if (ExprMatch(fname,argv[arg++])) |
377 | return 1; |
378 | |
379 | return 0; /* ignore this for the moment being */ |
380 | } |
381 | |
382 | |
383 | /* tar file list or extract */ |
384 | |
385 | int tar (gzFile in,int action,int arg,int argc,char **argv) |
386 | { |
387 | union tar_buffer buffer; |
388 | int len; |
389 | int err; |
390 | int getheader = 1; |
391 | int remaining = 0; |
392 | FILE *outfile = NULL; |
393 | char fname[BLOCKSIZE]; |
394 | int tarmode; |
395 | time_t tartime; |
396 | struct attr_item *attributes = NULL; |
397 | |
398 | if (action == TGZ_LIST) |
399 | printf(" date time size file\n" |
400 | " ---------- -------- --------- -------------------------------------\n"); |
401 | while (1) |
402 | { |
403 | len = gzread(in, &buffer, BLOCKSIZE); |
404 | if (len < 0) |
405 | error(gzerror(in, &err)); |
406 | /* |
407 | * Always expect complete blocks to process |
408 | * the tar information. |
409 | */ |
410 | if (len != BLOCKSIZE) |
411 | { |
412 | action = TGZ_INVALID; /* force error exit */ |
413 | remaining = 0; /* force I/O cleanup */ |
414 | } |
415 | |
416 | /* |
417 | * If we have to get a tar header |
418 | */ |
419 | if (getheader >= 1) |
420 | { |
421 | /* |
422 | * if we met the end of the tar |
423 | * or the end-of-tar block, |
424 | * we are done |
425 | */ |
426 | if (len == 0 || buffer.header.name[0] == 0) |
427 | break; |
428 | |
429 | tarmode = getoct(buffer.header.mode,8); |
430 | tartime = (time_t)getoct(buffer.header.mtime,12); |
431 | if (tarmode == -1 || tartime == (time_t)-1) |
432 | { |
433 | buffer.header.name[0] = 0; |
434 | action = TGZ_INVALID; |
435 | } |
436 | |
437 | if (getheader == 1) |
438 | { |
439 | strncpy(fname,buffer.header.name,SHORTNAMESIZE); |
440 | if (fname[SHORTNAMESIZE-1] != 0) |
441 | fname[SHORTNAMESIZE] = 0; |
442 | } |
443 | else |
444 | { |
445 | /* |
446 | * The file name is longer than SHORTNAMESIZE |
447 | */ |
448 | if (strncmp(fname,buffer.header.name,SHORTNAMESIZE-1) != 0) |
449 | error("bad long name"); |
450 | getheader = 1; |
451 | } |
452 | |
453 | /* |
454 | * Act according to the type flag |
455 | */ |
456 | switch (buffer.header.typeflag) |
457 | { |
458 | case DIRTYPE: |
459 | if (action == TGZ_LIST) |
460 | printf(" %s <dir> %s\n",strtime(&tartime),fname); |
461 | if (action == TGZ_EXTRACT) |
462 | { |
463 | makedir(fname); |
464 | push_attr(&attributes,fname,tarmode,tartime); |
465 | } |
466 | break; |
467 | case REGTYPE: |
468 | case AREGTYPE: |
469 | remaining = getoct(buffer.header.size,12); |
470 | if (remaining == -1) |
471 | { |
472 | action = TGZ_INVALID; |
473 | break; |
474 | } |
475 | if (action == TGZ_LIST) |
476 | printf(" %s %9d %s\n",strtime(&tartime),remaining,fname); |
477 | else if (action == TGZ_EXTRACT) |
478 | { |
479 | if (matchname(arg,argc,argv,fname)) |
480 | { |
481 | outfile = fopen(fname,"wb"); |
482 | if (outfile == NULL) { |
483 | /* try creating directory */ |
484 | char *p = strrchr(fname, '/'); |
485 | if (p != NULL) { |
486 | *p = '\0'; |
487 | makedir(fname); |
488 | *p = '/'; |
489 | outfile = fopen(fname,"wb"); |
490 | } |
491 | } |
492 | if (outfile != NULL) |
493 | printf("Extracting %s\n",fname); |
494 | else |
495 | fprintf(stderr, "%s: Couldn't create %s",prog,fname); |
496 | } |
497 | else |
498 | outfile = NULL; |
499 | } |
500 | getheader = 0; |
501 | break; |
502 | case GNUTYPE_LONGLINK: |
503 | case GNUTYPE_LONGNAME: |
504 | remaining = getoct(buffer.header.size,12); |
505 | if (remaining < 0 || remaining >= BLOCKSIZE) |
506 | { |
507 | action = TGZ_INVALID; |
508 | break; |
509 | } |
510 | len = gzread(in, fname, BLOCKSIZE); |
511 | if (len < 0) |
512 | error(gzerror(in, &err)); |
513 | if (fname[BLOCKSIZE-1] != 0 || (int)strlen(fname) > remaining) |
514 | { |
515 | action = TGZ_INVALID; |
516 | break; |
517 | } |
518 | getheader = 2; |
519 | break; |
520 | default: |
521 | if (action == TGZ_LIST) |
522 | printf(" %s <---> %s\n",strtime(&tartime),fname); |
523 | break; |
524 | } |
525 | } |
526 | else |
527 | { |
528 | unsigned int bytes = (remaining > BLOCKSIZE) ? BLOCKSIZE : remaining; |
529 | |
530 | if (outfile != NULL) |
531 | { |
532 | if (fwrite(&buffer,sizeof(char),bytes,outfile) != bytes) |
533 | { |
534 | fprintf(stderr, |
535 | "%s: Error writing %s -- skipping\n",prog,fname); |
536 | fclose(outfile); |
537 | outfile = NULL; |
538 | remove(fname); |
539 | } |
540 | } |
541 | remaining -= bytes; |
542 | } |
543 | |
544 | if (remaining == 0) |
545 | { |
546 | getheader = 1; |
547 | if (outfile != NULL) |
548 | { |
549 | fclose(outfile); |
550 | outfile = NULL; |
551 | if (action != TGZ_INVALID) |
552 | push_attr(&attributes,fname,tarmode,tartime); |
553 | } |
554 | } |
555 | |
556 | /* |
557 | * Abandon if errors are found |
558 | */ |
559 | if (action == TGZ_INVALID) |
560 | { |
561 | error("broken archive"); |
562 | break; |
563 | } |
564 | } |
565 | |
566 | /* |
567 | * Restore file modes and time stamps |
568 | */ |
569 | restore_attr(&attributes); |
570 | |
571 | if (gzclose(in) != Z_OK) |
572 | error("failed gzclose"); |
573 | |
574 | return 0; |
575 | } |
576 | |
577 | |
578 | /* ============================================================ */ |
579 | |
580 | void help(int exitval) |
581 | { |
582 | printf("untgz version 0.2.1\n" |
583 | " using zlib version %s\n\n", |
584 | zlibVersion()); |
585 | printf("Usage: untgz file.tgz extract all files\n" |
586 | " untgz file.tgz fname ... extract selected files\n" |
587 | " untgz -l file.tgz list archive contents\n" |
588 | " untgz -h display this help\n"); |
589 | exit(exitval); |
590 | } |
591 | |
9e052883 |
592 | |
593 | /* ============================================================ */ |
594 | |
595 | #if defined(WIN32) && defined(__GNUC__) |
596 | int _CRT_glob = 0; /* disable argument globbing in MinGW */ |
597 | #endif |
598 | |
599 | int main(int argc,char **argv) |
600 | { |
601 | int action = TGZ_EXTRACT; |
602 | int arg = 1; |
603 | char *TGZfile; |
648db22b |
604 | gzFile f; |
9e052883 |
605 | |
606 | prog = strrchr(argv[0],'\\'); |
607 | if (prog == NULL) |
608 | { |
609 | prog = strrchr(argv[0],'/'); |
610 | if (prog == NULL) |
611 | { |
612 | prog = strrchr(argv[0],':'); |
613 | if (prog == NULL) |
614 | prog = argv[0]; |
615 | else |
616 | prog++; |
617 | } |
618 | else |
619 | prog++; |
620 | } |
621 | else |
622 | prog++; |
623 | |
624 | if (argc == 1) |
625 | help(0); |
626 | |
627 | if (strcmp(argv[arg],"-l") == 0) |
628 | { |
629 | action = TGZ_LIST; |
630 | if (argc == ++arg) |
631 | help(0); |
632 | } |
633 | else if (strcmp(argv[arg],"-h") == 0) |
634 | { |
635 | help(0); |
636 | } |
637 | |
638 | if ((TGZfile = TGZfname(argv[arg])) == NULL) |
639 | TGZnotfound(argv[arg]); |
640 | |
641 | ++arg; |
642 | if ((action == TGZ_LIST) && (arg != argc)) |
643 | help(1); |
644 | |
645 | /* |
646 | * Process the TGZ file |
647 | */ |
648 | switch(action) |
649 | { |
650 | case TGZ_LIST: |
651 | case TGZ_EXTRACT: |
652 | f = gzopen(TGZfile,"rb"); |
653 | if (f == NULL) |
654 | { |
655 | fprintf(stderr,"%s: Couldn't gzopen %s\n",prog,TGZfile); |
656 | return 1; |
657 | } |
658 | exit(tar(f, action, arg, argc, argv)); |
659 | break; |
660 | |
661 | default: |
662 | error("Unknown option"); |
663 | exit(1); |
664 | } |
665 | |
666 | return 0; |
667 | } |