xref: /illumos-gate/usr/src/cmd/compress/compress.c (revision bc54f855)
1 /*
2  * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
3  * Use is subject to license terms.
4  */
5 
6 /*	Copyright (c) 1984, 1986, 1987, 1988, 1989 AT&T	*/
7 /*	  All Rights Reserved  	*/
8 
9 
10 /*
11  * Copyright (c) 1986 Regents of the University of California.
12  * All rights reserved.  The Berkeley software License Agreement
13  * specifies the terms and conditions for redistribution.
14  */
15 
16 /*
17  * Copyright (c) 2018, Joyent, Inc.
18  */
19 
20 /*
21  * Compress - data compression program
22  */
23 #define	min(a, b)	((a > b) ? b : a)
24 
25 /*
26  * machine variants which require cc -Dmachine:  pdp11, z8000, pcxt
27  */
28 
29 /*
30  * Set USERMEM to the maximum amount of physical user memory available
31  * in bytes.  USERMEM is used to determine the maximum BITS that can be used
32  * for compression.
33  *
34  * SACREDMEM is the amount of physical memory saved for others; compress
35  * will hog the rest.
36  */
37 #ifndef SACREDMEM
38 #define	SACREDMEM	0
39 #endif
40 
41 #ifndef USERMEM
42 #define	USERMEM 	450000	/* default user memory */
43 #endif
44 
45 #ifdef USERMEM
46 #if USERMEM >= (433484+SACREDMEM)
47 #define	PBITS	16
48 #else
49 #if USERMEM >= (229600+SACREDMEM)
50 #define	PBITS	15
51 #else
52 #if USERMEM >= (127536+SACREDMEM)
53 #define	PBITS	14
54 #else
55 #if USERMEM >= (73464+SACREDMEM)
56 #define	PBITS	13
57 #else
58 #define	PBITS	12
59 #endif
60 #endif
61 #endif
62 #endif
63 #undef USERMEM
64 #endif /* USERMEM */
65 
66 #ifdef PBITS		/* Preferred BITS for this memory size */
67 #ifndef BITS
68 #define	BITS PBITS
69 #endif /* BITS */
70 #endif /* PBITS */
71 
72 #if BITS == 16
73 #define	HSIZE	69001		/* 95% occupancy */
74 #endif
75 #if BITS == 15
76 #define	HSIZE	35023		/* 94% occupancy */
77 #endif
78 #if BITS == 14
79 #define	HSIZE	18013		/* 91% occupancy */
80 #endif
81 #if BITS == 13
82 #define	HSIZE	9001		/* 91% occupancy */
83 #endif
84 #if BITS <= 12
85 #define	HSIZE	5003		/* 80% occupancy */
86 #endif
87 
88 #define	OUTSTACKSIZE	(2<<BITS)
89 
90 /*
91  * a code_int must be able to hold 2**BITS values of type int, and also -1
92  */
93 #if BITS > 15
94 typedef long int	code_int;
95 #else
96 typedef int		code_int;
97 #endif
98 
99 typedef long int	count_int;
100 typedef long long	count_long;
101 
102 typedef	unsigned char	char_type;
103 
104 static char_type magic_header[] = { "\037\235" }; /* 1F 9D */
105 
106 /* Defines for third byte of header */
107 #define	BIT_MASK	0x1f
108 #define	BLOCK_MASK	0x80
109 /*
110  * Masks 0x40 and 0x20 are free.  I think 0x20 should mean that there is
111  * a fourth header byte(for expansion).
112  */
113 #define	INIT_BITS 9			/* initial number of bits/code */
114 
115 /*
116  * compress.c - File compression ala IEEE Computer, June 1984.
117  */
118 static char rcs_ident[] =
119 	"$Header: compress.c,v 4.0 85/07/30 12:50:00 joe Release $";
120 
121 #include <ctype.h>
122 #include <signal.h>
123 #include <sys/param.h>
124 #include <locale.h>
125 #include <langinfo.h>
126 #include <sys/acl.h>
127 #include <utime.h>
128 #include <libgen.h>
129 #include <setjmp.h>
130 #include <aclutils.h>
131 #include <libcmdutils.h>
132 #include "getresponse.h"
133 
134 
135 static int n_bits;			/* number of bits/code */
136 static int maxbits = BITS;	/* user settable max # bits/code */
137 static code_int maxcode;	/* maximum code, given n_bits */
138 			/* should NEVER generate this code */
139 static code_int maxmaxcode = 1 << BITS;
140 #define	MAXCODE(n_bits)	((1 << (n_bits)) - 1)
141 
142 static count_int htab [OUTSTACKSIZE];
143 static unsigned short codetab [OUTSTACKSIZE];
144 
145 #define	htabof(i)	htab[i]
146 #define	codetabof(i)	codetab[i]
147 static code_int hsize = HSIZE; /* for dynamic table sizing */
148 static off_t	fsize;	/* file size of input file */
149 
150 /*
151  * To save much memory, we overlay the table used by compress() with those
152  * used by decompress().  The tab_prefix table is the same size and type
153  * as the codetab.  The tab_suffix table needs 2**BITS characters.  We
154  * get this from the beginning of htab.  The output stack uses the rest
155  * of htab, and contains characters.  There is plenty of room for any
156  * possible stack (stack used to be 8000 characters).
157  */
158 
159 #define	tab_prefixof(i)		codetabof(i)
160 #define	tab_suffixof(i)		((char_type *)(htab))[i]
161 #define	de_stack		((char_type *)&tab_suffixof(1<<BITS))
162 #define	stack_max		((char_type *)&tab_suffixof(OUTSTACKSIZE))
163 
164 static code_int free_ent = 0; /* first unused entry */
165 static int newline_needed = 0;
166 static int didnt_shrink = 0;
167 static int perm_stat = 0;	/* permanent status */
168 
169 static code_int getcode();
170 
171 	/* Use a 3-byte magic number header, unless old file */
172 static int nomagic = 0;
173 	/* Write output on stdout, suppress messages */
174 static int zcat_flg = 0;	/* use stdout on all files */
175 static int zcat_cmd = 0;	/* zcat cmd */
176 static int use_stdout = 0;	/* set for each file processed */
177 	/* Don't unlink output file on interrupt */
178 static int precious = 1;
179 static int quiet = 1;	/* don't tell me about compression */
180 
181 /*
182  * block compression parameters -- after all codes are used up,
183  * and compression rate changes, start over.
184  */
185 static int block_compress = BLOCK_MASK;
186 static int clear_flg = 0;
187 static long int ratio = 0;
188 #define	CHECK_GAP 10000	/* ratio check interval */
189 static count_long checkpoint = CHECK_GAP;
190 /*
191  * the next two codes should not be changed lightly, as they must not
192  * lie within the contiguous general code space.
193  */
194 #define	FIRST	257	/* first free entry */
195 #define	CLEAR	256	/* table clear output code */
196 
197 static int force = 0;
198 static char ofname [MAXPATHLEN];
199 
200 static int Vflg = 0;
201 static int vflg = 0;
202 static int qflg = 0;
203 static int bflg = 0;
204 static int Fflg = 0;
205 static int dflg = 0;
206 static int cflg = 0;
207 static int Cflg = 0;
208 
209 #ifdef DEBUG
210 int verbose = 0;
211 int debug = 0;
212 #endif /* DEBUG */
213 
214 static void (*oldint)();
215 static int bgnd_flag;
216 
217 static int do_decomp = 0;
218 
219 static char *progname;
220 static char *optstr;
221 /*
222  * Fix lint errors
223  */
224 
225 static char *local_basename(char *);
226 
227 static int  addDotZ(char *, size_t);
228 
229 static void Usage(void);
230 static void cl_block(count_long);
231 static void cl_hash(count_int);
232 static void compress(void);
233 static void copystat(char *, struct stat *, char *);
234 static void decompress(void);
235 static void ioerror(void);
236 static void onintr();
237 static void oops();
238 static void output(code_int);
239 static void prratio(FILE *, count_long, count_long);
240 static void version(void);
241 
242 #ifdef DEBUG
243 static int in_stack(int, int);
244 static void dump_tab(void);
245 static void printcodes(void);
246 #endif
247 
248 /* For error-handling */
249 
250 static jmp_buf env;
251 
252 /* For input and ouput */
253 
254 static FILE *inp;		/* the current input file */
255 static FILE *infile;		/* disk-based input stream */
256 static FILE *outp;		/* current output file */
257 static FILE *outfile;		/* disk-based output stream */
258 
259 /* For output() */
260 
261 static char buf[BITS];
262 
263 static char_type lmask[9] =
264 	{0xff, 0xfe, 0xfc, 0xf8, 0xf0, 0xe0, 0xc0, 0x80, 0x00};
265 static char_type rmask[9] =
266 	{0x00, 0x01, 0x03, 0x07, 0x0f, 0x1f, 0x3f, 0x7f, 0xff};
267 
268 /* For compress () */
269 
270 static int offset;
271 static count_long bytes_out;	/* length of compressed output */
272 	/* # of codes output (for debugging) */
273 
274 /* For dump_tab() */
275 
276 #define	STACK_SIZE	15000
277 #ifdef DEBUG
278 code_int sorttab[1<<BITS];	/* sorted pointers into htab */
279 #endif
280 
281 /* Extended system attribute support */
282 
283 static int saflg = 0;
284 
285 /*
286  * *************************************************************
287  * TAG( main )
288  *
289  * Algorithm from "A Technique for High Performance Data Compression",
290  * Terry A. Welch, IEEE Computer Vol 17, No 6 (June 1984), pp 8-19.
291  *
292  * Usage: compress [-dfvc/] [-b bits] [file ...]
293  * Inputs:
294  *	-d:	    If given, decompression is done instead.
295  *
296  *	-c:	    Write output on stdout, don't remove original.
297  *
298  *	-b:	    Parameter limits the max number of bits/code.
299  *
300  *	-f:	    Forces output file to be generated, even if one already
301  *		    exists, and even if no space is saved by compressing.
302  *		    If -f is not used, the user will be prompted if stdin is
303  *		    a tty, otherwise, the output file will not be overwritten.
304  *
305  *	-/	    Copies extended attributes and extended system attributes.
306  *
307  *  -v:	    Write compression statistics
308  *
309  * 	file ...:   Files to be compressed.  If none specified, stdin
310  *		    is used.
311  * Outputs:
312  *	file.Z:	    Compressed form of file with same mode, owner, and utimes
313  * 	or stdout   (if stdin used as input)
314  *
315  * Assumptions:
316  * When filenames are given, replaces with the compressed version
317  * (.Z suffix) only if the file decreases in size.
318  * Algorithm:
319  * Modified Lempel-Ziv method (LZW).  Basically finds common
320  * substrings and replaces them with a variable size code.  This is
321  * deterministic, and can be done on the fly.  Thus, the decompression
322  * procedure needs no input table, but tracks the way the table was built.
323  */
324 
325 int
main(int argc,char * argv[])326 main(int argc, char *argv[])
327 {
328 	int overwrite = 0;	/* Do not overwrite unless given -f flag */
329 	char tempname[MAXPATHLEN];
330 	char line[LINE_MAX];
331 	char **filelist, **fileptr;
332 	char *cp;
333 	struct stat statbuf;
334 	struct stat ostatbuf;
335 	int ch;				/* XCU4 */
336 	char	*p;
337 	extern int optind, optopt;
338 	extern char *optarg;
339 	int dash_count = 0;		/* times "-" is on cmdline */
340 
341 	/* XCU4 changes */
342 	(void) setlocale(LC_ALL, "");
343 #if !defined(TEXT_DOMAIN)	/* Should be defined by cc -D */
344 #define	TEXT_DOMAIN "SYS_TEST"	/* Use this only if it weren't */
345 #endif
346 	(void) textdomain(TEXT_DOMAIN);
347 
348 	if (init_yes() < 0) {
349 		(void) fprintf(stderr, gettext(ERR_MSG_INIT_YES),
350 		    strerror(errno));
351 		exit(1);
352 	}
353 
354 	/* This bg check only works for sh. */
355 	if ((oldint = signal(SIGINT, SIG_IGN)) != SIG_IGN) {
356 		(void) signal(SIGINT, onintr);
357 		(void) signal(SIGSEGV, oops);
358 	}
359 	bgnd_flag = oldint != SIG_DFL;
360 
361 	/* Allocate room for argv + "-" (if stdin needs to be added) */
362 
363 	filelist = fileptr = (char **)(malloc((argc + 1) * sizeof (*argv)));
364 	*filelist = NULL;
365 
366 	if ((cp = rindex(argv[0], '/')) != 0) {
367 		cp++;
368 	} else {
369 		cp = argv[0];
370 	}
371 
372 	if (strcmp(cp, "uncompress") == 0) {
373 		do_decomp = 1;
374 	} else if (strcmp(cp, "zcat") == 0) {
375 		do_decomp = 1;
376 		zcat_cmd = zcat_flg = 1;
377 	}
378 
379 	progname = local_basename(argv[0]);
380 
381 	/*
382 	 * Argument Processing
383 	 * All flags are optional.
384 	 * -D = > debug
385 	 * -V = > print Version; debug verbose
386 	 * -d = > do_decomp
387 	 * -v = > unquiet
388 	 * -f = > force overwrite of output file
389 	 * -n = > no header: useful to uncompress old files
390 	 * -b	  maxbits => maxbits.  If -b is specified,
391 	 *	  then maxbits MUST be given also.
392 	 * -c = > cat all output to stdout
393 	 * -C = > generate output compatible with compress 2.0.
394 	 * if a string is left, must be an input filename.
395 	 */
396 #ifdef DEBUG
397 	optstr = "b:cCdDfFnqvV/";
398 #else
399 	optstr = "b:cCdfFnqvV/";
400 #endif
401 
402 	while ((ch = getopt(argc, argv, optstr)) != EOF) {
403 		/* Process all flags in this arg */
404 		switch (ch) {
405 #ifdef DEBUG
406 			case 'D':
407 				debug = 1;
408 				break;
409 			case 'V':
410 				verbose = 1;
411 				version();
412 				break;
413 #else
414 			case 'V':
415 				version();
416 				Vflg++;
417 				break;
418 #endif /* DEBUG */
419 			case 'v':
420 				quiet = 0;
421 				vflg++;
422 				break;
423 			case 'd':
424 				do_decomp = 1;
425 				dflg++;
426 				break;
427 			case 'f':
428 			case 'F':
429 				Fflg++;
430 				overwrite = 1;
431 				force = 1;
432 				break;
433 			case 'n':
434 				nomagic = 1;
435 				break;
436 			case 'C':
437 				Cflg++;
438 				block_compress = 0;
439 				break;
440 			case 'b':
441 				bflg++;
442 				p = optarg;
443 				if (!p) {
444 					(void) fprintf(stderr, gettext(
445 					    "Missing maxbits\n"));
446 					Usage();
447 					exit(1);
448 				}
449 				maxbits = strtoul(optarg, &p, 10);
450 				if (*p) {
451 					(void) fprintf(stderr, gettext(
452 					    "Missing maxbits\n"));
453 					Usage();
454 					exit(1);
455 				}
456 				break;
457 
458 			case 'c':
459 				cflg++;
460 				zcat_flg = 1;
461 				break;
462 			case 'q':
463 				qflg++;
464 				quiet = 1;
465 				break;
466 			case '/':
467 				saflg++;
468 				break;
469 			default:
470 				(void) fprintf(stderr, gettext(
471 				    "Unknown flag: '%c'\n"), optopt);
472 				Usage();
473 				exit(1);
474 		}
475 	} /* while */
476 
477 	/*
478 	 * Validate zcat syntax
479 	 */
480 
481 	if (zcat_cmd && (Fflg | Cflg | cflg |
482 	    bflg | qflg | dflg | nomagic)) {
483 		(void) fprintf(stderr, gettext(
484 		    "Invalid Option\n"));
485 		Usage();
486 		exit(1);
487 	}
488 
489 	/*
490 	 * Process the file list
491 	 */
492 
493 	for (; optind < argc; optind++) {
494 		if (strcmp(argv[optind], "-") == 0) {
495 			dash_count++;
496 		}
497 
498 		*fileptr++ = argv[optind];	/* Build input file list */
499 		*fileptr = NULL;
500 	}
501 
502 	if (dash_count > 1) {
503 		(void) fprintf(stderr,
504 		    gettext("%s may only appear once in the file"
505 		    " list\n"), "\"-\"");
506 		exit(1);
507 	}
508 
509 	if (fileptr - filelist == 0) {
510 		*fileptr++ = "-";
511 		*fileptr = NULL;
512 	}
513 
514 	if (fileptr - filelist > 1 && cflg && !do_decomp) {
515 		(void) fprintf(stderr,
516 		    gettext("compress: only one file may be compressed"
517 		    " to stdout\n"));
518 		exit(1);
519 	}
520 
521 	if (maxbits < INIT_BITS)
522 		maxbits = INIT_BITS;
523 	if (maxbits > BITS)
524 		maxbits = BITS;
525 	maxmaxcode = 1 << maxbits;
526 
527 	/* Need to open something to close with freopen later */
528 
529 	if ((infile = fopen("/dev/null", "r")) == NULL) {
530 		(void) fprintf(stderr, gettext("Error opening /dev/null for "
531 		    "input\n"));
532 		exit(1);
533 	}
534 
535 	if ((outfile = fopen("/dev/null", "w")) == NULL) {
536 		(void) fprintf(stderr, gettext("Error opening /dev/null for "
537 		    "output\n"));
538 		exit(1);
539 	}
540 
541 	for (fileptr = filelist; *fileptr; fileptr++) {
542 		int jmpval = 0;
543 		didnt_shrink = 0;
544 		newline_needed = 0;
545 
546 		if (do_decomp) {
547 			/* DECOMPRESSION */
548 
549 			if (strcmp(*fileptr, "-") == 0) {
550 				/* process stdin */
551 				inp = stdin;
552 				outp = stdout;
553 				use_stdout = 1;
554 				*fileptr = "stdin"; /* for error messages */
555 			} else {
556 				/* process the named file */
557 
558 				inp = infile;
559 				outp = outfile;
560 				use_stdout = 0;
561 
562 				if (zcat_flg) {
563 					use_stdout = 1;
564 					outp = stdout;
565 				}
566 
567 				/* Check for .Z suffix */
568 
569 				if (strcmp(*fileptr +
570 				    strlen(*fileptr) - 2, ".Z") != 0) {
571 					/* No .Z: tack one on */
572 
573 					if (strlcpy(tempname, *fileptr,
574 					    sizeof (tempname)) >=
575 					    sizeof (tempname)) {
576 						(void) fprintf(stderr,
577 						    gettext("%s: filename "
578 						    "too long\n"),
579 						    *fileptr);
580 						perm_stat = 1;
581 						continue;
582 					}
583 
584 					if (addDotZ(tempname,
585 					    sizeof (tempname)) < 0) {
586 						perm_stat = 1;
587 						continue;
588 					}
589 
590 					*fileptr = tempname;
591 				}
592 
593 				/* Open input file */
594 
595 				if (stat(*fileptr, &statbuf) < 0) {
596 					perror(*fileptr);
597 					perm_stat = 1;
598 					continue;
599 				}
600 
601 				if ((freopen(*fileptr, "r", inp)) == NULL) {
602 					perror(*fileptr);
603 					perm_stat = 1;
604 					continue;
605 				}
606 			}
607 
608 			/* Check the magic number */
609 
610 			if (nomagic == 0) {
611 				if ((getc(inp) !=
612 				    (magic_header[0] & 0xFF)) ||
613 				    (getc(inp) !=
614 				    (magic_header[1] & 0xFF))) {
615 					(void) fprintf(stderr, gettext(
616 					    "%s: not in compressed "
617 					    "format\n"),
618 					    *fileptr);
619 					perm_stat = 1;
620 					continue;
621 				}
622 
623 				/* set -b from file */
624 				if ((maxbits = getc(inp)) == EOF &&
625 				    ferror(inp)) {
626 					perror(*fileptr);
627 					perm_stat = 1;
628 					continue;
629 				}
630 
631 				block_compress = maxbits & BLOCK_MASK;
632 				maxbits &= BIT_MASK;
633 				maxmaxcode = 1 << maxbits;
634 
635 				if (maxbits > BITS) {
636 					(void) fprintf(stderr,
637 					    gettext("%s: compressed "
638 					    "with %d bits, "
639 					    "can only handle"
640 					    " %d bits\n"),
641 					    *fileptr, maxbits, BITS);
642 					perm_stat = 1;
643 					continue;
644 				}
645 			}
646 
647 			if (!use_stdout) {
648 				/* Generate output filename */
649 
650 				if (strlcpy(ofname, *fileptr,
651 				    sizeof (ofname)) >=
652 				    sizeof (ofname)) {
653 					(void) fprintf(stderr,
654 					    gettext("%s: filename "
655 					    "too long\n"),
656 					    *fileptr);
657 					perm_stat = 1;
658 					continue;
659 				}
660 
661 				/* Strip off .Z */
662 
663 				ofname[strlen(*fileptr) - 2] = '\0';
664 			}
665 		} else {
666 			/* COMPRESSION */
667 
668 			if (strcmp(*fileptr, "-") == 0) {
669 				/* process stdin */
670 				inp = stdin;
671 				outp = stdout;
672 				use_stdout = 1;
673 				*fileptr = "stdin"; /* for error messages */
674 
675 				/* Use the largest possible hash table */
676 				hsize =  HSIZE;
677 			} else {
678 				/* process the named file */
679 
680 				inp = infile;
681 				outp = outfile;
682 				use_stdout = 0;
683 
684 				if (zcat_flg) {
685 					use_stdout = 1;
686 					outp = stdout;
687 				}
688 
689 				if (strcmp(*fileptr +
690 				    strlen(*fileptr) - 2, ".Z") == 0) {
691 					(void) fprintf(stderr, gettext(
692 					    "%s: already has .Z "
693 					    "suffix -- no change\n"),
694 					    *fileptr);
695 					perm_stat = 1;
696 					continue;
697 				}
698 				/* Open input file */
699 
700 				if (stat(*fileptr, &statbuf) < 0) {
701 					perror(*fileptr);
702 					perm_stat = 1;
703 					continue;
704 				}
705 
706 				if ((freopen(*fileptr, "r", inp)) == NULL) {
707 					perror(*fileptr);
708 					perm_stat = 1;
709 					continue;
710 				}
711 
712 				fsize = (off_t)statbuf.st_size;
713 
714 				/*
715 				 * tune hash table size for small
716 				 * files -- ad hoc,
717 				 * but the sizes match earlier #defines, which
718 				 * serve as upper bounds on the number of
719 				 * output codes.
720 				 */
721 				hsize = HSIZE;
722 				if (fsize < (1 << 12))
723 					hsize = min(5003, HSIZE);
724 				else if (fsize < (1 << 13))
725 					hsize = min(9001, HSIZE);
726 				else if (fsize < (1 << 14))
727 					hsize = min(18013, HSIZE);
728 				else if (fsize < (1 << 15))
729 					hsize = min(35023, HSIZE);
730 				else if (fsize < 47000)
731 					hsize = min(50021, HSIZE);
732 
733 				if (!use_stdout) {
734 					/* Generate output filename */
735 
736 					if (strlcpy(ofname, *fileptr,
737 					    sizeof (ofname)) >=
738 					    sizeof (ofname)) {
739 						(void) fprintf(stderr,
740 						    gettext("%s: filename "
741 						    "too long\n"),
742 						    *fileptr);
743 						perm_stat = 1;
744 						continue;
745 					}
746 
747 					if (addDotZ(ofname,
748 					    sizeof (ofname)) < 0) {
749 						perm_stat = 1;
750 						continue;
751 					}
752 				}
753 			}
754 		}	/* if (do_decomp) */
755 
756 		/* Check for overwrite of existing file */
757 
758 		if (!overwrite && !use_stdout) {
759 			if (stat(ofname, &ostatbuf) == 0) {
760 				(void) fprintf(stderr, gettext(
761 				    "%s already exists;"), ofname);
762 				if (bgnd_flag == 0 && isatty(2)) {
763 					int cin;
764 
765 					(void) fprintf(stderr, gettext(
766 					    " do you wish to overwr"
767 					    "ite %s (%s or %s)? "),
768 					    ofname, yesstr, nostr);
769 					(void) fflush(stderr);
770 					for (cin = 0; cin < LINE_MAX; cin++)
771 						line[cin] = 0;
772 					(void) read(2, line, LINE_MAX);
773 
774 					if (yes_check(line) == 0) {
775 						(void) fprintf(stderr,
776 						    gettext(
777 						    "\tnot overwri"
778 						    "tten\n"));
779 						continue;
780 					}
781 				} else {
782 					/*
783 					 * XPG4: Assertion 1009
784 					 * Standard input is not
785 					 * terminal, and no '-f',
786 					 * and file exists.
787 					 */
788 
789 					(void) fprintf(stderr, gettext(
790 					    "%s: File exists, -f not"
791 					    " specified, and ru"
792 					    "nning in the backgro"
793 					    "und.\n"), *fileptr);
794 					perm_stat = 1;
795 					continue;
796 				}
797 			}
798 		}
799 		if (!use_stdout) {
800 			if ((pathconf(ofname, _PC_XATTR_EXISTS) == 1) ||
801 			    (saflg && sysattr_support(ofname,
802 			    _PC_SATTR_EXISTS) == 1)) {
803 				(void) unlink(ofname);
804 			}
805 			/* Open output file */
806 			if (freopen(ofname, "w", outp) == NULL) {
807 				perror(ofname);
808 				perm_stat = 1;
809 				continue;
810 			}
811 			precious = 0;
812 			if (!quiet) {
813 				(void) fprintf(stderr, "%s: ",
814 				    *fileptr);
815 				newline_needed = 1;
816 			}
817 		} else if (!quiet && !do_decomp) {
818 			(void) fprintf(stderr, "%s: ",
819 			    *fileptr);
820 			newline_needed = 1;
821 		}
822 
823 		/* Actually do the compression/decompression */
824 
825 		if ((jmpval = setjmp(env)) == 0) {
826 			/* We'll see how things go */
827 #ifndef DEBUG
828 			if (do_decomp == 0)  {
829 				compress();
830 			} else {
831 				decompress();
832 			}
833 #else
834 			if (do_decomp == 0)  {
835 				compress();
836 			} else if (debug == 0)  {
837 				decompress();
838 			} else {
839 				printcodes();
840 			}
841 
842 			if (verbose) {
843 				dump_tab();
844 			}
845 #endif
846 		} else {
847 			/*
848 			 * Things went badly - clean up and go on.
849 			 * jmpval's values break down as follows:
850 			 *   1 == message determined by ferror() values.
851 			 *   2 == input problem message needed.
852 			 *   3 == output problem message needed.
853 			 */
854 
855 			if (ferror(inp) || jmpval == 2) {
856 				if (do_decomp) {
857 					(void) fprintf(stderr, gettext(
858 					    "uncompress: %s: corrupt"
859 					    " input\n"), *fileptr);
860 				} else {
861 					perror(*fileptr);
862 				}
863 			}
864 
865 			if (ferror(outp) || jmpval == 3) {
866 				/* handle output errors */
867 
868 				if (use_stdout) {
869 					perror("");
870 				} else {
871 					perror(ofname);
872 				}
873 			}
874 
875 			if (ofname[0] != '\0') {
876 				if (unlink(ofname) < 0)  {
877 					perror(ofname);
878 				}
879 
880 				ofname[0] = '\0';
881 			}
882 
883 			perm_stat = 1;
884 			continue;
885 		}
886 
887 		/* Things went well */
888 
889 		if (!use_stdout) {
890 				/* Copy stats */
891 			copystat(*fileptr, &statbuf, ofname);
892 			precious = 1;
893 			if (newline_needed) {
894 				(void) putc('\n', stderr);
895 			}
896 			/*
897 			 * Print the info. for unchanged file
898 			 * when no -v
899 			 */
900 
901 			if (didnt_shrink) {
902 				if (!force && perm_stat == 0) {
903 					if (quiet) {
904 						(void) fprintf(stderr, gettext(
905 						    "%s: -- file "
906 						    "unchanged\n"),
907 						    *fileptr);
908 					}
909 
910 					perm_stat = 2;
911 				}
912 			}
913 		} else {
914 			if (didnt_shrink && !force && perm_stat == 0) {
915 				perm_stat = 2;
916 			}
917 
918 			if (newline_needed) {
919 				(void) fprintf(stderr, "\n");
920 			}
921 		}
922 	}	/* for */
923 
924 	return (perm_stat);
925 }
926 
927 static void
cinterr(int hshift)928 cinterr(int hshift)
929 {
930 	/* we have exceeded the hash table */
931 	(void) fprintf(stderr,
932 	    "internal error: hashtable exceeded - hsize = %ld\n", hsize);
933 	(void) fprintf(stderr, "hshift = %d, %d\n", hshift, (1 << hshift) -1);
934 	(void) fprintf(stderr, "maxbits = %d\n", maxbits);
935 	(void) fprintf(stderr, "n_bits = %d\n", n_bits);
936 	(void) fprintf(stderr, "maxcode = %ld\n", maxcode);
937 	longjmp(env, 1);
938 }
939 
940 static code_int
adjusti(code_int i,code_int hsize_reg)941 adjusti(code_int i, code_int hsize_reg)
942 {
943 	while (i < 0) {
944 		i += hsize_reg;
945 	}
946 
947 	while (i >= hsize_reg) {
948 		i -= hsize_reg;
949 	}
950 	return (i);
951 }
952 
953 /*
954  * compress inp to outp
955  *
956  * Algorithm:  use open addressing double hashing(no chaining) on the
957  * prefix code / next character combination.  We do a variant of Knuth's
958  * algorithm D (vol. 3, sec. 6.4) along with G. Knott's relatively-prime
959  * secondary probe.  Here, the modular division first probe is gives way
960  * to a faster exclusive-or manipulation.  Also do block compression with
961  * an adaptive reset, whereby the code table is cleared when the compression
962  * ratio decreases, but after the table fills.  The variable-length output
963  * codes are re-sized at this point, and a special CLEAR code is generated
964  * for the decompressor.  Late addition:  construct the table according to
965  * file size for noticeable speed improvement on small files.  Please direct
966  * questions about this implementation to ames!jaw.
967  */
968 
969 static void
compress()970 compress()
971 {
972 	long fcode;
973 	code_int i = 0;
974 	int c;
975 	code_int ent;
976 	int disp;
977 	code_int hsize_reg;
978 	int hshift;
979 	int probecnt;
980 	count_long in_count;
981 	uint32_t inchi, inclo;
982 	int maxbits_reg;
983 	FILE *fin = inp;
984 #ifdef DEBUG
985 	count_long out_count = 0;
986 #endif
987 
988 	if (nomagic == 0) {
989 		if ((putc(magic_header[0], outp) == EOF ||
990 		    putc(magic_header[1], outp) == EOF ||
991 		    putc((char)(maxbits | block_compress),
992 		    outp) == EOF) &&
993 		    ferror(outp)) {
994 			ioerror();
995 		}
996 	}
997 
998 	offset = 0;
999 	bytes_out = 3;		/* includes 3-byte header mojo */
1000 	clear_flg = 0;
1001 	ratio = 0;
1002 	in_count = 1;
1003 	inchi = 0;
1004 	inclo = 1;
1005 	checkpoint = CHECK_GAP;
1006 	maxcode = MAXCODE(n_bits = INIT_BITS);
1007 	free_ent = ((block_compress) ? FIRST : 256);
1008 
1009 	if ((ent = getc(fin)) == EOF && ferror(fin)) {
1010 		ioerror();
1011 	}
1012 
1013 	hshift = 0;
1014 
1015 	for (fcode = (long)hsize;  fcode < 65536L; fcode *= 2L)
1016 		hshift++;
1017 
1018 	hshift = 8 - hshift;		/* set hash code range bound */
1019 
1020 	hsize_reg = hsize;
1021 	maxbits_reg = maxbits;
1022 
1023 	cl_hash((count_int) hsize_reg);		/* clear hash table */
1024 
1025 	while ((c = getc(fin)) != EOF) {
1026 		if (++inclo == 0)
1027 			inchi++;
1028 		fcode = (long)(((long)c << maxbits_reg) + ent);
1029 		i = ((c << hshift) ^ ent);	/* xor hashing */
1030 
1031 		if ((unsigned int)i >= hsize_reg)
1032 			i = adjusti(i, hsize_reg);
1033 
1034 		if (htabof(i) == fcode) {
1035 			ent = codetabof(i);
1036 			continue;
1037 		} else if ((long)htabof(i) < 0) {
1038 			/* empty slot */
1039 			goto nomatch;
1040 		}
1041 
1042 		/* secondary hash (after G. Knott) */
1043 		disp = hsize_reg - i;
1044 
1045 		if (i == 0) {
1046 			disp = 1;
1047 		}
1048 
1049 		probecnt = 0;
1050 	probe:
1051 		if (++probecnt > hsize_reg)
1052 			cinterr(hshift);
1053 
1054 		if ((i -= disp) < 0) {
1055 			while (i < 0)
1056 				i += hsize_reg;
1057 		}
1058 
1059 		if (htabof(i) == fcode) {
1060 			ent = codetabof(i);
1061 			continue;
1062 		}
1063 
1064 		if ((long)htabof(i) > 0) {
1065 			goto probe;
1066 		}
1067 	nomatch:
1068 		output((code_int) ent);
1069 #ifdef DEBUG
1070 		out_count++;
1071 #endif
1072 		ent = c;
1073 		if (free_ent < maxmaxcode) {
1074 			codetabof(i) = free_ent++;
1075 			/* code -> hashtable */
1076 			htabof(i) = fcode;
1077 		} else {
1078 			in_count = ((long long)inchi<<32|inclo);
1079 			if ((count_long)in_count >=
1080 			    (count_long)checkpoint && block_compress) {
1081 				cl_block(in_count);
1082 			}
1083 		}
1084 	}
1085 
1086 	in_count = ((long long)inchi<<32|inclo);
1087 
1088 	if (ferror(fin) != 0) {
1089 		ioerror();
1090 	}
1091 
1092 	/*
1093 	 * Put out the final code.
1094 	 */
1095 	output((code_int)ent);
1096 #ifdef DEBUG
1097 	out_count++;
1098 #endif
1099 
1100 	output((code_int)-1);
1101 
1102 	/*
1103 	 * Print out stats on stderr
1104 	 */
1105 	if (!quiet) {
1106 #ifdef DEBUG
1107 		(void) fprintf(stderr,
1108 		    "%lld chars in, %lld codes (%lld bytes) out, "
1109 		    "compression factor: ",
1110 		    (count_long)in_count, (count_long)out_count,
1111 		    (count_long) bytes_out);
1112 		prratio(stderr, (count_long)in_count,
1113 		    (count_long)bytes_out);
1114 		(void) fprintf(stderr, "\n");
1115 		(void) fprintf(stderr, "\tCompression as in compact: ");
1116 		prratio(stderr,
1117 		    (count_long)in_count-(count_long)bytes_out,
1118 		    (count_long)in_count);
1119 		(void) fprintf(stderr, "\n");
1120 		(void) fprintf(stderr,
1121 		    "\tLargest code (of last block) was %d"
1122 		    " (%d bits)\n",
1123 		    free_ent - 1, n_bits);
1124 #else /* !DEBUG */
1125 		(void) fprintf(stderr, gettext("Compression: "));
1126 		prratio(stderr,
1127 		    (count_long)in_count-(count_long)bytes_out,
1128 		    (count_long)in_count);
1129 #endif /* DEBUG */
1130 	}
1131 	/* report if no savings */
1132 	if ((count_long)bytes_out > (count_long)in_count) {
1133 		didnt_shrink = 1;
1134 	}
1135 }
1136 
1137 /*
1138  * **************************************************************
1139  * TAG(output)
1140  *
1141  * Output the given code.
1142  * Inputs:
1143  * 	code:	A n_bits-bit integer.  If == -1, then EOF.  This assumes
1144  *		that n_bits = < (long)wordsize - 1.
1145  * Outputs:
1146  * 	Outputs code to the file.
1147  * Assumptions:
1148  *	Chars are 8 bits long.
1149  * Algorithm:
1150  * 	Maintain a BITS character long buffer(so that 8 codes will
1151  * fit in it exactly).  Use the VAX insv instruction to insert each
1152  * code in turn.  When the buffer fills up empty it and start over.
1153  */
1154 
1155 static void
output(code_int code)1156 output(code_int code)
1157 {
1158 #ifdef DEBUG
1159 	static int col = 0;
1160 #endif /* DEBUG */
1161 
1162 	int r_off = offset, bits = n_bits;
1163 	char *bp = buf;
1164 
1165 #ifdef DEBUG
1166 	if (verbose)
1167 		(void) fprintf(stderr, "%5d%c", code,
1168 		    (col += 6) >= 74 ? (col = 0, '\n') : ' ');
1169 #endif /* DEBUG */
1170 	if (code >= 0) {
1171 		/*
1172 		 * byte/bit numbering on the VAX is simulated
1173 		 * by the following code
1174 		 */
1175 		/*
1176 		 * Get to the first byte.
1177 		 */
1178 		bp += (r_off >> 3);
1179 		r_off &= 7;
1180 		/*
1181 		 * Since code is always >= 8 bits, only need to mask the first
1182 		 * hunk on the left.
1183 		 */
1184 		*bp = (*bp & rmask[r_off]) | (code << r_off) & lmask[r_off];
1185 		bp++;
1186 		bits -= (8 - r_off);
1187 		code >>= 8 - r_off;
1188 		/*
1189 		 * Get any 8 bit parts in the middle (<=1 for up to 16
1190 		 * bits).
1191 		 */
1192 		if (bits >= 8) {
1193 			*bp++ = code;
1194 			code >>= 8;
1195 			bits -= 8;
1196 		}
1197 		/* Last bits. */
1198 		if (bits)
1199 			*bp = code;
1200 		offset += n_bits;
1201 		if (offset == (n_bits << 3)) {
1202 			bp = buf;
1203 			bits = n_bits;
1204 			bytes_out += bits;
1205 			do {
1206 				if (putc(*bp, outp) == EOF &&
1207 				    ferror(outp)) {
1208 					ioerror();
1209 				}
1210 				bp++;
1211 			} while (--bits);
1212 			offset = 0;
1213 		}
1214 
1215 		/*
1216 		 * If the next entry is going to be too big for the code size,
1217 		 * then increase it, if possible.
1218 		 */
1219 		if (free_ent > maxcode || (clear_flg > 0)) {
1220 			/*
1221 			 * Write the whole buffer, because the input
1222 			 * side won't discover the size increase until
1223 			 * after it has read it.
1224 			 */
1225 			if (offset > 0) {
1226 				if (fwrite(buf, 1, n_bits, outp) != n_bits) {
1227 					longjmp(env, 3);
1228 				}
1229 				bytes_out += n_bits;
1230 			}
1231 			offset = 0;
1232 
1233 			if (clear_flg) {
1234 				maxcode = MAXCODE(n_bits = INIT_BITS);
1235 				clear_flg = 0;
1236 			} else {
1237 				n_bits++;
1238 				if (n_bits == maxbits)
1239 					maxcode = maxmaxcode;
1240 				else
1241 					maxcode = MAXCODE(n_bits);
1242 			}
1243 #ifdef DEBUG
1244 			if (debug) {
1245 				(void) fprintf(stderr,
1246 				    "\nChange to %d bits\n", n_bits);
1247 				col = 0;
1248 			}
1249 #endif /* DEBUG */
1250 		}
1251 	} else {
1252 		/*
1253 		 * At EOF, write the rest of the buffer.
1254 		 */
1255 		if (offset > 0) {
1256 			if (fwrite(buf, 1, (offset + 7) / 8, outp) == 0 &&
1257 			    ferror(outp)) {
1258 				ioerror();
1259 			}
1260 			bytes_out += (offset + 7) / 8;
1261 		}
1262 		offset = 0;
1263 		(void) fflush(outp);
1264 #ifdef DEBUG
1265 		if (verbose)
1266 			(void) fprintf(stderr, "\n");
1267 #endif /* DEBUG */
1268 		if (ferror(outp))
1269 			ioerror();
1270 	}
1271 }
1272 
1273 /*
1274  * Decompress inp to outp.  This routine adapts to the codes in the
1275  * file building the "string" table on-the-fly; requiring no table to
1276  * be stored in the compressed file.  The tables used herein are shared
1277  * with those of the compress() routine.  See the definitions above.
1278  */
1279 
1280 static void
decompress()1281 decompress()
1282 {
1283 	char_type *stackp, *stack_lim;
1284 	int finchar;
1285 	code_int code, oldcode, incode;
1286 	FILE *fout = outp;
1287 
1288 	/*
1289 	 * As above, initialize the first 256 entries in the table.
1290 	 */
1291 	maxcode = MAXCODE(n_bits = INIT_BITS);
1292 	for (code = 255; code >= 0; code--) {
1293 		tab_prefixof(code) = 0;
1294 		tab_suffixof(code) = (char_type)code;
1295 	}
1296 	free_ent = ((block_compress) ? FIRST : 256);
1297 
1298 	finchar = oldcode = getcode();
1299 	if (oldcode == -1)	/* EOF already? */
1300 		return;			/* Get out of here */
1301 	/* first code must be 8 bits = char */
1302 	if (putc((char)finchar, outp) == EOF && ferror(outp)) {
1303 		/* Crash if can't write */
1304 		ioerror();
1305 	}
1306 	stackp = de_stack;
1307 	stack_lim = stack_max;
1308 
1309 	while ((code = getcode()) > -1) {
1310 
1311 		if ((code == CLEAR) && block_compress) {
1312 			for (code = 255; code >= 0; code--)
1313 				tab_prefixof(code) = 0;
1314 			clear_flg = 1;
1315 			free_ent = FIRST - 1;
1316 			if ((code = getcode()) == -1)	/* O, untimely death! */
1317 				break;
1318 		}
1319 		incode = code;
1320 		/*
1321 		 * Special case for KwKwK string.
1322 		 */
1323 		if (code >= free_ent) {
1324 			if (stackp < stack_lim) {
1325 				*stackp++ = (char_type) finchar;
1326 				code = oldcode;
1327 			} else {
1328 				/* badness */
1329 				longjmp(env, 2);
1330 			}
1331 		}
1332 
1333 		/*
1334 		 * Generate output characters in reverse order
1335 		 */
1336 		while (code >= 256) {
1337 			if (stackp < stack_lim) {
1338 				*stackp++ = tab_suffixof(code);
1339 				code = tab_prefixof(code);
1340 			} else {
1341 				/* badness */
1342 				longjmp(env, 2);
1343 			}
1344 		}
1345 		*stackp++ = finchar = tab_suffixof(code);
1346 
1347 		/*
1348 		 * And put them out in forward order
1349 		 */
1350 		do {
1351 			stackp--;
1352 			(void) putc(*stackp, fout);
1353 		} while (stackp > de_stack);
1354 
1355 		if (ferror(fout))
1356 			ioerror();
1357 
1358 		/*
1359 		 * Generate the new entry.
1360 		 */
1361 		if ((code = free_ent) < maxmaxcode) {
1362 			tab_prefixof(code) = (unsigned short) oldcode;
1363 			tab_suffixof(code) = (char_type) finchar;
1364 			free_ent = code+1;
1365 		}
1366 		/*
1367 		 * Remember previous code.
1368 		 */
1369 		oldcode = incode;
1370 	}
1371 	(void) fflush(outp);
1372 	if (ferror(outp))
1373 		ioerror();
1374 }
1375 
1376 /*
1377  * **************************************************************
1378  * TAG( getcode )
1379  *
1380  * Read one code from the standard input.  If EOF, return -1.
1381  * Inputs:
1382  * 	inp
1383  * Outputs:
1384  * 	code or -1 is returned.
1385  */
1386 
1387 code_int
getcode()1388 getcode() {
1389 	code_int code;
1390 	static int offset = 0, size = 0;
1391 	static char_type buf[BITS];
1392 	int r_off, bits;
1393 	char_type *bp = buf;
1394 
1395 	if (clear_flg > 0 || offset >= size || free_ent > maxcode) {
1396 		/*
1397 		 * If the next entry will be too big for the current code
1398 		 * size, then we must increase the size.  This implies reading
1399 		 * a new buffer full, too.
1400 		 */
1401 		if (free_ent > maxcode) {
1402 			n_bits++;
1403 			if (n_bits == maxbits)
1404 				/* won't get any bigger now */
1405 				maxcode = maxmaxcode;
1406 			else
1407 				maxcode = MAXCODE(n_bits);
1408 		}
1409 		if (clear_flg > 0) {
1410 			maxcode = MAXCODE(n_bits = INIT_BITS);
1411 			clear_flg = 0;
1412 		}
1413 		size = fread(buf, 1, n_bits, inp);
1414 
1415 		if (size <= 0) {
1416 			if (feof(inp)) {
1417 				/* end of file */
1418 				return (-1);
1419 			} else if (ferror(inp)) {
1420 				ioerror();
1421 			}
1422 		}
1423 
1424 		offset = 0;
1425 		/* Round size down to integral number of codes */
1426 		size = (size << 3) - (n_bits - 1);
1427 	}
1428 	r_off = offset;
1429 	bits = n_bits;
1430 	/*
1431 	 * Get to the first byte.
1432 	 */
1433 	bp += (r_off >> 3);
1434 	r_off &= 7;
1435 	/* Get first part (low order bits) */
1436 	code = (*bp++ >> r_off);
1437 	bits -= (8 - r_off);
1438 	r_off = 8 - r_off;		/* now, offset into code word */
1439 	/* Get any 8 bit parts in the middle (<=1 for up to 16 bits). */
1440 	if (bits >= 8) {
1441 		code |= *bp++ << r_off;
1442 		r_off += 8;
1443 		bits -= 8;
1444 	}
1445 	/* high order bits. */
1446 	code |= (*bp & rmask[bits]) << r_off;
1447 	offset += n_bits;
1448 
1449 	return (code);
1450 }
1451 
1452 #ifdef DEBUG
1453 static void
printcodes()1454 printcodes()
1455 {
1456 	/*
1457 	 * Just print out codes from input file.  For debugging.
1458 	 */
1459 	code_int code;
1460 	int col = 0, bits;
1461 
1462 	bits = n_bits = INIT_BITS;
1463 	maxcode = MAXCODE(n_bits);
1464 	free_ent = ((block_compress) ? FIRST : 256);
1465 	while ((code = getcode()) >= 0) {
1466 		if ((code == CLEAR) && block_compress) {
1467 			free_ent = FIRST - 1;
1468 			clear_flg = 1;
1469 		} else if (free_ent < maxmaxcode)
1470 			free_ent++;
1471 		if (bits != n_bits) {
1472 			(void) fprintf(stderr, "\nChange to %d bits\n", n_bits);
1473 			bits = n_bits;
1474 			col = 0;
1475 		}
1476 		(void) fprintf(stderr, "%5d%c",
1477 		    code, (col += 6) >= 74 ? (col = 0, '\n') : ' ');
1478 	}
1479 	(void) putc('\n', stderr);
1480 }
1481 
1482 #endif /* DEBUG */
1483 
1484 #ifdef DEBUG
1485 static void
dump_tab()1486 dump_tab()	/* dump string table */
1487 {
1488 	int i, first;
1489 	int ent;
1490 	int stack_top = STACK_SIZE;
1491 	int c;
1492 
1493 	if (do_decomp == 0) {	/* compressing */
1494 		int flag = 1;
1495 
1496 		for (i = 0; i < hsize; i++) {	/* build sort pointers */
1497 			if ((long)htabof(i) >= 0) {
1498 				sorttab[codetabof(i)] = i;
1499 			}
1500 		}
1501 		first = block_compress ? FIRST : 256;
1502 		for (i = first; i < free_ent; i++) {
1503 			(void) fprintf(stderr, "%5d: \"", i);
1504 			de_stack[--stack_top] = '\n';
1505 			de_stack[--stack_top] = '"';
1506 			stack_top =
1507 			    in_stack((htabof(sorttab[i]) >> maxbits) & 0xff,
1508 			    stack_top);
1509 			for (ent = htabof(sorttab[i]) & ((1 << maxbits) -1);
1510 			    ent > 256;
1511 			    ent = htabof(sorttab[ent]) & ((1<<maxbits)-1)) {
1512 				stack_top = in_stack(
1513 				    htabof(sorttab[ent]) >> maxbits,
1514 				    stack_top);
1515 			}
1516 			stack_top = in_stack(ent, stack_top);
1517 			(void) fwrite(&de_stack[stack_top], 1,
1518 			    STACK_SIZE - stack_top, stderr);
1519 			stack_top = STACK_SIZE;
1520 		}
1521 	} else if (!debug) {	/* decompressing */
1522 
1523 		for (i = 0; i < free_ent; i++) {
1524 			ent = i;
1525 			c = tab_suffixof(ent);
1526 			if (isascii(c) && isprint(c))
1527 				(void) fprintf(stderr, "%5d: %5d/'%c'  \"",
1528 				    ent, tab_prefixof(ent), c);
1529 			else
1530 				(void) fprintf(stderr, "%5d: %5d/\\%03o \"",
1531 				    ent, tab_prefixof(ent), c);
1532 			de_stack[--stack_top] = '\n';
1533 			de_stack[--stack_top] = '"';
1534 			for (; ent != NULL;
1535 			    ent = (ent >= FIRST ? tab_prefixof(ent) :
1536 			    NULL)) {
1537 				stack_top = in_stack(tab_suffixof(ent),
1538 				    stack_top);
1539 			}
1540 			(void) fwrite(&de_stack[stack_top], 1,
1541 			    STACK_SIZE - stack_top, stderr);
1542 			stack_top = STACK_SIZE;
1543 		}
1544 	}
1545 }
1546 
1547 #endif /* DEBUG */
1548 #ifdef DEBUG
1549 static int
in_stack(int c,int stack_top)1550 in_stack(int c, int stack_top)
1551 {
1552 	if ((isascii(c) && isprint(c) && c != '\\') || c == ' ') {
1553 		de_stack[--stack_top] = c;
1554 	} else {
1555 		switch (c) {
1556 		case '\n': de_stack[--stack_top] = 'n'; break;
1557 		case '\t': de_stack[--stack_top] = 't'; break;
1558 		case '\b': de_stack[--stack_top] = 'b'; break;
1559 		case '\f': de_stack[--stack_top] = 'f'; break;
1560 		case '\r': de_stack[--stack_top] = 'r'; break;
1561 		case '\\': de_stack[--stack_top] = '\\'; break;
1562 		default:
1563 			de_stack[--stack_top] = '0' + c % 8;
1564 			de_stack[--stack_top] = '0' + (c / 8) % 8;
1565 			de_stack[--stack_top] = '0' + c / 64;
1566 			break;
1567 		}
1568 		de_stack[--stack_top] = '\\';
1569 	}
1570 	return (stack_top);
1571 }
1572 
1573 #endif /* DEBUG */
1574 static void
ioerror()1575 ioerror()
1576 {
1577 	longjmp(env, 1);
1578 }
1579 
1580 static void
copystat(char * ifname,struct stat * ifstat,char * ofname)1581 copystat(char *ifname, struct stat *ifstat, char *ofname)
1582 {
1583 	mode_t mode;
1584 	struct utimbuf timep;
1585 	acl_t *aclp = NULL;
1586 	int error;
1587 	int sattr_exist = 0;
1588 	int xattr_exist = 0;
1589 
1590 	if (pathconf(ifname, _PC_XATTR_EXISTS) == 1)
1591 		xattr_exist = 1;
1592 	if (saflg && sysattr_support(ifname, _PC_SATTR_EXISTS) == 1)
1593 		sattr_exist = 1;
1594 
1595 	if (fclose(outp)) {
1596 		perror(ofname);
1597 		if (!quiet) {
1598 			(void) fprintf(stderr, gettext(" -- file unchanged"));
1599 			newline_needed = 1;
1600 		}
1601 		perm_stat = 1;
1602 	} else if (ifstat == NULL) {	/* Get stat on input file */
1603 		perror(ifname);
1604 		return;
1605 	} else if ((ifstat->st_mode &
1606 	    S_IFMT /* 0170000 */) != S_IFREG /* 0100000 */) {
1607 		if (quiet) {
1608 			(void) fprintf(stderr, "%s: ", ifname);
1609 		}
1610 		(void) fprintf(stderr, gettext(
1611 		    " -- not a regular file: unchanged"));
1612 		newline_needed = 1;
1613 		perm_stat = 1;
1614 	} else if (ifstat->st_nlink > 1) {
1615 		if (quiet) {
1616 			(void) fprintf(stderr, "%s: ", ifname);
1617 		}
1618 		(void) fprintf(stderr, gettext(
1619 		    " -- has %d other links: unchanged"),
1620 		    (uint_t)ifstat->st_nlink - 1);
1621 		newline_needed = 1;
1622 		perm_stat = 1;
1623 	} else if (didnt_shrink && !force) {
1624 		/* No compression: remove file.Z */
1625 		if (!quiet) {
1626 			(void) fprintf(stderr, gettext(
1627 			    " -- file unchanged"));
1628 			newline_needed = 1;
1629 		}
1630 	} else 	if ((xattr_exist || sattr_exist) &&
1631 	    (mv_xattrs(progname, ifname, ofname, sattr_exist, 0)
1632 	    != 0)) {
1633 		(void) fprintf(stderr, gettext(
1634 		    "%s: -- cannot preserve extended attributes or "
1635 		    "system attributes, file unchanged"), ifname);
1636 		newline_needed = 1;
1637 		/* Move attributes back ... */
1638 		xattr_exist = 0;
1639 		sattr_exist = 0;
1640 		if (pathconf(ofname, _PC_XATTR_EXISTS) == 1)
1641 			xattr_exist = 1;
1642 		if (saflg && sysattr_support(ofname, _PC_SATTR_EXISTS) == 1)
1643 			sattr_exist = 1;
1644 		if (sattr_exist || xattr_exist)
1645 			(void) mv_xattrs(progname, ofname, ifname,
1646 			    sattr_exist, 1);
1647 		perm_stat = 1;
1648 	} else { /* ***** Successful Compression ***** */
1649 		mode = ifstat->st_mode & 07777;
1650 		if (chmod(ofname, mode)) {	 /* Copy modes */
1651 			if (errno == EPERM) {
1652 				(void) fprintf(stderr,
1653 				    gettext("failed to chmod %s"
1654 				    "- permisssion denied\n"), ofname);
1655 			}
1656 			perror(ofname);
1657 		}
1658 		error = acl_get(ifname, ACL_NO_TRIVIAL, &aclp);
1659 		if (error != 0) {
1660 			(void) fprintf(stderr, gettext(
1661 			    "%s: failed to retrieve acl : %s\n"),
1662 			    ifname, acl_strerror(error));
1663 			perm_stat = 1;
1664 		}
1665 		if (aclp && (acl_set(ofname, aclp) < 0)) {
1666 			(void) fprintf(stderr,
1667 			    gettext("%s: failed to set acl "
1668 			    "entries\n"), ofname);
1669 			perm_stat = 1;
1670 		}
1671 		if (aclp) {
1672 			acl_free(aclp);
1673 			aclp = NULL;
1674 		}
1675 
1676 		/* Copy ownership */
1677 		(void) chown(ofname, ifstat->st_uid, ifstat->st_gid);
1678 		timep.actime = ifstat->st_atime;
1679 		timep.modtime = ifstat->st_mtime;
1680 		/* Update last accessed and modified times */
1681 		(void) utime(ofname, &timep);
1682 		if (unlink(ifname)) { /* Remove input file */
1683 			if (errno == EPERM) {
1684 				(void) fprintf(stderr,
1685 				    gettext("failed to remove %s"
1686 				    "- permisssion denied\n"), ifname);
1687 			}
1688 			perror(ifname);
1689 		}
1690 		if (!quiet) {
1691 			(void) fprintf(stderr, gettext(
1692 			    " -- replaced with %s"), ofname);
1693 			newline_needed = 1;
1694 		}
1695 		return;		/* Successful return */
1696 	}
1697 
1698 	/* Unsuccessful return -- one of the tests failed */
1699 	if (ofname[0] != '\0') {
1700 		if (unlink(ofname)) {
1701 			if (errno == EPERM) {
1702 				(void) fprintf(stderr,
1703 				    gettext("failed to remove %s"
1704 				    "- permisssion denied\n"), ifname);
1705 			}
1706 			perror(ofname);
1707 		}
1708 
1709 		ofname[0] = '\0';
1710 	}
1711 }
1712 
1713 static void
onintr()1714 onintr()
1715 {
1716 	if (!precious && !use_stdout && ofname[0] != '\0')
1717 		(void) unlink(ofname);
1718 	exit(1);
1719 }
1720 
1721 static void
oops()1722 oops()	/* wild pointer -- assume bad input */
1723 {
1724 	if (do_decomp) {
1725 		(void) fprintf(stderr, gettext("uncompress: corrupt input\n"));
1726 	}
1727 
1728 	if (!use_stdout && ofname[0] != '\0') {
1729 		(void) unlink(ofname);
1730 	}
1731 
1732 	exit(1);
1733 }
1734 
1735 static void
cl_block(count_long in_count)1736 cl_block(count_long in_count)	/* table clear for block compress */
1737 {
1738 	count_long rat;
1739 
1740 	checkpoint = (count_long)in_count + (count_long)CHECK_GAP;
1741 #ifdef DEBUG
1742 	if (debug) {
1743 		(void) fprintf(stderr, "count: %lld, ratio: ",
1744 		    (count_long)in_count);
1745 		prratio(stderr, (count_long)in_count, (count_long)bytes_out);
1746 		(void) fprintf(stderr, "\n");
1747 	}
1748 #endif /* DEBUG */
1749 
1750 	/* shift will overflow */
1751 	if ((count_long)in_count > 0x007fffffffffffffLL) {
1752 		rat = (count_long)bytes_out >> 8;
1753 		if (rat == 0) {		/* Don't divide by zero */
1754 			rat = 0x7fffffffffffffffLL;
1755 		} else {
1756 			rat = (count_long)in_count / (count_long)rat;
1757 		}
1758 	} else {
1759 		/* 8 fractional bits */
1760 		rat = ((count_long)in_count << 8) /(count_long)bytes_out;
1761 	}
1762 	if (rat > ratio) {
1763 		ratio = rat;
1764 	} else {
1765 		ratio = 0;
1766 #ifdef DEBUG
1767 		if (verbose)
1768 			dump_tab();	/* dump string table */
1769 #endif
1770 		cl_hash((count_int) hsize);
1771 		free_ent = FIRST;
1772 		clear_flg = 1;
1773 		output((code_int) CLEAR);
1774 #ifdef DEBUG
1775 		if (debug)
1776 			(void) fprintf(stderr, "clear\n");
1777 #endif /* DEBUG */
1778 	}
1779 }
1780 
1781 static void
cl_hash(count_int hsize)1782 cl_hash(count_int hsize)		/* reset code table */
1783 {
1784 	count_int *htab_p = htab+hsize;
1785 	long i;
1786 	long m1 = -1;
1787 
1788 	i = hsize - 16;
1789 	do {				/* might use Sys V memset(3) here */
1790 		*(htab_p-16) = m1;
1791 		*(htab_p-15) = m1;
1792 		*(htab_p-14) = m1;
1793 		*(htab_p-13) = m1;
1794 		*(htab_p-12) = m1;
1795 		*(htab_p-11) = m1;
1796 		*(htab_p-10) = m1;
1797 		*(htab_p-9) = m1;
1798 		*(htab_p-8) = m1;
1799 		*(htab_p-7) = m1;
1800 		*(htab_p-6) = m1;
1801 		*(htab_p-5) = m1;
1802 		*(htab_p-4) = m1;
1803 		*(htab_p-3) = m1;
1804 		*(htab_p-2) = m1;
1805 		*(htab_p-1) = m1;
1806 		htab_p -= 16;
1807 	} while ((i -= 16) >= 0);
1808 
1809 	for (i += 16; i > 0; i--)
1810 		*--htab_p = m1;
1811 }
1812 
1813 static void
prratio(FILE * stream,count_long num,count_long den)1814 prratio(FILE *stream, count_long num, count_long den)
1815 {
1816 	int q;  /* store percentage */
1817 
1818 	q = (int)(10000LL * (count_long)num / (count_long)den);
1819 	if (q < 0) {
1820 		(void) putc('-', stream);
1821 		q = -q;
1822 	}
1823 	(void) fprintf(stream, "%d%s%02d%%", q / 100,
1824 	    localeconv()->decimal_point, q % 100);
1825 }
1826 
1827 static void
version()1828 version()
1829 {
1830 	(void) fprintf(stderr, "%s, Berkeley 5.9 5/11/86\n", rcs_ident);
1831 	(void) fprintf(stderr, "Options: ");
1832 #ifdef DEBUG
1833 	(void) fprintf(stderr, "DEBUG, ");
1834 #endif
1835 	(void) fprintf(stderr, "BITS = %d\n", BITS);
1836 }
1837 
1838 static void
Usage()1839 Usage()
1840 {
1841 #ifdef DEBUG
1842 	(void) fprintf(stderr,
1843 	"Usage: compress [-dDVfc/] [-b maxbits] [file ...]\n");
1844 #else
1845 	if (strcmp(progname, "compress") == 0) {
1846 		(void) fprintf(stderr,
1847 		    gettext(
1848 		    "Usage: compress [-fv/] [-b maxbits] [file ...]\n"\
1849 		    "       compress c [-fv] [-b maxbits] [file]\n"));
1850 	} else if (strcmp(progname, "uncompress") == 0)
1851 		(void) fprintf(stderr, gettext(
1852 		    "Usage: uncompress [-fv] [-c || -/] [file ...]\n"));
1853 	else if (strcmp(progname, "zcat") == 0)
1854 		(void) fprintf(stderr, gettext("Usage: zcat [file ...]\n"));
1855 
1856 #endif /* DEBUG */
1857 }
1858 
1859 static char *
local_basename(char * path)1860 local_basename(char *path)
1861 {
1862 	char *p;
1863 	char *ret = (char *)path;
1864 
1865 	while ((p = (char *)strpbrk(ret, "/")) != NULL)
1866 		ret = p + 1;
1867 	return (ret);
1868 }
1869 
1870 static int
addDotZ(char * fn,size_t fnsize)1871 addDotZ(char *fn, size_t fnsize)
1872 {
1873 	char *fn_dup;
1874 	char *dir;
1875 	long int max_name;
1876 	long int max_path;
1877 
1878 	fn_dup = strdup(fn);
1879 	dir = dirname(fn_dup);
1880 	max_name = pathconf(dir, _PC_NAME_MAX);
1881 	max_path = pathconf(dir, _PC_PATH_MAX);
1882 	free(fn_dup);
1883 
1884 	/* Check for component length too long */
1885 
1886 	if ((strlen(local_basename(fn)) + 2) > (size_t)max_name) {
1887 		(void) fprintf(stderr,
1888 		    gettext("%s: filename too long to tack on .Z:"
1889 		    " %s\n"), progname, fn);
1890 		return (-1);
1891 	}
1892 
1893 	/* Check for path length too long */
1894 
1895 	if ((strlen(fn) + 2) > (size_t)max_path - 1) {
1896 		(void) fprintf(stderr,
1897 		    gettext("%s: Pathname too long to tack on .Z:"
1898 		    " %s\n"), progname, fn);
1899 		return (-1);
1900 	}
1901 
1902 	if (strlcat(fn, ".Z", fnsize) >= fnsize) {
1903 		(void) fprintf(stderr,
1904 		    gettext("%s: Buffer overflow adding .Z to %s\n"),
1905 		    progname, fn);
1906 		return (-1);
1907 	}
1908 
1909 	return (0);
1910 }
1911