![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to compress.c CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [CSRG BSD Unix] / 43BSDTahoe / ucb / compress|
Return to compress.c CVS log | Up to [CSRG BSD Unix] / 43BSDTahoe / ucb / compress |
1.1 root 1: #ifndef lint
2: static char sccsid[] = "@(#)compress.c @(#)compress.c 5.9 (Berkeley) 5/11/86";
3: #endif not lint
4:
5: /*
6: * Compress - data compression program
7: */
8: #define min(a,b) ((a>b) ? b : a)
9:
10: /*
11: * machine variants which require cc -Dmachine: pdp11, z8000, pcxt
12: */
13:
14: /*
15: * Set USERMEM to the maximum amount of physical user memory available
16: * in bytes. USERMEM is used to determine the maximum BITS that can be used
17: * for compression.
18: *
19: * SACREDMEM is the amount of physical memory saved for others; compress
20: * will hog the rest.
21: */
22: #ifndef SACREDMEM
23: #define SACREDMEM 0
24: #endif
25:
26: #ifndef USERMEM
27: # define USERMEM 450000 /* default user memory */
28: #endif
29:
30: #ifdef interdata /* (Perkin-Elmer) */
31: #define SIGNED_COMPARE_SLOW /* signed compare is slower than unsigned */
32: #endif
33:
34: #ifdef pdp11
35: # define BITS 12 /* max bits/code for 16-bit machine */
36: # define NO_UCHAR /* also if "unsigned char" functions as signed char */
37: # undef USERMEM
38: #endif /* pdp11 */ /* don't forget to compile with -i */
39:
40: #ifdef z8000
41: # define BITS 12
42: # undef vax /* weird preprocessor */
43: # undef USERMEM
44: #endif /* z8000 */
45:
46: #ifdef pcxt
47: # define BITS 12
48: # undef USERMEM
49: #endif /* pcxt */
50:
51: #ifdef USERMEM
52: # if USERMEM >= (433484+SACREDMEM)
53: # define PBITS 16
54: # else
55: # if USERMEM >= (229600+SACREDMEM)
56: # define PBITS 15
57: # else
58: # if USERMEM >= (127536+SACREDMEM)
59: # define PBITS 14
60: # else
61: # if USERMEM >= (73464+SACREDMEM)
62: # define PBITS 13
63: # else
64: # define PBITS 12
65: # endif
66: # endif
67: # endif
68: # endif
69: # undef USERMEM
70: #endif /* USERMEM */
71:
72: #ifdef PBITS /* Preferred BITS for this memory size */
73: # ifndef BITS
74: # define BITS PBITS
75: # endif BITS
76: #endif /* PBITS */
77:
78: #if BITS == 16
79: # define HSIZE 69001 /* 95% occupancy */
80: #endif
81: #if BITS == 15
82: # define HSIZE 35023 /* 94% occupancy */
83: #endif
84: #if BITS == 14
85: # define HSIZE 18013 /* 91% occupancy */
86: #endif
87: #if BITS == 13
88: # define HSIZE 9001 /* 91% occupancy */
89: #endif
90: #if BITS <= 12
91: # define HSIZE 5003 /* 80% occupancy */
92: #endif
93:
94: #ifdef M_XENIX /* Stupid compiler can't handle arrays with */
95: # if BITS == 16 /* more than 65535 bytes - so we fake it */
96: # define XENIX_16
97: # else
98: # if BITS > 13 /* Code only handles BITS = 12, 13, or 16 */
99: # define BITS 13
100: # endif
101: # endif
102: #endif
103:
104: /*
105: * a code_int must be able to hold 2**BITS values of type int, and also -1
106: */
107: #if BITS > 15
108: typedef long int code_int;
109: #else
110: typedef int code_int;
111: #endif
112:
113: #ifdef SIGNED_COMPARE_SLOW
114: typedef unsigned long int count_int;
115: typedef unsigned short int count_short;
116: #else
117: typedef long int count_int;
118: #endif
119:
120: #ifdef NO_UCHAR
121: typedef char char_type;
122: #else
123: typedef unsigned char char_type;
124: #endif /* UCHAR */
125: char_type magic_header[] = { "\037\235" }; /* 1F 9D */
126:
127: /* Defines for third byte of header */
128: #define BIT_MASK 0x1f
129: #define BLOCK_MASK 0x80
130: /* Masks 0x40 and 0x20 are free. I think 0x20 should mean that there is
131: a fourth header byte (for expansion).
132: */
133: #define INIT_BITS 9 /* initial number of bits/code */
134:
135: /*
136: * compress.c - File compression ala IEEE Computer, June 1984.
137: *
138: * Authors: Spencer W. Thomas (decvax!utah-cs!thomas)
139: * Jim McKie (decvax!mcvax!jim)
140: * Steve Davies (decvax!vax135!petsd!peora!srd)
141: * Ken Turkowski (decvax!decwrl!turtlevax!ken)
142: * James A. Woods (decvax!ihnp4!ames!jaw)
143: * Joe Orost (decvax!vax135!petsd!joe)
144: *
145: * $Header: compress.c,v 4.0 85/07/30 12:50:00 joe Release $
146: * $Log: compress.c,v $
147: * Revision 4.0 85/07/30 12:50:00 joe
148: * Removed ferror() calls in output routine on every output except first.
149: * Prepared for release to the world.
150: *
151: * Revision 3.6 85/07/04 01:22:21 joe
152: * Remove much wasted storage by overlaying hash table with the tables
153: * used by decompress: tab_suffix[1<<BITS], stack[8000]. Updated USERMEM
154: * computations. Fixed dump_tab() DEBUG routine.
155: *
156: * Revision 3.5 85/06/30 20:47:21 jaw
157: * Change hash function to use exclusive-or. Rip out hash cache. These
158: * speedups render the megamemory version defunct, for now. Make decoder
159: * stack global. Parts of the RCS trunks 2.7, 2.6, and 2.1 no longer apply.
160: *
161: * Revision 3.4 85/06/27 12:00:00 ken
162: * Get rid of all floating-point calculations by doing all compression ratio
163: * calculations in fixed point.
164: *
165: * Revision 3.3 85/06/24 21:53:24 joe
166: * Incorporate portability suggestion for M_XENIX. Got rid of text on #else
167: * and #endif lines. Cleaned up #ifdefs for vax and interdata.
168: *
169: * Revision 3.2 85/06/06 21:53:24 jaw
170: * Incorporate portability suggestions for Z8000, IBM PC/XT from mailing list.
171: * Default to "quiet" output (no compression statistics).
172: *
173: * Revision 3.1 85/05/12 18:56:13 jaw
174: * Integrate decompress() stack speedups (from early pointer mods by McKie).
175: * Repair multi-file USERMEM gaffe. Unify 'force' flags to mimic semantics
176: * of SVR2 'pack'. Streamline block-compress table clear logic. Increase
177: * output byte count by magic number size.
178: *
179: * Revision 3.0 84/11/27 11:50:00 petsd!joe
180: * Set HSIZE depending on BITS. Set BITS depending on USERMEM. Unrolled
181: * loops in clear routines. Added "-C" flag for 2.0 compatibility. Used
182: * unsigned compares on Perkin-Elmer. Fixed foreground check.
183: *
184: * Revision 2.7 84/11/16 19:35:39 ames!jaw
185: * Cache common hash codes based on input statistics; this improves
186: * performance for low-density raster images. Pass on #ifdef bundle
187: * from Turkowski.
188: *
189: * Revision 2.6 84/11/05 19:18:21 ames!jaw
190: * Vary size of hash tables to reduce time for small files.
191: * Tune PDP-11 hash function.
192: *
193: * Revision 2.5 84/10/30 20:15:14 ames!jaw
194: * Junk chaining; replace with the simpler (and, on the VAX, faster)
195: * double hashing, discussed within. Make block compression standard.
196: *
197: * Revision 2.4 84/10/16 11:11:11 ames!jaw
198: * Introduce adaptive reset for block compression, to boost the rate
199: * another several percent. (See mailing list notes.)
200: *
201: * Revision 2.3 84/09/22 22:00:00 petsd!joe
202: * Implemented "-B" block compress. Implemented REVERSE sorting of tab_next.
203: * Bug fix for last bits. Changed fwrite to putchar loop everywhere.
204: *
205: * Revision 2.2 84/09/18 14:12:21 ames!jaw
206: * Fold in news changes, small machine typedef from thomas,
207: * #ifdef interdata from joe.
208: *
209: * Revision 2.1 84/09/10 12:34:56 ames!jaw
210: * Configured fast table lookup for 32-bit machines.
211: * This cuts user time in half for b <= FBITS, and is useful for news batching
212: * from VAX to PDP sites. Also sped up decompress() [fwrite->putc] and
213: * added signal catcher [plus beef in writeerr()] to delete effluvia.
214: *
215: * Revision 2.0 84/08/28 22:00:00 petsd!joe
216: * Add check for foreground before prompting user. Insert maxbits into
217: * compressed file. Force file being uncompressed to end with ".Z".
218: * Added "-c" flag and "zcat". Prepared for release.
219: *
220: * Revision 1.10 84/08/24 18:28:00 turtlevax!ken
221: * Will only compress regular files (no directories), added a magic number
222: * header (plus an undocumented -n flag to handle old files without headers),
223: * added -f flag to force overwriting of possibly existing destination file,
224: * otherwise the user is prompted for a response. Will tack on a .Z to a
225: * filename if it doesn't have one when decompressing. Will only replace
226: * file if it was compressed.
227: *
228: * Revision 1.9 84/08/16 17:28:00 turtlevax!ken
229: * Removed scanargs(), getopt(), added .Z extension and unlimited number of
230: * filenames to compress. Flags may be clustered (-Ddvb12) or separated
231: * (-D -d -v -b 12), or combination thereof. Modes and other status is
232: * copied with copystat(). -O bug for 4.2 seems to have disappeared with
233: * 1.8.
234: *
235: * Revision 1.8 84/08/09 23:15:00 joe
236: * Made it compatible with vax version, installed jim's fixes/enhancements
237: *
238: * Revision 1.6 84/08/01 22:08:00 joe
239: * Sped up algorithm significantly by sorting the compress chain.
240: *
241: * Revision 1.5 84/07/13 13:11:00 srd
242: * Added C version of vax asm routines. Changed structure to arrays to
243: * save much memory. Do unsigned compares where possible (faster on
244: * Perkin-Elmer)
245: *
246: * Revision 1.4 84/07/05 03:11:11 thomas
247: * Clean up the code a little and lint it. (Lint complains about all
248: * the regs used in the asm, but I'm not going to "fix" this.)
249: *
250: * Revision 1.3 84/07/05 02:06:54 thomas
251: * Minor fixes.
252: *
253: * Revision 1.2 84/07/05 00:27:27 thomas
254: * Add variable bit length output.
255: *
256: */
257: static char rcs_ident[] = "$Header: compress.c,v 4.0 85/07/30 12:50:00 joe Release $";
258:
259: #include <stdio.h>
260: #include <ctype.h>
261: #include <signal.h>
262: #include <sys/types.h>
263: #include <sys/stat.h>
264: #ifdef notdef
265: #include <sys/ioctl.h>
266: #endif
267:
268: #define ARGVAL() (*++(*argv) || (--argc && *++argv))
269:
270: int n_bits; /* number of bits/code */
271: int maxbits = BITS; /* user settable max # bits/code */
272: code_int maxcode; /* maximum code, given n_bits */
273: code_int maxmaxcode = 1 << BITS; /* should NEVER generate this code */
274: #ifdef COMPATIBLE /* But wrong! */
275: # define MAXCODE(n_bits) (1 << (n_bits) - 1)
276: #else
277: # define MAXCODE(n_bits) ((1 << (n_bits)) - 1)
278: #endif /* COMPATIBLE */
279:
280: #ifdef XENIX_16
281: count_int htab0[8192];
282: count_int htab1[8192];
283: count_int htab2[8192];
284: count_int htab3[8192];
285: count_int htab4[8192];
286: count_int htab5[8192];
287: count_int htab6[8192];
288: count_int htab7[8192];
289: count_int htab8[HSIZE-65536];
290: count_int * htab[9] = {
291: htab0, htab1, htab2, htab3, htab4, htab5, htab6, htab7, htab8 };
292:
293: #define htabof(i) (htab[(i) >> 13][(i) & 0x1fff])
294: unsigned short code0tab[16384];
295: unsigned short code1tab[16384];
296: unsigned short code2tab[16384];
297: unsigned short code3tab[16384];
298: unsigned short code4tab[16384];
299: unsigned short * codetab[5] = {
300: code0tab, code1tab, code2tab, code3tab, code4tab };
301:
302: #define codetabof(i) (codetab[(i) >> 14][(i) & 0x3fff])
303:
304: #else /* Normal machine */
305:
306: #ifdef sel /* gould base register braindamage */
307: /*NOBASE*/
308: count_int htab [HSIZE];
309: unsigned short codetab [HSIZE];
310: /*NOBASE*/
311: #else
312: count_int htab [HSIZE];
313: unsigned short codetab [HSIZE];
314: #endif sel
315:
316: #define htabof(i) htab[i]
317: #define codetabof(i) codetab[i]
318: #endif /* XENIX_16 */
319: code_int hsize = HSIZE; /* for dynamic table sizing */
320: count_int fsize;
321:
322: /*
323: * To save much memory, we overlay the table used by compress() with those
324: * used by decompress(). The tab_prefix table is the same size and type
325: * as the codetab. The tab_suffix table needs 2**BITS characters. We
326: * get this from the beginning of htab. The output stack uses the rest
327: * of htab, and contains characters. There is plenty of room for any
328: * possible stack (stack used to be 8000 characters).
329: */
330:
331: #define tab_prefixof(i) codetabof(i)
332: #ifdef XENIX_16
333: # define tab_suffixof(i) ((char_type *)htab[(i)>>15])[(i) & 0x7fff]
334: # define de_stack ((char_type *)(htab2))
335: #else /* Normal machine */
336: # define tab_suffixof(i) ((char_type *)(htab))[i]
337: # define de_stack ((char_type *)&tab_suffixof(1<<BITS))
338: #endif /* XENIX_16 */
339:
340: code_int free_ent = 0; /* first unused entry */
341: int exit_stat = 0; /* per-file status */
342: int perm_stat = 0; /* permanent status */
343:
344: code_int getcode();
345:
346: Usage() {
347: #ifdef DEBUG
348: fprintf(stderr,"Usage: compress [-dDVfc] [-b maxbits] [file ...]\n");
349: }
350: int debug = 0;
351: #else
352: fprintf(stderr,"Usage: compress [-fvc] [-b maxbits] [file ...]\n");
353: }
354: #endif /* DEBUG */
355: int nomagic = 0; /* Use a 3-byte magic number header, unless old file */
356: int zcat_flg = 0; /* Write output on stdout, suppress messages */
357: int precious = 1; /* Don't unlink output file on interrupt */
358: int quiet = 1; /* don't tell me about compression */
359:
360: /*
361: * block compression parameters -- after all codes are used up,
362: * and compression rate changes, start over.
363: */
364: int block_compress = BLOCK_MASK;
365: int clear_flg = 0;
366: long int ratio = 0;
367: #define CHECK_GAP 10000 /* ratio check interval */
368: count_int checkpoint = CHECK_GAP;
369: /*
370: * the next two codes should not be changed lightly, as they must not
371: * lie within the contiguous general code space.
372: */
373: #define FIRST 257 /* first free entry */
374: #define CLEAR 256 /* table clear output code */
375:
376: int force = 0;
377: char ofname [100];
378: #ifdef DEBUG
379: int verbose = 0;
380: #endif /* DEBUG */
381: int (*oldint)();
382: int bgnd_flag;
383:
384: int do_decomp = 0;
385:
386: /*****************************************************************
387: * TAG( main )
388: *
389: * Algorithm from "A Technique for High Performance Data Compression",
390: * Terry A. Welch, IEEE Computer Vol 17, No 6 (June 1984), pp 8-19.
391: *
392: * Usage: compress [-dfvc] [-b bits] [file ...]
393: * Inputs:
394: * -d: If given, decompression is done instead.
395: *
396: * -c: Write output on stdout, don't remove original.
397: *
398: * -b: Parameter limits the max number of bits/code.
399: *
400: * -f: Forces output file to be generated, even if one already
401: * exists, and even if no space is saved by compressing.
402: * If -f is not used, the user will be prompted if stdin is
403: * a tty, otherwise, the output file will not be overwritten.
404: *
405: * -v: Write compression statistics
406: *
407: * file ...: Files to be compressed. If none specified, stdin
408: * is used.
409: * Outputs:
410: * file.Z: Compressed form of file with same mode, owner, and utimes
411: * or stdout (if stdin used as input)
412: *
413: * Assumptions:
414: * When filenames are given, replaces with the compressed version
415: * (.Z suffix) only if the file decreases in size.
416: * Algorithm:
417: * Modified Lempel-Ziv method (LZW). Basically finds common
418: * substrings and replaces them with a variable size code. This is
419: * deterministic, and can be done on the fly. Thus, the decompression
420: * procedure needs no input table, but tracks the way the table was built.
421: */
422:
423: main( argc, argv )
424: register int argc; char **argv;
425: {
426: int overwrite = 0; /* Do not overwrite unless given -f flag */
427: char tempname[100];
428: char **filelist, **fileptr;
429: char *cp, *rindex(), *malloc();
430: struct stat statbuf;
431: extern onintr(), oops();
432:
433: /* This bg check only works for sh. */
434: if ( (oldint = signal ( SIGINT, SIG_IGN )) != SIG_IGN ) {
435: signal ( SIGINT, onintr );
436: signal ( SIGSEGV, oops );
437: }
438: bgnd_flag = oldint != SIG_DFL;
439: #ifdef notdef /* This works for csh but we don't want it. */
440: { int tgrp;
441: if (bgnd_flag == 0 && ioctl(2, TIOCGPGRP, (char *)&tgrp) == 0 &&
442: getpgrp(0) != tgrp)
443: bgnd_flag = 1;
444: }
445: #endif
446:
447: #ifdef COMPATIBLE
448: nomagic = 1; /* Original didn't have a magic number */
449: #endif /* COMPATIBLE */
450:
451: filelist = fileptr = (char **)(malloc(argc * sizeof(*argv)));
452: *filelist = NULL;
453:
454: if((cp = rindex(argv[0], '/')) != 0) {
455: cp++;
456: } else {
457: cp = argv[0];
458: }
459: if(strcmp(cp, "uncompress") == 0) {
460: do_decomp = 1;
461: } else if(strcmp(cp, "zcat") == 0) {
462: do_decomp = 1;
463: zcat_flg = 1;
464: }
465:
466: #ifdef BSD4_2
467: /* 4.2BSD dependent - take it out if not */
468: setlinebuf( stderr );
469: #endif /* BSD4_2 */
470:
471: /* Argument Processing
472: * All flags are optional.
473: * -D => debug
474: * -V => print Version; debug verbose
475: * -d => do_decomp
476: * -v => unquiet
477: * -f => force overwrite of output file
478: * -n => no header: useful to uncompress old files
479: * -b maxbits => maxbits. If -b is specified, then maxbits MUST be
480: * given also.
481: * -c => cat all output to stdout
482: * -C => generate output compatible with compress 2.0.
483: * if a string is left, must be an input filename.
484: */
485: for (argc--, argv++; argc > 0; argc--, argv++) {
486: if (**argv == '-') { /* A flag argument */
487: while (*++(*argv)) { /* Process all flags in this arg */
488: switch (**argv) {
489: #ifdef DEBUG
490: case 'D':
491: debug = 1;
492: break;
493: case 'V':
494: verbose = 1;
495: version();
496: break;
497: #else
498: case 'V':
499: version();
500: break;
501: #endif /* DEBUG */
502: case 'v':
503: quiet = 0;
504: break;
505: case 'd':
506: do_decomp = 1;
507: break;
508: case 'f':
509: case 'F':
510: overwrite = 1;
511: force = 1;
512: break;
513: case 'n':
514: nomagic = 1;
515: break;
516: case 'C':
517: block_compress = 0;
518: break;
519: case 'b':
520: if (!ARGVAL()) {
521: fprintf(stderr, "Missing maxbits\n");
522: Usage();
523: exit(1);
524: }
525: maxbits = atoi(*argv);
526: goto nextarg;
527: case 'c':
528: zcat_flg = 1;
529: break;
530: case 'q':
531: quiet = 1;
532: break;
533: default:
534: fprintf(stderr, "Unknown flag: '%c'; ", **argv);
535: Usage();
536: exit(1);
537: }
538: }
539: }
540: else { /* Input file name */
541: *fileptr++ = *argv; /* Build input file list */
542: *fileptr = NULL;
543: /* process nextarg; */
544: }
545: nextarg: continue;
546: }
547:
548: if(maxbits < INIT_BITS) maxbits = INIT_BITS;
549: if (maxbits > BITS) maxbits = BITS;
550: maxmaxcode = 1 << maxbits;
551:
552: if (*filelist != NULL) {
553: for (fileptr = filelist; *fileptr; fileptr++) {
554: exit_stat = 0;
555: if (do_decomp) { /* DECOMPRESSION */
556: /* Check for .Z suffix */
557: if (strcmp(*fileptr + strlen(*fileptr) - 2, ".Z") != 0) {
558: /* No .Z: tack one on */
559: strcpy(tempname, *fileptr);
560: strcat(tempname, ".Z");
561: *fileptr = tempname;
562: }
563: /* Open input file */
564: if ((freopen(*fileptr, "r", stdin)) == NULL) {
565: perror(*fileptr);
566: perm_stat = 1;
567: continue;
568: }
569: /* Check the magic number */
570: if (nomagic == 0) {
571: if ((getchar() != (magic_header[0] & 0xFF))
572: || (getchar() != (magic_header[1] & 0xFF))) {
573: fprintf(stderr, "%s: not in compressed format\n",
574: *fileptr);
575: continue;
576: }
577: maxbits = getchar(); /* set -b from file */
578: block_compress = maxbits & BLOCK_MASK;
579: maxbits &= BIT_MASK;
580: maxmaxcode = 1 << maxbits;
581: if(maxbits > BITS) {
582: fprintf(stderr,
583: "%s: compressed with %d bits, can only handle %d bits\n",
584: *fileptr, maxbits, BITS);
585: continue;
586: }
587: }
588: /* Generate output filename */
589: strcpy(ofname, *fileptr);
590: ofname[strlen(*fileptr) - 2] = '\0'; /* Strip off .Z */
591: } else { /* COMPRESSION */
592: if (strcmp(*fileptr + strlen(*fileptr) - 2, ".Z") == 0) {
593: fprintf(stderr, "%s: already has .Z suffix -- no change\n",
594: *fileptr);
595: continue;
596: }
597: /* Open input file */
598: if ((freopen(*fileptr, "r", stdin)) == NULL) {
599: perror(*fileptr);
600: perm_stat = 1;
601: continue;
602: }
603: stat ( *fileptr, &statbuf );
604: fsize = (long) statbuf.st_size;
605: /*
606: * tune hash table size for small files -- ad hoc,
607: * but the sizes match earlier #defines, which
608: * serve as upper bounds on the number of output codes.
609: */
610: hsize = HSIZE;
611: if ( fsize < (1 << 12) )
612: hsize = min ( 5003, HSIZE );
613: else if ( fsize < (1 << 13) )
614: hsize = min ( 9001, HSIZE );
615: else if ( fsize < (1 << 14) )
616: hsize = min ( 18013, HSIZE );
617: else if ( fsize < (1 << 15) )
618: hsize = min ( 35023, HSIZE );
619: else if ( fsize < 47000 )
620: hsize = min ( 50021, HSIZE );
621:
622: /* Generate output filename */
623: strcpy(ofname, *fileptr);
624: #ifndef BSD4_2 /* Short filenames */
625: if ((cp=rindex(ofname,'/')) != NULL) cp++;
626: else cp = ofname;
627: if (strlen(cp) > 12) {
628: fprintf(stderr,"%s: filename too long to tack on .Z\n",cp);
629: continue;
630: }
631: #endif /* BSD4_2 Long filenames allowed */
632: strcat(ofname, ".Z");
633: }
634: /* Check for overwrite of existing file */
635: if (overwrite == 0 && zcat_flg == 0) {
636: if (stat(ofname, &statbuf) == 0) {
637: char response[2];
638: response[0] = 'n';
639: fprintf(stderr, "%s already exists;", ofname);
640: if (bgnd_flag == 0 && isatty(2)) {
641: fprintf(stderr, " do you wish to overwrite %s (y or n)? ",
642: ofname);
643: fflush(stderr);
644: read(2, response, 2);
645: while (response[1] != '\n') {
646: if (read(2, response+1, 1) < 0) { /* Ack! */
647: perror("stderr"); break;
648: }
649: }
650: }
651: if (response[0] != 'y') {
652: fprintf(stderr, "\tnot overwritten\n");
653: continue;
654: }
655: }
656: }
657: if(zcat_flg == 0) { /* Open output file */
658: if (freopen(ofname, "w", stdout) == NULL) {
659: perror(ofname);
660: perm_stat = 1;
661: continue;
662: }
663: precious = 0;
664: if(!quiet)
665: fprintf(stderr, "%s: ", *fileptr);
666: }
667:
668: /* Actually do the compression/decompression */
669: if (do_decomp == 0) compress();
670: #ifndef DEBUG
671: else decompress();
672: #else
673: else if (debug == 0) decompress();
674: else printcodes();
675: if (verbose) dump_tab();
676: #endif /* DEBUG */
677: if(zcat_flg == 0) {
678: copystat(*fileptr, ofname); /* Copy stats */
679: precious = 1;
680: if((exit_stat == 1) || (!quiet))
681: putc('\n', stderr);
682: }
683: }
684: } else { /* Standard input */
685: if (do_decomp == 0) {
686: compress();
687: #ifdef DEBUG
688: if(verbose) dump_tab();
689: #endif /* DEBUG */
690: if(!quiet)
691: putc('\n', stderr);
692: } else {
693: /* Check the magic number */
694: if (nomagic == 0) {
695: if ((getchar()!=(magic_header[0] & 0xFF))
696: || (getchar()!=(magic_header[1] & 0xFF))) {
697: fprintf(stderr, "stdin: not in compressed format\n");
698: exit(1);
699: }
700: maxbits = getchar(); /* set -b from file */
701: block_compress = maxbits & BLOCK_MASK;
702: maxbits &= BIT_MASK;
703: maxmaxcode = 1 << maxbits;
704: fsize = 100000; /* assume stdin large for USERMEM */
705: if(maxbits > BITS) {
706: fprintf(stderr,
707: "stdin: compressed with %d bits, can only handle %d bits\n",
708: maxbits, BITS);
709: exit(1);
710: }
711: }
712: #ifndef DEBUG
713: decompress();
714: #else
715: if (debug == 0) decompress();
716: else printcodes();
717: if (verbose) dump_tab();
718: #endif /* DEBUG */
719: }
720: }
721: exit(perm_stat ? perm_stat : exit_stat);
722: }
723:
724: static int offset;
725: long int in_count = 1; /* length of input */
726: long int bytes_out; /* length of compressed output */
727: long int out_count = 0; /* # of codes output (for debugging) */
728:
729: /*
730: * compress stdin to stdout
731: *
732: * Algorithm: use open addressing double hashing (no chaining) on the
733: * prefix code / next character combination. We do a variant of Knuth's
734: * algorithm D (vol. 3, sec. 6.4) along with G. Knott's relatively-prime
735: * secondary probe. Here, the modular division first probe is gives way
736: * to a faster exclusive-or manipulation. Also do block compression with
737: * an adaptive reset, whereby the code table is cleared when the compression
738: * ratio decreases, but after the table fills. The variable-length output
739: * codes are re-sized at this point, and a special CLEAR code is generated
740: * for the decompressor. Late addition: construct the table according to
741: * file size for noticeable speed improvement on small files. Please direct
742: * questions about this implementation to ames!jaw.
743: */
744:
745: compress() {
746: register long fcode;
747: register code_int i = 0;
748: register int c;
749: register code_int ent;
750: #ifdef XENIX_16
751: register code_int disp;
752: #else /* Normal machine */
753: register int disp;
754: #endif
755: register code_int hsize_reg;
756: register int hshift;
757:
758: #ifndef COMPATIBLE
759: if (nomagic == 0) {
760: putchar(magic_header[0]); putchar(magic_header[1]);
761: putchar((char)(maxbits | block_compress));
762: if(ferror(stdout))
763: writeerr();
764: }
765: #endif /* COMPATIBLE */
766:
767: offset = 0;
768: bytes_out = 3; /* includes 3-byte header mojo */
769: out_count = 0;
770: clear_flg = 0;
771: ratio = 0;
772: in_count = 1;
773: checkpoint = CHECK_GAP;
774: maxcode = MAXCODE(n_bits = INIT_BITS);
775: free_ent = ((block_compress) ? FIRST : 256 );
776:
777: ent = getchar ();
778:
779: hshift = 0;
780: for ( fcode = (long) hsize; fcode < 65536L; fcode *= 2L )
781: hshift++;
782: hshift = 8 - hshift; /* set hash code range bound */
783:
784: hsize_reg = hsize;
785: cl_hash( (count_int) hsize_reg); /* clear hash table */
786:
787: #ifdef SIGNED_COMPARE_SLOW
788: while ( (c = getchar()) != (unsigned) EOF ) {
789: #else
790: while ( (c = getchar()) != EOF ) {
791: #endif
792: in_count++;
793: fcode = (long) (((long) c << maxbits) + ent);
794: i = ((c << hshift) ^ ent); /* xor hashing */
795:
796: if ( htabof (i) == fcode ) {
797: ent = codetabof (i);
798: continue;
799: } else if ( (long)htabof (i) < 0 ) /* empty slot */
800: goto nomatch;
801: disp = hsize_reg - i; /* secondary hash (after G. Knott) */
802: if ( i == 0 )
803: disp = 1;
804: probe:
805: if ( (i -= disp) < 0 )
806: i += hsize_reg;
807:
808: if ( htabof (i) == fcode ) {
809: ent = codetabof (i);
810: continue;
811: }
812: if ( (long)htabof (i) > 0 )
813: goto probe;
814: nomatch:
815: output ( (code_int) ent );
816: out_count++;
817: ent = c;
818: #ifdef SIGNED_COMPARE_SLOW
819: if ( (unsigned) free_ent < (unsigned) maxmaxcode) {
820: #else
821: if ( free_ent < maxmaxcode ) {
822: #endif
823: codetabof (i) = free_ent++; /* code -> hashtable */
824: htabof (i) = fcode;
825: }
826: else if ( (count_int)in_count >= checkpoint && block_compress )
827: cl_block ();
828: }
829: /*
830: * Put out the final code.
831: */
832: output( (code_int)ent );
833: out_count++;
834: output( (code_int)-1 );
835:
836: /*
837: * Print out stats on stderr
838: */
839: if(zcat_flg == 0 && !quiet) {
840: #ifdef DEBUG
841: fprintf( stderr,
842: "%ld chars in, %ld codes (%ld bytes) out, compression factor: ",
843: in_count, out_count, bytes_out );
844: prratio( stderr, in_count, bytes_out );
845: fprintf( stderr, "\n");
846: fprintf( stderr, "\tCompression as in compact: " );
847: prratio( stderr, in_count-bytes_out, in_count );
848: fprintf( stderr, "\n");
849: fprintf( stderr, "\tLargest code (of last block) was %d (%d bits)\n",
850: free_ent - 1, n_bits );
851: #else /* !DEBUG */
852: fprintf( stderr, "Compression: " );
853: prratio( stderr, in_count-bytes_out, in_count );
854: #endif /* DEBUG */
855: }
856: if(bytes_out > in_count) /* exit(2) if no savings */
857: exit_stat = 2;
858: return;
859: }
860:
861: /*****************************************************************
862: * TAG( output )
863: *
864: * Output the given code.
865: * Inputs:
866: * code: A n_bits-bit integer. If == -1, then EOF. This assumes
867: * that n_bits =< (long)wordsize - 1.
868: * Outputs:
869: * Outputs code to the file.
870: * Assumptions:
871: * Chars are 8 bits long.
872: * Algorithm:
873: * Maintain a BITS character long buffer (so that 8 codes will
874: * fit in it exactly). Use the VAX insv instruction to insert each
875: * code in turn. When the buffer fills up empty it and start over.
876: */
877:
878: static char buf[BITS];
879:
880: #ifndef vax
881: char_type lmask[9] = {0xff, 0xfe, 0xfc, 0xf8, 0xf0, 0xe0, 0xc0, 0x80, 0x00};
882: char_type rmask[9] = {0x00, 0x01, 0x03, 0x07, 0x0f, 0x1f, 0x3f, 0x7f, 0xff};
883: #endif /* vax */
884:
885: output( code )
886: code_int code;
887: {
888: #ifdef DEBUG
889: static int col = 0;
890: #endif /* DEBUG */
891:
892: /*
893: * On the VAX, it is important to have the register declarations
894: * in exactly the order given, or the asm will break.
895: */
896: register int r_off = offset, bits= n_bits;
897: register char * bp = buf;
898:
899: #ifdef DEBUG
900: if ( verbose )
901: fprintf( stderr, "%5d%c", code,
902: (col+=6) >= 74 ? (col = 0, '\n') : ' ' );
903: #endif /* DEBUG */
904: if ( code >= 0 ) {
905: #ifdef vax
906: /* VAX DEPENDENT!! Implementation on other machines is below.
907: *
908: * Translation: Insert BITS bits from the argument starting at
909: * offset bits from the beginning of buf.
910: */
911: 0; /* Work around for pcc -O bug with asm and if stmt */
912: asm( "insv 4(ap),r11,r10,(r9)" );
913: #else /* not a vax */
914: /*
915: * byte/bit numbering on the VAX is simulated by the following code
916: */
917: /*
918: * Get to the first byte.
919: */
920: bp += (r_off >> 3);
921: r_off &= 7;
922: /*
923: * Since code is always >= 8 bits, only need to mask the first
924: * hunk on the left.
925: */
926: *bp = (*bp & rmask[r_off]) | (code << r_off) & lmask[r_off];
927: bp++;
928: bits -= (8 - r_off);
929: code >>= 8 - r_off;
930: /* Get any 8 bit parts in the middle (<=1 for up to 16 bits). */
931: if ( bits >= 8 ) {
932: *bp++ = code;
933: code >>= 8;
934: bits -= 8;
935: }
936: /* Last bits. */
937: if(bits)
938: *bp = code;
939: #endif /* vax */
940: offset += n_bits;
941: if ( offset == (n_bits << 3) ) {
942: bp = buf;
943: bits = n_bits;
944: bytes_out += bits;
945: do
946: putchar(*bp++);
947: while(--bits);
948: offset = 0;
949: }
950:
951: /*
952: * If the next entry is going to be too big for the code size,
953: * then increase it, if possible.
954: */
955: if ( free_ent > maxcode || (clear_flg > 0))
956: {
957: /*
958: * Write the whole buffer, because the input side won't
959: * discover the size increase until after it has read it.
960: */
961: if ( offset > 0 ) {
962: if( fwrite( buf, 1, n_bits, stdout ) != n_bits)
963: writeerr();
964: bytes_out += n_bits;
965: }
966: offset = 0;
967:
968: if ( clear_flg ) {
969: maxcode = MAXCODE (n_bits = INIT_BITS);
970: clear_flg = 0;
971: }
972: else {
973: n_bits++;
974: if ( n_bits == maxbits )
975: maxcode = maxmaxcode;
976: else
977: maxcode = MAXCODE(n_bits);
978: }
979: #ifdef DEBUG
980: if ( debug ) {
981: fprintf( stderr, "\nChange to %d bits\n", n_bits );
982: col = 0;
983: }
984: #endif /* DEBUG */
985: }
986: } else {
987: /*
988: * At EOF, write the rest of the buffer.
989: */
990: if ( offset > 0 )
991: fwrite( buf, 1, (offset + 7) / 8, stdout );
992: bytes_out += (offset + 7) / 8;
993: offset = 0;
994: fflush( stdout );
995: #ifdef DEBUG
996: if ( verbose )
997: fprintf( stderr, "\n" );
998: #endif /* DEBUG */
999: if( ferror( stdout ) )
1000: writeerr();
1001: }
1002: }
1003:
1004: /*
1005: * Decompress stdin to stdout. This routine adapts to the codes in the
1006: * file building the "string" table on-the-fly; requiring no table to
1007: * be stored in the compressed file. The tables used herein are shared
1008: * with those of the compress() routine. See the definitions above.
1009: */
1010:
1011: decompress() {
1012: register char_type *stackp;
1013: register int finchar;
1014: register code_int code, oldcode, incode;
1015:
1016: /*
1017: * As above, initialize the first 256 entries in the table.
1018: */
1019: maxcode = MAXCODE(n_bits = INIT_BITS);
1020: for ( code = 255; code >= 0; code-- ) {
1021: tab_prefixof(code) = 0;
1022: tab_suffixof(code) = (char_type)code;
1023: }
1024: free_ent = ((block_compress) ? FIRST : 256 );
1025:
1026: finchar = oldcode = getcode();
1027: if(oldcode == -1) /* EOF already? */
1028: return; /* Get out of here */
1029: putchar( (char)finchar ); /* first code must be 8 bits = char */
1030: if(ferror(stdout)) /* Crash if can't write */
1031: writeerr();
1032: stackp = de_stack;
1033:
1034: while ( (code = getcode()) > -1 ) {
1035:
1036: if ( (code == CLEAR) && block_compress ) {
1037: for ( code = 255; code >= 0; code-- )
1038: tab_prefixof(code) = 0;
1039: clear_flg = 1;
1040: free_ent = FIRST - 1;
1041: if ( (code = getcode ()) == -1 ) /* O, untimely death! */
1042: break;
1043: }
1044: incode = code;
1045: /*
1046: * Special case for KwKwK string.
1047: */
1048: if ( code >= free_ent ) {
1049: *stackp++ = finchar;
1050: code = oldcode;
1051: }
1052:
1053: /*
1054: * Generate output characters in reverse order
1055: */
1056: #ifdef SIGNED_COMPARE_SLOW
1057: while ( ((unsigned long)code) >= ((unsigned long)256) ) {
1058: #else
1059: while ( code >= 256 ) {
1060: #endif
1061: *stackp++ = tab_suffixof(code);
1062: code = tab_prefixof(code);
1063: }
1064: *stackp++ = finchar = tab_suffixof(code);
1065:
1066: /*
1067: * And put them out in forward order
1068: */
1069: do
1070: putchar ( *--stackp );
1071: while ( stackp > de_stack );
1072:
1073: /*
1074: * Generate the new entry.
1075: */
1076: if ( (code=free_ent) < maxmaxcode ) {
1077: tab_prefixof(code) = (unsigned short)oldcode;
1078: tab_suffixof(code) = finchar;
1079: free_ent = code+1;
1080: }
1081: /*
1082: * Remember previous code.
1083: */
1084: oldcode = incode;
1085: }
1086: fflush( stdout );
1087: if(ferror(stdout))
1088: writeerr();
1089: }
1090:
1091: /*****************************************************************
1092: * TAG( getcode )
1093: *
1094: * Read one code from the standard input. If EOF, return -1.
1095: * Inputs:
1096: * stdin
1097: * Outputs:
1098: * code or -1 is returned.
1099: */
1100:
1101: code_int
1102: getcode() {
1103: /*
1104: * On the VAX, it is important to have the register declarations
1105: * in exactly the order given, or the asm will break.
1106: */
1107: register code_int code;
1108: static int offset = 0, size = 0;
1109: static char_type buf[BITS];
1110: register int r_off, bits;
1111: register char_type *bp = buf;
1112:
1113: if ( clear_flg > 0 || offset >= size || free_ent > maxcode ) {
1114: /*
1115: * If the next entry will be too big for the current code
1116: * size, then we must increase the size. This implies reading
1117: * a new buffer full, too.
1118: */
1119: if ( free_ent > maxcode ) {
1120: n_bits++;
1121: if ( n_bits == maxbits )
1122: maxcode = maxmaxcode; /* won't get any bigger now */
1123: else
1124: maxcode = MAXCODE(n_bits);
1125: }
1126: if ( clear_flg > 0) {
1127: maxcode = MAXCODE (n_bits = INIT_BITS);
1128: clear_flg = 0;
1129: }
1130: size = fread( buf, 1, n_bits, stdin );
1131: if ( size <= 0 )
1132: return -1; /* end of file */
1133: offset = 0;
1134: /* Round size down to integral number of codes */
1135: size = (size << 3) - (n_bits - 1);
1136: }
1137: r_off = offset;
1138: bits = n_bits;
1139: #ifdef vax
1140: asm( "extzv r10,r9,(r8),r11" );
1141: #else /* not a vax */
1142: /*
1143: * Get to the first byte.
1144: */
1145: bp += (r_off >> 3);
1146: r_off &= 7;
1147: /* Get first part (low order bits) */
1148: #ifdef NO_UCHAR
1149: code = ((*bp++ >> r_off) & rmask[8 - r_off]) & 0xff;
1150: #else
1151: code = (*bp++ >> r_off);
1152: #endif /* NO_UCHAR */
1153: bits -= (8 - r_off);
1154: r_off = 8 - r_off; /* now, offset into code word */
1155: /* Get any 8 bit parts in the middle (<=1 for up to 16 bits). */
1156: if ( bits >= 8 ) {
1157: #ifdef NO_UCHAR
1158: code |= (*bp++ & 0xff) << r_off;
1159: #else
1160: code |= *bp++ << r_off;
1161: #endif /* NO_UCHAR */
1162: r_off += 8;
1163: bits -= 8;
1164: }
1165: /* high order bits. */
1166: code |= (*bp & rmask[bits]) << r_off;
1167: #endif /* vax */
1168: offset += n_bits;
1169:
1170: return code;
1171: }
1172:
1173: char *
1174: rindex(s, c) /* For those who don't have it in libc.a */
1175: register char *s, c;
1176: {
1177: char *p;
1178: for (p = NULL; *s; s++)
1179: if (*s == c)
1180: p = s;
1181: return(p);
1182: }
1183:
1184: #ifdef DEBUG
1185: printcodes()
1186: {
1187: /*
1188: * Just print out codes from input file. For debugging.
1189: */
1190: code_int code;
1191: int col = 0, bits;
1192:
1193: bits = n_bits = INIT_BITS;
1194: maxcode = MAXCODE(n_bits);
1195: free_ent = ((block_compress) ? FIRST : 256 );
1196: while ( ( code = getcode() ) >= 0 ) {
1197: if ( (code == CLEAR) && block_compress ) {
1198: free_ent = FIRST - 1;
1199: clear_flg = 1;
1200: }
1201: else if ( free_ent < maxmaxcode )
1202: free_ent++;
1203: if ( bits != n_bits ) {
1204: fprintf(stderr, "\nChange to %d bits\n", n_bits );
1205: bits = n_bits;
1206: col = 0;
1207: }
1208: fprintf(stderr, "%5d%c", code, (col+=6) >= 74 ? (col = 0, '\n') : ' ' );
1209: }
1210: putc( '\n', stderr );
1211: exit( 0 );
1212: }
1213:
1214: code_int sorttab[1<<BITS]; /* sorted pointers into htab */
1215:
1216: dump_tab() /* dump string table */
1217: {
1218: register int i, first;
1219: register ent;
1220: #define STACK_SIZE 15000
1221: int stack_top = STACK_SIZE;
1222: register c;
1223:
1224: if(do_decomp == 0) { /* compressing */
1225: register int flag = 1;
1226:
1227: for(i=0; i<hsize; i++) { /* build sort pointers */
1228: if((long)htabof(i) >= 0) {
1229: sorttab[codetabof(i)] = i;
1230: }
1231: }
1232: first = block_compress ? FIRST : 256;
1233: for(i = first; i < free_ent; i++) {
1234: fprintf(stderr, "%5d: \"", i);
1235: de_stack[--stack_top] = '\n';
1236: de_stack[--stack_top] = '"';
1237: stack_top = in_stack((htabof(sorttab[i])>>maxbits)&0xff,
1238: stack_top);
1239: for(ent=htabof(sorttab[i]) & ((1<<maxbits)-1);
1240: ent > 256;
1241: ent=htabof(sorttab[ent]) & ((1<<maxbits)-1)) {
1242: stack_top = in_stack(htabof(sorttab[ent]) >> maxbits,
1243: stack_top);
1244: }
1245: stack_top = in_stack(ent, stack_top);
1246: fwrite( &de_stack[stack_top], 1, STACK_SIZE-stack_top, stderr);
1247: stack_top = STACK_SIZE;
1248: }
1249: } else if(!debug) { /* decompressing */
1250:
1251: for ( i = 0; i < free_ent; i++ ) {
1252: ent = i;
1253: c = tab_suffixof(ent);
1254: if ( isascii(c) && isprint(c) )
1255: fprintf( stderr, "%5d: %5d/'%c' \"",
1256: ent, tab_prefixof(ent), c );
1257: else
1258: fprintf( stderr, "%5d: %5d/\\%03o \"",
1259: ent, tab_prefixof(ent), c );
1260: de_stack[--stack_top] = '\n';
1261: de_stack[--stack_top] = '"';
1262: for ( ; ent != NULL;
1263: ent = (ent >= FIRST ? tab_prefixof(ent) : NULL) ) {
1264: stack_top = in_stack(tab_suffixof(ent), stack_top);
1265: }
1266: fwrite( &de_stack[stack_top], 1, STACK_SIZE - stack_top, stderr );
1267: stack_top = STACK_SIZE;
1268: }
1269: }
1270: }
1271:
1272: int
1273: in_stack(c, stack_top)
1274: register c, stack_top;
1275: {
1276: if ( (isascii(c) && isprint(c) && c != '\\') || c == ' ' ) {
1277: de_stack[--stack_top] = c;
1278: } else {
1279: switch( c ) {
1280: case '\n': de_stack[--stack_top] = 'n'; break;
1281: case '\t': de_stack[--stack_top] = 't'; break;
1282: case '\b': de_stack[--stack_top] = 'b'; break;
1283: case '\f': de_stack[--stack_top] = 'f'; break;
1284: case '\r': de_stack[--stack_top] = 'r'; break;
1285: case '\\': de_stack[--stack_top] = '\\'; break;
1286: default:
1287: de_stack[--stack_top] = '0' + c % 8;
1288: de_stack[--stack_top] = '0' + (c / 8) % 8;
1289: de_stack[--stack_top] = '0' + c / 64;
1290: break;
1291: }
1292: de_stack[--stack_top] = '\\';
1293: }
1294: return stack_top;
1295: }
1296: #endif /* DEBUG */
1297:
1298: writeerr()
1299: {
1300: perror ( ofname );
1301: unlink ( ofname );
1302: exit ( 1 );
1303: }
1304:
1305: copystat(ifname, ofname)
1306: char *ifname, *ofname;
1307: {
1308: struct stat statbuf;
1309: int mode;
1310: time_t timep[2];
1311:
1312: fclose(stdout);
1313: if (stat(ifname, &statbuf)) { /* Get stat on input file */
1314: perror(ifname);
1315: return;
1316: }
1317: if ((statbuf.st_mode & S_IFMT/*0170000*/) != S_IFREG/*0100000*/) {
1318: if(quiet)
1319: fprintf(stderr, "%s: ", ifname);
1320: fprintf(stderr, " -- not a regular file: unchanged");
1321: exit_stat = 1;
1322: perm_stat = 1;
1323: } else if (statbuf.st_nlink > 1) {
1324: if(quiet)
1325: fprintf(stderr, "%s: ", ifname);
1326: fprintf(stderr, " -- has %d other links: unchanged",
1327: statbuf.st_nlink - 1);
1328: exit_stat = 1;
1329: perm_stat = 1;
1330: } else if (exit_stat == 2 && (!force)) { /* No compression: remove file.Z */
1331: if(!quiet)
1332: fprintf(stderr, " -- file unchanged");
1333: } else { /* ***** Successful Compression ***** */
1334: exit_stat = 0;
1335: mode = statbuf.st_mode & 07777;
1336: if (chmod(ofname, mode)) /* Copy modes */
1337: perror(ofname);
1338: chown(ofname, statbuf.st_uid, statbuf.st_gid); /* Copy ownership */
1339: timep[0] = statbuf.st_atime;
1340: timep[1] = statbuf.st_mtime;
1341: utime(ofname, timep); /* Update last accessed and modified times */
1342: if (unlink(ifname)) /* Remove input file */
1343: perror(ifname);
1344: if(!quiet)
1345: fprintf(stderr, " -- replaced with %s", ofname);
1346: return; /* Successful return */
1347: }
1348:
1349: /* Unsuccessful return -- one of the tests failed */
1350: if (unlink(ofname))
1351: perror(ofname);
1352: }
1353:
1354: onintr ( )
1355: {
1356: if (!precious)
1357: unlink ( ofname );
1358: exit ( 1 );
1359: }
1360:
1361: oops ( ) /* wild pointer -- assume bad input */
1362: {
1363: if ( do_decomp )
1364: fprintf ( stderr, "uncompress: corrupt input\n" );
1365: unlink ( ofname );
1366: exit ( 1 );
1367: }
1368:
1369: cl_block () /* table clear for block compress */
1370: {
1371: register long int rat;
1372:
1373: checkpoint = in_count + CHECK_GAP;
1374: #ifdef DEBUG
1375: if ( debug ) {
1376: fprintf ( stderr, "count: %ld, ratio: ", in_count );
1377: prratio ( stderr, in_count, bytes_out );
1378: fprintf ( stderr, "\n");
1379: }
1380: #endif /* DEBUG */
1381:
1382: if(in_count > 0x007fffff) { /* shift will overflow */
1383: rat = bytes_out >> 8;
1384: if(rat == 0) { /* Don't divide by zero */
1385: rat = 0x7fffffff;
1386: } else {
1387: rat = in_count / rat;
1388: }
1389: } else {
1390: rat = (in_count << 8) / bytes_out; /* 8 fractional bits */
1391: }
1392: if ( rat > ratio ) {
1393: ratio = rat;
1394: } else {
1395: ratio = 0;
1396: #ifdef DEBUG
1397: if(verbose)
1398: dump_tab(); /* dump string table */
1399: #endif
1400: cl_hash ( (count_int) hsize );
1401: free_ent = FIRST;
1402: clear_flg = 1;
1403: output ( (code_int) CLEAR );
1404: #ifdef DEBUG
1405: if(debug)
1406: fprintf ( stderr, "clear\n" );
1407: #endif /* DEBUG */
1408: }
1409: }
1410:
1411: cl_hash(hsize) /* reset code table */
1412: register count_int hsize;
1413: {
1414: #ifndef XENIX_16 /* Normal machine */
1415: register count_int *htab_p = htab+hsize;
1416: #else
1417: register j;
1418: register long k = hsize;
1419: register count_int *htab_p;
1420: #endif
1421: register long i;
1422: register long m1 = -1;
1423:
1424: #ifdef XENIX_16
1425: for(j=0; j<=8 && k>=0; j++,k-=8192) {
1426: i = 8192;
1427: if(k < 8192) {
1428: i = k;
1429: }
1430: htab_p = &(htab[j][i]);
1431: i -= 16;
1432: if(i > 0) {
1433: #else
1434: i = hsize - 16;
1435: #endif
1436: do { /* might use Sys V memset(3) here */
1437: *(htab_p-16) = m1;
1438: *(htab_p-15) = m1;
1439: *(htab_p-14) = m1;
1440: *(htab_p-13) = m1;
1441: *(htab_p-12) = m1;
1442: *(htab_p-11) = m1;
1443: *(htab_p-10) = m1;
1444: *(htab_p-9) = m1;
1445: *(htab_p-8) = m1;
1446: *(htab_p-7) = m1;
1447: *(htab_p-6) = m1;
1448: *(htab_p-5) = m1;
1449: *(htab_p-4) = m1;
1450: *(htab_p-3) = m1;
1451: *(htab_p-2) = m1;
1452: *(htab_p-1) = m1;
1453: htab_p -= 16;
1454: } while ((i -= 16) >= 0);
1455: #ifdef XENIX_16
1456: }
1457: }
1458: #endif
1459: for ( i += 16; i > 0; i-- )
1460: *--htab_p = m1;
1461: }
1462:
1463: prratio(stream, num, den)
1464: FILE *stream;
1465: long int num, den;
1466: {
1467: register int q; /* Doesn't need to be long */
1468:
1469: if(num > 214748L) { /* 2147483647/10000 */
1470: q = num / (den / 10000L);
1471: } else {
1472: q = 10000L * num / den; /* Long calculations, though */
1473: }
1474: if (q < 0) {
1475: putc('-', stream);
1476: q = -q;
1477: }
1478: fprintf(stream, "%d.%02d%%", q / 100, q % 100);
1479: }
1480:
1481: version()
1482: {
1483: fprintf(stderr, "%s, Berkeley 5.9 5/11/86\n", rcs_ident);
1484: fprintf(stderr, "Options: ");
1485: #ifdef vax
1486: fprintf(stderr, "vax, ");
1487: #endif
1488: #ifdef NO_UCHAR
1489: fprintf(stderr, "NO_UCHAR, ");
1490: #endif
1491: #ifdef SIGNED_COMPARE_SLOW
1492: fprintf(stderr, "SIGNED_COMPARE_SLOW, ");
1493: #endif
1494: #ifdef XENIX_16
1495: fprintf(stderr, "XENIX_16, ");
1496: #endif
1497: #ifdef COMPATIBLE
1498: fprintf(stderr, "COMPATIBLE, ");
1499: #endif
1500: #ifdef DEBUG
1501: fprintf(stderr, "DEBUG, ");
1502: #endif
1503: #ifdef BSD4_2
1504: fprintf(stderr, "BSD4_2, ");
1505: #endif
1506: fprintf(stderr, "BITS = %d\n", BITS);
1507: }
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.