Coverage Report

Created: 2025-12-31 07:53

next uncovered line (L), next uncovered region (R), next uncovered branch (B)
/src/bzip2/blocksort.c
Line
Count
Source
1
2
/*-------------------------------------------------------------*/
3
/*--- Block sorting machinery                               ---*/
4
/*---                                           blocksort.c ---*/
5
/*-------------------------------------------------------------*/
6
7
/* ------------------------------------------------------------------
8
   This file is part of bzip2/libbzip2, a program and library for
9
   lossless, block-sorting data compression.
10
11
   bzip2/libbzip2 version 1.0.6 of 6 September 2010
12
   Copyright (C) 1996-2010 Julian Seward <jseward@acm.org>
13
14
   Please read the WARNING, DISCLAIMER and PATENTS sections in the 
15
   README file.
16
17
   This program is released under the terms of the license contained
18
   in the file LICENSE.
19
   ------------------------------------------------------------------ */
20
21
22
#include "bzlib_private.h"
23
24
/*---------------------------------------------*/
25
/*--- Fallback O(N log(N)^2) sorting        ---*/
26
/*--- algorithm, for repetitive blocks      ---*/
27
/*---------------------------------------------*/
28
29
/*---------------------------------------------*/
30
static 
31
__inline__
32
void fallbackSimpleSort ( UInt32* fmap, 
33
                          UInt32* eclass, 
34
                          Int32   lo, 
35
                          Int32   hi )
36
101M
{
37
101M
   Int32 i, j, tmp;
38
101M
   UInt32 ec_tmp;
39
40
101M
   if (lo == hi) return;
41
42
98.6M
   if (hi - lo > 3) {
43
86.8M
      for ( i = hi-4; i >= lo; i-- ) {
44
65.8M
         tmp = fmap[i];
45
65.8M
         ec_tmp = eclass[tmp];
46
87.1M
         for ( j = i+4; j <= hi && ec_tmp > eclass[fmap[j]]; j += 4 )
47
21.2M
            fmap[j-4] = fmap[j];
48
65.8M
         fmap[j-4] = tmp;
49
65.8M
      }
50
21.0M
   }
51
52
328M
   for ( i = hi-1; i >= lo; i-- ) {
53
230M
      tmp = fmap[i];
54
230M
      ec_tmp = eclass[tmp];
55
333M
      for ( j = i+1; j <= hi && ec_tmp > eclass[fmap[j]]; j++ )
56
103M
         fmap[j-1] = fmap[j];
57
230M
      fmap[j-1] = tmp;
58
230M
   }
59
98.6M
}
60
61
62
/*---------------------------------------------*/
63
#define fswap(zz1, zz2) \
64
1.19G
   { Int32 zztmp = zz1; zz1 = zz2; zz2 = zztmp; }
65
66
41.4M
#define fvswap(zzp1, zzp2, zzn)       \
67
41.4M
{                                     \
68
41.4M
   Int32 yyp1 = (zzp1);               \
69
41.4M
   Int32 yyp2 = (zzp2);               \
70
41.4M
   Int32 yyn  = (zzn);                \
71
221M
   while (yyn > 0) {                  \
72
180M
      fswap(fmap[yyp1], fmap[yyp2]);  \
73
180M
      yyp1++; yyp2++; yyn--;          \
74
180M
   }                                  \
75
41.4M
}
76
77
78
41.4M
#define fmin(a,b) ((a) < (b)) ? (a) : (b)
79
80
123M
#define fpush(lz,hz) { stackLo[sp] = lz; \
81
123M
                       stackHi[sp] = hz; \
82
123M
                       sp++; }
83
84
123M
#define fpop(lz,hz) { sp--;              \
85
123M
                      lz = stackLo[sp];  \
86
123M
                      hz = stackHi[sp]; }
87
88
123M
#define FALLBACK_QSORT_SMALL_THRESH 10
89
#define FALLBACK_QSORT_STACK_SIZE   100
90
91
92
static
93
void fallbackQSort3 ( UInt32* fmap, 
94
                      UInt32* eclass,
95
                      Int32   loSt, 
96
                      Int32   hiSt )
97
82.0M
{
98
82.0M
   Int32 unLo, unHi, ltLo, gtHi, n, m;
99
82.0M
   Int32 sp, lo, hi;
100
82.0M
   UInt32 med, r, r3;
101
82.0M
   Int32 stackLo[FALLBACK_QSORT_STACK_SIZE];
102
82.0M
   Int32 stackHi[FALLBACK_QSORT_STACK_SIZE];
103
104
82.0M
   r = 0;
105
106
82.0M
   sp = 0;
107
82.0M
   fpush ( loSt, hiSt );
108
109
205M
   while (sp > 0) {
110
111
123M
      AssertH ( sp < FALLBACK_QSORT_STACK_SIZE - 1, 1004 );
112
113
123M
      fpop ( lo, hi );
114
123M
      if (hi - lo < FALLBACK_QSORT_SMALL_THRESH) {
115
101M
         fallbackSimpleSort ( fmap, eclass, lo, hi );
116
101M
         continue;
117
101M
      }
118
119
      /* Random partitioning.  Median of 3 sometimes fails to
120
         avoid bad cases.  Median of 9 seems to help but 
121
         looks rather expensive.  This too seems to work but
122
         is cheaper.  Guidance for the magic constants 
123
         7621 and 32768 is taken from Sedgewick's algorithms
124
         book, chapter 35.
125
      */
126
22.1M
      r = ((r * 7621) + 1) % 32768;
127
22.1M
      r3 = r % 3;
128
22.1M
      if (r3 == 0) med = eclass[fmap[lo]]; else
129
19.6M
      if (r3 == 1) med = eclass[fmap[(lo+hi)>>1]]; else
130
7.53M
                   med = eclass[fmap[hi]];
131
132
22.1M
      unLo = ltLo = lo;
133
22.1M
      unHi = gtHi = hi;
134
135
113M
      while (1) {
136
1.15G
         while (1) {
137
1.15G
            if (unLo > unHi) break;
138
1.14G
            n = (Int32)eclass[fmap[unLo]] - (Int32)med;
139
1.14G
            if (n == 0) { 
140
742M
               fswap(fmap[unLo], fmap[ltLo]); 
141
742M
               ltLo++; unLo++; 
142
742M
               continue; 
143
742M
            };
144
403M
            if (n > 0) break;
145
302M
            unLo++;
146
302M
         }
147
604M
         while (1) {
148
604M
            if (unLo > unHi) break;
149
581M
            n = (Int32)eclass[fmap[unHi]] - (Int32)med;
150
581M
            if (n == 0) { 
151
184M
               fswap(fmap[unHi], fmap[gtHi]); 
152
184M
               gtHi--; unHi--; 
153
184M
               continue; 
154
397M
            };
155
397M
            if (n < 0) break;
156
305M
            unHi--;
157
305M
         }
158
113M
         if (unLo > unHi) break;
159
91.7M
         fswap(fmap[unLo], fmap[unHi]); unLo++; unHi--;
160
91.7M
      }
161
162
22.1M
      AssertD ( unHi == unLo-1, "fallbackQSort3(2)" );
163
164
22.1M
      if (gtHi < ltLo) continue;
165
166
20.7M
      n = fmin(ltLo-lo, unLo-ltLo); fvswap(lo, unLo-n, n);
167
20.7M
      m = fmin(hi-gtHi, gtHi-unHi); fvswap(unLo, hi-m+1, m);
168
169
20.7M
      n = lo + unLo - ltLo - 1;
170
20.7M
      m = hi - (gtHi - unHi) + 1;
171
172
20.7M
      if (n - lo > hi - m) {
173
10.3M
         fpush ( lo, n );
174
10.3M
         fpush ( m, hi );
175
10.3M
      } else {
176
10.3M
         fpush ( m, hi );
177
10.3M
         fpush ( lo, n );
178
10.3M
      }
179
20.7M
   }
180
82.0M
}
181
182
#undef fmin
183
#undef fpush
184
#undef fpop
185
#undef fswap
186
#undef fvswap
187
#undef FALLBACK_QSORT_SMALL_THRESH
188
#undef FALLBACK_QSORT_STACK_SIZE
189
190
191
/*---------------------------------------------*/
192
/* Pre:
193
      nblock > 0
194
      eclass exists for [0 .. nblock-1]
195
      ((UChar*)eclass) [0 .. nblock-1] holds block
196
      ptr exists for [0 .. nblock-1]
197
198
   Post:
199
      ((UChar*)eclass) [0 .. nblock-1] holds block
200
      All other areas of eclass destroyed
201
      fmap [0 .. nblock-1] holds sorted order
202
      bhtab [ 0 .. 2+(nblock/32) ] destroyed
203
*/
204
205
245M
#define       SET_BH(zz)  bhtab[(zz) >> 5] |= ((UInt32)1 << ((zz) & 31))
206
2.71M
#define     CLEAR_BH(zz)  bhtab[(zz) >> 5] &= ~((UInt32)1 << ((zz) & 31))
207
3.21G
#define     ISSET_BH(zz)  (bhtab[(zz) >> 5] & ((UInt32)1 << ((zz) & 31)))
208
53.4M
#define      WORD_BH(zz)  bhtab[(zz) >> 5]
209
458M
#define UNALIGNED_BH(zz)  ((zz) & 0x01f)
210
211
static
212
void fallbackSort ( UInt32* fmap, 
213
                    UInt32* eclass, 
214
                    UInt32* bhtab,
215
                    Int32   nblock,
216
                    Int32   verb )
217
84.7k
{
218
84.7k
   Int32 ftab[257];
219
84.7k
   Int32 ftabCopy[256];
220
84.7k
   Int32 H, i, j, k, l, r, cc, cc1;
221
84.7k
   Int32 nNotDone;
222
84.7k
   Int32 nBhtab;
223
84.7k
   UChar* eclass8 = (UChar*)eclass;
224
225
   /*--
226
      Initial 1-char radix sort to generate
227
      initial fmap and initial BH bits.
228
   --*/
229
84.7k
   if (verb >= 4)
230
0
      VPrintf0 ( "        bucket sorting ...\n" );
231
21.8M
   for (i = 0; i < 257;    i++) ftab[i] = 0;
232
146M
   for (i = 0; i < nblock; i++) ftab[eclass8[i]]++;
233
21.7M
   for (i = 0; i < 256;    i++) ftabCopy[i] = ftab[i];
234
21.7M
   for (i = 1; i < 257;    i++) ftab[i] += ftab[i-1];
235
236
146M
   for (i = 0; i < nblock; i++) {
237
146M
      j = eclass8[i];
238
146M
      k = ftab[j] - 1;
239
146M
      ftab[j] = k;
240
146M
      fmap[k] = i;
241
146M
   }
242
243
84.7k
   nBhtab = 2 + (nblock / 32);
244
4.80M
   for (i = 0; i < nBhtab; i++) bhtab[i] = 0;
245
21.7M
   for (i = 0; i < 256; i++) SET_BH(ftab[i]);
246
247
   /*--
248
      Inductively refine the buckets.  Kind-of an
249
      "exponential radix sort" (!), inspired by the
250
      Manber-Myers suffix array construction algorithm.
251
   --*/
252
253
   /*-- set sentinel bits for block-end detection --*/
254
2.79M
   for (i = 0; i < 32; i++) { 
255
2.71M
      SET_BH(nblock + 2*i);
256
2.71M
      CLEAR_BH(nblock + 2*i + 1);
257
2.71M
   }
258
259
   /*-- the log(N) loop --*/
260
84.7k
   H = 1;
261
515k
   while (1) {
262
263
515k
      if (verb >= 4) 
264
0
         VPrintf1 ( "        depth %6d has ", H );
265
266
515k
      j = 0;
267
1.69G
      for (i = 0; i < nblock; i++) {
268
1.69G
         if (ISSET_BH(i)) j = i;
269
1.69G
         k = fmap[i] - H; if (k < 0) k += nblock;
270
1.69G
         eclass[k] = j;
271
1.69G
      }
272
273
515k
      nNotDone = 0;
274
515k
      r = -1;
275
82.5M
      while (1) {
276
277
   /*-- find the next non-singleton bucket --*/
278
82.5M
         k = r + 1;
279
296M
         while (ISSET_BH(k) && UNALIGNED_BH(k)) k++;
280
82.5M
         if (ISSET_BH(k)) {
281
16.8M
            while (WORD_BH(k) == 0xffffffff) k += 32;
282
63.7M
            while (ISSET_BH(k)) k++;
283
8.86M
         }
284
82.5M
         l = k - 1;
285
82.5M
         if (l >= nblock) break;
286
308M
         while (!ISSET_BH(k) && UNALIGNED_BH(k)) k++;
287
82.0M
         if (!ISSET_BH(k)) {
288
36.5M
            while (WORD_BH(k) == 0x00000000) k += 32;
289
81.7M
            while (!ISSET_BH(k)) k++;
290
9.28M
         }
291
82.0M
         r = k - 1;
292
82.0M
         if (r >= nblock) break;
293
294
         /*-- now [l, r] bracket current bucket --*/
295
82.0M
         if (r > l) {
296
82.0M
            nNotDone += (r - l + 1);
297
82.0M
            fallbackQSort3 ( fmap, eclass, l, r );
298
299
            /*-- scan bucket and generate header bits-- */
300
82.0M
            cc = -1;
301
1.33G
            for (i = l; i <= r; i++) {
302
1.25G
               cc1 = eclass[fmap[i]];
303
1.25G
               if (cc != cc1) { SET_BH(i); cc = cc1; };
304
1.25G
            }
305
82.0M
         }
306
82.0M
      }
307
308
515k
      if (verb >= 4) 
309
0
         VPrintf1 ( "%6d unresolved strings\n", nNotDone );
310
311
515k
      H *= 2;
312
515k
      if (H > nblock || nNotDone == 0) break;
313
515k
   }
314
315
   /*-- 
316
      Reconstruct the original block in
317
      eclass8 [0 .. nblock-1], since the
318
      previous phase destroyed it.
319
   --*/
320
84.7k
   if (verb >= 4)
321
0
      VPrintf0 ( "        reconstructing block ...\n" );
322
84.7k
   j = 0;
323
146M
   for (i = 0; i < nblock; i++) {
324
161M
      while (ftabCopy[j] == 0) j++;
325
146M
      ftabCopy[j]--;
326
146M
      eclass8[fmap[i]] = (UChar)j;
327
146M
   }
328
84.7k
   AssertH ( j < 256, 1005 );
329
84.7k
}
330
331
#undef       SET_BH
332
#undef     CLEAR_BH
333
#undef     ISSET_BH
334
#undef      WORD_BH
335
#undef UNALIGNED_BH
336
337
338
/*---------------------------------------------*/
339
/*--- The main, O(N^2 log(N)) sorting       ---*/
340
/*--- algorithm.  Faster for "normal"       ---*/
341
/*--- non-repetitive blocks.                ---*/
342
/*---------------------------------------------*/
343
344
/*---------------------------------------------*/
345
static
346
__inline__
347
Bool mainGtU ( UInt32  i1, 
348
               UInt32  i2,
349
               UChar*  block, 
350
               UInt16* quadrant,
351
               UInt32  nblock,
352
               Int32*  budget )
353
101M
{
354
101M
   Int32  k;
355
101M
   UChar  c1, c2;
356
101M
   UInt16 s1, s2;
357
358
101M
   AssertD ( i1 != i2, "mainGtU" );
359
   /* 1 */
360
101M
   c1 = block[i1]; c2 = block[i2];
361
101M
   if (c1 != c2) return (c1 > c2);
362
98.7M
   i1++; i2++;
363
   /* 2 */
364
98.7M
   c1 = block[i1]; c2 = block[i2];
365
98.7M
   if (c1 != c2) return (c1 > c2);
366
96.8M
   i1++; i2++;
367
   /* 3 */
368
96.8M
   c1 = block[i1]; c2 = block[i2];
369
96.8M
   if (c1 != c2) return (c1 > c2);
370
95.6M
   i1++; i2++;
371
   /* 4 */
372
95.6M
   c1 = block[i1]; c2 = block[i2];
373
95.6M
   if (c1 != c2) return (c1 > c2);
374
94.2M
   i1++; i2++;
375
   /* 5 */
376
94.2M
   c1 = block[i1]; c2 = block[i2];
377
94.2M
   if (c1 != c2) return (c1 > c2);
378
92.9M
   i1++; i2++;
379
   /* 6 */
380
92.9M
   c1 = block[i1]; c2 = block[i2];
381
92.9M
   if (c1 != c2) return (c1 > c2);
382
91.5M
   i1++; i2++;
383
   /* 7 */
384
91.5M
   c1 = block[i1]; c2 = block[i2];
385
91.5M
   if (c1 != c2) return (c1 > c2);
386
90.1M
   i1++; i2++;
387
   /* 8 */
388
90.1M
   c1 = block[i1]; c2 = block[i2];
389
90.1M
   if (c1 != c2) return (c1 > c2);
390
89.0M
   i1++; i2++;
391
   /* 9 */
392
89.0M
   c1 = block[i1]; c2 = block[i2];
393
89.0M
   if (c1 != c2) return (c1 > c2);
394
87.9M
   i1++; i2++;
395
   /* 10 */
396
87.9M
   c1 = block[i1]; c2 = block[i2];
397
87.9M
   if (c1 != c2) return (c1 > c2);
398
86.8M
   i1++; i2++;
399
   /* 11 */
400
86.8M
   c1 = block[i1]; c2 = block[i2];
401
86.8M
   if (c1 != c2) return (c1 > c2);
402
85.7M
   i1++; i2++;
403
   /* 12 */
404
85.7M
   c1 = block[i1]; c2 = block[i2];
405
85.7M
   if (c1 != c2) return (c1 > c2);
406
84.8M
   i1++; i2++;
407
408
84.8M
   k = nblock + 8;
409
410
767M
   do {
411
      /* 1 */
412
767M
      c1 = block[i1]; c2 = block[i2];
413
767M
      if (c1 != c2) return (c1 > c2);
414
763M
      s1 = quadrant[i1]; s2 = quadrant[i2];
415
763M
      if (s1 != s2) return (s1 > s2);
416
752M
      i1++; i2++;
417
      /* 2 */
418
752M
      c1 = block[i1]; c2 = block[i2];
419
752M
      if (c1 != c2) return (c1 > c2);
420
748M
      s1 = quadrant[i1]; s2 = quadrant[i2];
421
748M
      if (s1 != s2) return (s1 > s2);
422
735M
      i1++; i2++;
423
      /* 3 */
424
735M
      c1 = block[i1]; c2 = block[i2];
425
735M
      if (c1 != c2) return (c1 > c2);
426
732M
      s1 = quadrant[i1]; s2 = quadrant[i2];
427
732M
      if (s1 != s2) return (s1 > s2);
428
727M
      i1++; i2++;
429
      /* 4 */
430
727M
      c1 = block[i1]; c2 = block[i2];
431
727M
      if (c1 != c2) return (c1 > c2);
432
723M
      s1 = quadrant[i1]; s2 = quadrant[i2];
433
723M
      if (s1 != s2) return (s1 > s2);
434
718M
      i1++; i2++;
435
      /* 5 */
436
718M
      c1 = block[i1]; c2 = block[i2];
437
718M
      if (c1 != c2) return (c1 > c2);
438
715M
      s1 = quadrant[i1]; s2 = quadrant[i2];
439
715M
      if (s1 != s2) return (s1 > s2);
440
710M
      i1++; i2++;
441
      /* 6 */
442
710M
      c1 = block[i1]; c2 = block[i2];
443
710M
      if (c1 != c2) return (c1 > c2);
444
706M
      s1 = quadrant[i1]; s2 = quadrant[i2];
445
706M
      if (s1 != s2) return (s1 > s2);
446
699M
      i1++; i2++;
447
      /* 7 */
448
699M
      c1 = block[i1]; c2 = block[i2];
449
699M
      if (c1 != c2) return (c1 > c2);
450
696M
      s1 = quadrant[i1]; s2 = quadrant[i2];
451
696M
      if (s1 != s2) return (s1 > s2);
452
692M
      i1++; i2++;
453
      /* 8 */
454
692M
      c1 = block[i1]; c2 = block[i2];
455
692M
      if (c1 != c2) return (c1 > c2);
456
688M
      s1 = quadrant[i1]; s2 = quadrant[i2];
457
688M
      if (s1 != s2) return (s1 > s2);
458
682M
      i1++; i2++;
459
460
682M
      if (i1 >= nblock) i1 -= nblock;
461
682M
      if (i2 >= nblock) i2 -= nblock;
462
463
682M
      k -= 8;
464
682M
      (*budget)--;
465
682M
   }
466
682M
      while (k >= 0);
467
468
35.5k
   return False;
469
84.8M
}
470
471
472
/*---------------------------------------------*/
473
/*--
474
   Knuth's increments seem to work better
475
   than Incerpi-Sedgewick here.  Possibly
476
   because the number of elems to sort is
477
   usually small, typically <= 20.
478
--*/
479
static
480
Int32 incs[14] = { 1, 4, 13, 40, 121, 364, 1093, 3280,
481
                   9841, 29524, 88573, 265720,
482
                   797161, 2391484 };
483
484
static
485
void mainSimpleSort ( UInt32* ptr,
486
                      UChar*  block,
487
                      UInt16* quadrant,
488
                      Int32   nblock,
489
                      Int32   lo, 
490
                      Int32   hi, 
491
                      Int32   d,
492
                      Int32*  budget )
493
940k
{
494
940k
   Int32 i, j, h, bigN, hp;
495
940k
   UInt32 v;
496
497
940k
   bigN = hi - lo + 1;
498
940k
   if (bigN < 2) return;
499
500
703k
   hp = 0;
501
1.99M
   while (incs[hp] < bigN) hp++;
502
703k
   hp--;
503
504
1.97M
   for (; hp >= 0; hp--) {
505
1.27M
      h = incs[hp];
506
507
1.27M
      i = lo + h;
508
16.7M
      while (True) {
509
510
         /*-- copy 1 --*/
511
16.7M
         if (i > hi) break;
512
16.4M
         v = ptr[i];
513
16.4M
         j = i;
514
34.0M
         while ( mainGtU ( 
515
34.0M
                    ptr[j-h]+d, v+d, block, quadrant, nblock, budget 
516
34.0M
                 ) ) {
517
19.9M
            ptr[j] = ptr[j-h];
518
19.9M
            j = j - h;
519
19.9M
            if (j <= (lo + h - 1)) break;
520
19.9M
         }
521
16.4M
         ptr[j] = v;
522
16.4M
         i++;
523
524
         /*-- copy 2 --*/
525
16.4M
         if (i > hi) break;
526
15.8M
         v = ptr[i];
527
15.8M
         j = i;
528
33.7M
         while ( mainGtU ( 
529
33.7M
                    ptr[j-h]+d, v+d, block, quadrant, nblock, budget 
530
33.7M
                 ) ) {
531
19.9M
            ptr[j] = ptr[j-h];
532
19.9M
            j = j - h;
533
19.9M
            if (j <= (lo + h - 1)) break;
534
19.9M
         }
535
15.8M
         ptr[j] = v;
536
15.8M
         i++;
537
538
         /*-- copy 3 --*/
539
15.8M
         if (i > hi) break;
540
15.4M
         v = ptr[i];
541
15.4M
         j = i;
542
33.4M
         while ( mainGtU ( 
543
33.4M
                    ptr[j-h]+d, v+d, block, quadrant, nblock, budget 
544
33.4M
                 ) ) {
545
19.9M
            ptr[j] = ptr[j-h];
546
19.9M
            j = j - h;
547
19.9M
            if (j <= (lo + h - 1)) break;
548
19.9M
         }
549
15.4M
         ptr[j] = v;
550
15.4M
         i++;
551
552
15.4M
         if (*budget < 0) return;
553
15.4M
      }
554
1.27M
   }
555
703k
}
556
557
558
/*---------------------------------------------*/
559
/*--
560
   The following is an implementation of
561
   an elegant 3-way quicksort for strings,
562
   described in a paper "Fast Algorithms for
563
   Sorting and Searching Strings", by Robert
564
   Sedgewick and Jon L. Bentley.
565
--*/
566
567
#define mswap(zz1, zz2) \
568
215M
   { Int32 zztmp = zz1; zz1 = zz2; zz2 = zztmp; }
569
570
441k
#define mvswap(zzp1, zzp2, zzn)       \
571
441k
{                                     \
572
441k
   Int32 yyp1 = (zzp1);               \
573
441k
   Int32 yyp2 = (zzp2);               \
574
441k
   Int32 yyn  = (zzn);                \
575
9.07M
   while (yyn > 0) {                  \
576
8.62M
      mswap(ptr[yyp1], ptr[yyp2]);    \
577
8.62M
      yyp1++; yyp2++; yyn--;          \
578
8.62M
   }                                  \
579
441k
}
580
581
static 
582
__inline__
583
UChar mmed3 ( UChar a, UChar b, UChar c )
584
1.07M
{
585
1.07M
   UChar t;
586
1.07M
   if (a > b) { t = a; a = b; b = t; };
587
1.07M
   if (b > c) { 
588
43.6k
      b = c;
589
43.6k
      if (a > b) b = a;
590
43.6k
   }
591
1.07M
   return b;
592
1.07M
}
593
594
441k
#define mmin(a,b) ((a) < (b)) ? (a) : (b)
595
596
2.02M
#define mpush(lz,hz,dz) { stackLo[sp] = lz; \
597
2.02M
                          stackHi[sp] = hz; \
598
2.02M
                          stackD [sp] = dz; \
599
2.02M
                          sp++; }
600
601
2.02M
#define mpop(lz,hz,dz) { sp--;             \
602
2.02M
                         lz = stackLo[sp]; \
603
2.02M
                         hz = stackHi[sp]; \
604
2.02M
                         dz = stackD [sp]; }
605
606
607
1.32M
#define mnextsize(az) (nextHi[az]-nextLo[az])
608
609
#define mnextswap(az,bz)                                        \
610
471k
   { Int32 tz;                                                  \
611
471k
     tz = nextLo[az]; nextLo[az] = nextLo[bz]; nextLo[bz] = tz; \
612
471k
     tz = nextHi[az]; nextHi[az] = nextHi[bz]; nextHi[bz] = tz; \
613
471k
     tz = nextD [az]; nextD [az] = nextD [bz]; nextD [bz] = tz; }
614
615
616
4.04M
#define MAIN_QSORT_SMALL_THRESH 20
617
1.18M
#define MAIN_QSORT_DEPTH_THRESH (BZ_N_RADIX + BZ_N_QSORT)
618
#define MAIN_QSORT_STACK_SIZE 100
619
620
static
621
void mainQSort3 ( UInt32* ptr,
622
                  UChar*  block,
623
                  UInt16* quadrant,
624
                  Int32   nblock,
625
                  Int32   loSt, 
626
                  Int32   hiSt, 
627
                  Int32   dSt,
628
                  Int32*  budget )
629
500k
{
630
500k
   Int32 unLo, unHi, ltLo, gtHi, n, m, med;
631
500k
   Int32 sp, lo, hi, d;
632
633
500k
   Int32 stackLo[MAIN_QSORT_STACK_SIZE];
634
500k
   Int32 stackHi[MAIN_QSORT_STACK_SIZE];
635
500k
   Int32 stackD [MAIN_QSORT_STACK_SIZE];
636
637
500k
   Int32 nextLo[3];
638
500k
   Int32 nextHi[3];
639
500k
   Int32 nextD [3];
640
641
500k
   sp = 0;
642
500k
   mpush ( loSt, hiSt, dSt );
643
644
2.51M
   while (sp > 0) {
645
646
2.02M
      AssertH ( sp < MAIN_QSORT_STACK_SIZE - 2, 1001 );
647
648
2.02M
      mpop ( lo, hi, d );
649
2.02M
      if (hi - lo < MAIN_QSORT_SMALL_THRESH || 
650
1.18M
          d > MAIN_QSORT_DEPTH_THRESH) {
651
940k
         mainSimpleSort ( ptr, block, quadrant, nblock, lo, hi, d, budget );
652
940k
         if (*budget < 0) return;
653
935k
         continue;
654
940k
      }
655
656
1.07M
      med = (Int32) 
657
1.07M
            mmed3 ( block[ptr[ lo         ]+d],
658
1.07M
                    block[ptr[ hi         ]+d],
659
1.07M
                    block[ptr[ (lo+hi)>>1 ]+d] );
660
661
1.07M
      unLo = ltLo = lo;
662
1.07M
      unHi = gtHi = hi;
663
664
1.48M
      while (True) {
665
185M
         while (True) {
666
185M
            if (unLo > unHi) break;
667
184M
            n = ((Int32)block[ptr[unLo]+d]) - med;
668
184M
            if (n == 0) { 
669
178M
               mswap(ptr[unLo], ptr[ltLo]); 
670
178M
               ltLo++; unLo++; continue; 
671
178M
            };
672
5.76M
            if (n >  0) break;
673
5.27M
            unLo++;
674
5.27M
         }
675
34.4M
         while (True) {
676
34.4M
            if (unLo > unHi) break;
677
33.3M
            n = ((Int32)block[ptr[unHi]+d]) - med;
678
33.3M
            if (n == 0) { 
679
28.0M
               mswap(ptr[unHi], ptr[gtHi]); 
680
28.0M
               gtHi--; unHi--; continue; 
681
28.0M
            };
682
5.25M
            if (n <  0) break;
683
4.84M
            unHi--;
684
4.84M
         }
685
1.48M
         if (unLo > unHi) break;
686
408k
         mswap(ptr[unLo], ptr[unHi]); unLo++; unHi--;
687
408k
      }
688
689
1.07M
      AssertD ( unHi == unLo-1, "mainQSort3(2)" );
690
691
1.07M
      if (gtHi < ltLo) {
692
858k
         mpush(lo, hi, d+1 );
693
858k
         continue;
694
858k
      }
695
696
220k
      n = mmin(ltLo-lo, unLo-ltLo); mvswap(lo, unLo-n, n);
697
220k
      m = mmin(hi-gtHi, gtHi-unHi); mvswap(unLo, hi-m+1, m);
698
699
220k
      n = lo + unLo - ltLo - 1;
700
220k
      m = hi - (gtHi - unHi) + 1;
701
702
220k
      nextLo[0] = lo;  nextHi[0] = n;   nextD[0] = d;
703
220k
      nextLo[1] = m;   nextHi[1] = hi;  nextD[1] = d;
704
220k
      nextLo[2] = n+1; nextHi[2] = m-1; nextD[2] = d+1;
705
706
220k
      if (mnextsize(0) < mnextsize(1)) mnextswap(0,1);
707
220k
      if (mnextsize(1) < mnextsize(2)) mnextswap(1,2);
708
220k
      if (mnextsize(0) < mnextsize(1)) mnextswap(0,1);
709
710
220k
      AssertD (mnextsize(0) >= mnextsize(1), "mainQSort3(8)" );
711
220k
      AssertD (mnextsize(1) >= mnextsize(2), "mainQSort3(9)" );
712
713
220k
      mpush (nextLo[0], nextHi[0], nextD[0]);
714
220k
      mpush (nextLo[1], nextHi[1], nextD[1]);
715
220k
      mpush (nextLo[2], nextHi[2], nextD[2]);
716
220k
   }
717
500k
}
718
719
#undef mswap
720
#undef mvswap
721
#undef mpush
722
#undef mpop
723
#undef mmin
724
#undef mnextsize
725
#undef mnextswap
726
#undef MAIN_QSORT_SMALL_THRESH
727
#undef MAIN_QSORT_DEPTH_THRESH
728
#undef MAIN_QSORT_STACK_SIZE
729
730
731
/*---------------------------------------------*/
732
/* Pre:
733
      nblock > N_OVERSHOOT
734
      block32 exists for [0 .. nblock-1 +N_OVERSHOOT]
735
      ((UChar*)block32) [0 .. nblock-1] holds block
736
      ptr exists for [0 .. nblock-1]
737
738
   Post:
739
      ((UChar*)block32) [0 .. nblock-1] holds block
740
      All other areas of block32 destroyed
741
      ftab [0 .. 65536 ] destroyed
742
      ptr [0 .. nblock-1] holds sorted order
743
      if (*budget < 0), sorting was abandoned
744
*/
745
746
20.3M
#define BIGFREQ(b) (ftab[((b)+1) << 8] - ftab[(b) << 8])
747
3.07G
#define SETMASK (1 << 21)
748
1.54G
#define CLEARMASK (~(SETMASK))
749
750
static
751
void mainSort ( UInt32* ptr, 
752
                UChar*  block,
753
                UInt16* quadrant, 
754
                UInt32* ftab,
755
                Int32   nblock,
756
                Int32   verb,
757
                Int32*  budget )
758
7.90k
{
759
7.90k
   Int32  i, j, k, ss, sb;
760
7.90k
   Int32  runningOrder[256];
761
7.90k
   Bool   bigDone[256];
762
7.90k
   Int32  copyStart[256];
763
7.90k
   Int32  copyEnd  [256];
764
7.90k
   UChar  c1;
765
7.90k
   Int32  numQSorted;
766
7.90k
   UInt16 s;
767
7.90k
   if (verb >= 4) VPrintf0 ( "        main sort initialise ...\n" );
768
769
   /*-- set up the 2-byte frequency table --*/
770
517M
   for (i = 65536; i >= 0; i--) ftab[i] = 0;
771
772
7.90k
   j = block[0] << 8;
773
7.90k
   i = nblock-1;
774
22.6M
   for (; i >= 3; i -= 4) {
775
22.6M
      quadrant[i] = 0;
776
22.6M
      j = (j >> 8) | ( ((UInt16)block[i]) << 8);
777
22.6M
      ftab[j]++;
778
22.6M
      quadrant[i-1] = 0;
779
22.6M
      j = (j >> 8) | ( ((UInt16)block[i-1]) << 8);
780
22.6M
      ftab[j]++;
781
22.6M
      quadrant[i-2] = 0;
782
22.6M
      j = (j >> 8) | ( ((UInt16)block[i-2]) << 8);
783
22.6M
      ftab[j]++;
784
22.6M
      quadrant[i-3] = 0;
785
22.6M
      j = (j >> 8) | ( ((UInt16)block[i-3]) << 8);
786
22.6M
      ftab[j]++;
787
22.6M
   }
788
17.3k
   for (; i >= 0; i--) {
789
9.48k
      quadrant[i] = 0;
790
9.48k
      j = (j >> 8) | ( ((UInt16)block[i]) << 8);
791
9.48k
      ftab[j]++;
792
9.48k
   }
793
794
   /*-- (emphasises close relationship of block & quadrant) --*/
795
276k
   for (i = 0; i < BZ_N_OVERSHOOT; i++) {
796
268k
      block   [nblock+i] = block[i];
797
268k
      quadrant[nblock+i] = 0;
798
268k
   }
799
800
7.90k
   if (verb >= 4) VPrintf0 ( "        bucket sorting ...\n" );
801
802
   /*-- Complete the initial radix sort --*/
803
517M
   for (i = 1; i <= 65536; i++) ftab[i] += ftab[i-1];
804
805
7.90k
   s = block[0] << 8;
806
7.90k
   i = nblock-1;
807
22.6M
   for (; i >= 3; i -= 4) {
808
22.6M
      s = (s >> 8) | (block[i] << 8);
809
22.6M
      j = ftab[s] -1;
810
22.6M
      ftab[s] = j;
811
22.6M
      ptr[j] = i;
812
22.6M
      s = (s >> 8) | (block[i-1] << 8);
813
22.6M
      j = ftab[s] -1;
814
22.6M
      ftab[s] = j;
815
22.6M
      ptr[j] = i-1;
816
22.6M
      s = (s >> 8) | (block[i-2] << 8);
817
22.6M
      j = ftab[s] -1;
818
22.6M
      ftab[s] = j;
819
22.6M
      ptr[j] = i-2;
820
22.6M
      s = (s >> 8) | (block[i-3] << 8);
821
22.6M
      j = ftab[s] -1;
822
22.6M
      ftab[s] = j;
823
22.6M
      ptr[j] = i-3;
824
22.6M
   }
825
17.3k
   for (; i >= 0; i--) {
826
9.48k
      s = (s >> 8) | (block[i] << 8);
827
9.48k
      j = ftab[s] -1;
828
9.48k
      ftab[s] = j;
829
9.48k
      ptr[j] = i;
830
9.48k
   }
831
832
   /*--
833
      Now ftab contains the first loc of every small bucket.
834
      Calculate the running order, from smallest to largest
835
      big bucket.
836
   --*/
837
2.03M
   for (i = 0; i <= 255; i++) {
838
2.02M
      bigDone     [i] = False;
839
2.02M
      runningOrder[i] = i;
840
2.02M
   }
841
842
7.90k
   {
843
7.90k
      Int32 vv;
844
7.90k
      Int32 h = 1;
845
39.5k
      do h = 3 * h + 1; while (h <= 256);
846
39.5k
      do {
847
39.5k
         h = h / 3;
848
8.73M
         for (i = h; i <= 255; i++) {
849
8.69M
            vv = runningOrder[i];
850
8.69M
            j = i;
851
10.1M
            while ( BIGFREQ(runningOrder[j-h]) > BIGFREQ(vv) ) {
852
1.61M
               runningOrder[j] = runningOrder[j-h];
853
1.61M
               j = j - h;
854
1.61M
               if (j <= (h - 1)) goto zero;
855
1.61M
            }
856
8.69M
            zero:
857
8.69M
            runningOrder[j] = vv;
858
8.69M
         }
859
39.5k
      } while (h != 1);
860
7.90k
   }
861
862
   /*--
863
      The main sorting loop.
864
   --*/
865
866
7.90k
   numQSorted = 0;
867
868
2.00M
   for (i = 0; i <= 255; i++) {
869
870
      /*--
871
         Process big buckets, starting with the least full.
872
         Basically this is a 3-step process in which we call
873
         mainQSort3 to sort the small buckets [ss, j], but
874
         also make a big effort to avoid the calls if we can.
875
      --*/
876
1.99M
      ss = runningOrder[i];
877
878
      /*--
879
         Step 1:
880
         Complete the big bucket [ss] by quicksorting
881
         any unsorted small buckets [ss, j], for j != ss.  
882
         Hopefully previous pointer-scanning phases have already
883
         completed many of the small buckets [ss, j], so
884
         we don't have to sort them at all.
885
      --*/
886
512M
      for (j = 0; j <= 255; j++) {
887
510M
         if (j != ss) {
888
508M
            sb = (ss << 8) + j;
889
508M
            if ( ! (ftab[sb] & SETMASK) ) {
890
257M
               Int32 lo = ftab[sb]   & CLEARMASK;
891
257M
               Int32 hi = (ftab[sb+1] & CLEARMASK) - 1;
892
257M
               if (hi > lo) {
893
500k
                  if (verb >= 4)
894
0
                     VPrintf4 ( "        qsort [0x%x, 0x%x]   "
895
500k
                                "done %d   this %d\n",
896
500k
                                ss, j, numQSorted, hi - lo + 1 );
897
500k
                  mainQSort3 ( 
898
500k
                     ptr, block, quadrant, nblock, 
899
500k
                     lo, hi, BZ_N_RADIX, budget 
900
500k
                  );   
901
500k
                  numQSorted += (hi - lo + 1);
902
500k
                  if (*budget < 0) return;
903
500k
               }
904
257M
            }
905
508M
            ftab[sb] |= SETMASK;
906
508M
         }
907
510M
      }
908
909
1.99M
      AssertH ( !bigDone[ss], 1006 );
910
911
      /*--
912
         Step 2:
913
         Now scan this big bucket [ss] so as to synthesise the
914
         sorted order for small buckets [t, ss] for all t,
915
         including, magically, the bucket [ss,ss] too.
916
         This will avoid doing Real Work in subsequent Step 1's.
917
      --*/
918
1.99M
      {
919
512M
         for (j = 0; j <= 255; j++) {
920
510M
            copyStart[j] =  ftab[(j << 8) + ss]     & CLEARMASK;
921
510M
            copyEnd  [j] = (ftab[(j << 8) + ss + 1] & CLEARMASK) - 1;
922
510M
         }
923
16.1M
         for (j = ftab[ss << 8] & CLEARMASK; j < copyStart[ss]; j++) {
924
14.1M
            k = ptr[j]-1; if (k < 0) k += nblock;
925
14.1M
            c1 = block[k];
926
14.1M
            if (!bigDone[c1])
927
9.09M
               ptr[ copyStart[c1]++ ] = k;
928
14.1M
         }
929
18.7M
         for (j = (ftab[(ss+1) << 8] & CLEARMASK) - 1; j > copyEnd[ss]; j--) {
930
16.7M
            k = ptr[j]-1; if (k < 0) k += nblock;
931
16.7M
            c1 = block[k];
932
16.7M
            if (!bigDone[c1]) 
933
10.0M
               ptr[ copyEnd[c1]-- ] = k;
934
16.7M
         }
935
1.99M
      }
936
937
1.99M
      AssertH ( (copyStart[ss]-1 == copyEnd[ss])
938
1.99M
                || 
939
                /* Extremely rare case missing in bzip2-1.0.0 and 1.0.1.
940
                   Necessity for this case is demonstrated by compressing 
941
                   a sequence of approximately 48.5 million of character 
942
                   251; 1.0.0/1.0.1 will then die here. */
943
1.99M
                (copyStart[ss] == 0 && copyEnd[ss] == nblock-1),
944
1.99M
                1007 )
945
946
512M
      for (j = 0; j <= 255; j++) ftab[(j << 8) + ss] |= SETMASK;
947
948
      /*--
949
         Step 3:
950
         The [ss] big bucket is now done.  Record this fact,
951
         and update the quadrant descriptors.  Remember to
952
         update quadrants in the overshoot area too, if
953
         necessary.  The "if (i < 255)" test merely skips
954
         this updating for the last bucket processed, since
955
         updating for the last bucket is pointless.
956
957
         The quadrant array provides a way to incrementally
958
         cache sort orderings, as they appear, so as to 
959
         make subsequent comparisons in fullGtU() complete
960
         faster.  For repetitive blocks this makes a big
961
         difference (but not big enough to be able to avoid
962
         the fallback sorting mechanism, exponential radix sort).
963
964
         The precise meaning is: at all times:
965
966
            for 0 <= i < nblock and 0 <= j <= nblock
967
968
            if block[i] != block[j], 
969
970
               then the relative values of quadrant[i] and 
971
                    quadrant[j] are meaningless.
972
973
               else {
974
                  if quadrant[i] < quadrant[j]
975
                     then the string starting at i lexicographically
976
                     precedes the string starting at j
977
978
                  else if quadrant[i] > quadrant[j]
979
                     then the string starting at j lexicographically
980
                     precedes the string starting at i
981
982
                  else
983
                     the relative ordering of the strings starting
984
                     at i and j has not yet been determined.
985
               }
986
      --*/
987
1.99M
      bigDone[ss] = True;
988
989
1.99M
      if (i < 255) {
990
1.99M
         Int32 bbStart  = ftab[ss << 8] & CLEARMASK;
991
1.99M
         Int32 bbSize   = (ftab[(ss+1) << 8] & CLEARMASK) - bbStart;
992
1.99M
         Int32 shifts   = 0;
993
994
1.99M
         while ((bbSize >> shifts) > 65534) shifts++;
995
996
22.8M
         for (j = bbSize-1; j >= 0; j--) {
997
20.8M
            Int32 a2update     = ptr[bbStart + j];
998
20.8M
            UInt16 qVal        = (UInt16)(j >> shifts);
999
20.8M
            quadrant[a2update] = qVal;
1000
20.8M
            if (a2update < BZ_N_OVERSHOOT)
1001
62.2k
               quadrant[a2update + nblock] = qVal;
1002
20.8M
         }
1003
1.99M
         AssertH ( ((bbSize-1) >> shifts) <= 65535, 1002 );
1004
1.99M
      }
1005
1006
1.99M
   }
1007
1008
2.53k
   if (verb >= 4)
1009
0
      VPrintf3 ( "        %d pointers, %d sorted, %d scanned\n",
1010
2.53k
                 nblock, numQSorted, nblock - numQSorted );
1011
2.53k
}
1012
1013
#undef BIGFREQ
1014
#undef SETMASK
1015
#undef CLEARMASK
1016
1017
1018
/*---------------------------------------------*/
1019
/* Pre:
1020
      nblock > 0
1021
      arr2 exists for [0 .. nblock-1 +N_OVERSHOOT]
1022
      ((UChar*)arr2)  [0 .. nblock-1] holds block
1023
      arr1 exists for [0 .. nblock-1]
1024
1025
   Post:
1026
      ((UChar*)arr2) [0 .. nblock-1] holds block
1027
      All other areas of block destroyed
1028
      ftab [ 0 .. 65536 ] destroyed
1029
      arr1 [0 .. nblock-1] holds sorted order
1030
*/
1031
void BZ2_blockSort ( EState* s )
1032
87.2k
{
1033
87.2k
   UInt32* ptr    = s->ptr; 
1034
87.2k
   UChar*  block  = s->block;
1035
87.2k
   UInt32* ftab   = s->ftab;
1036
87.2k
   Int32   nblock = s->nblock;
1037
87.2k
   Int32   verb   = s->verbosity;
1038
87.2k
   Int32   wfact  = s->workFactor;
1039
87.2k
   UInt16* quadrant;
1040
87.2k
   Int32   budget;
1041
87.2k
   Int32   budgetInit;
1042
87.2k
   Int32   i;
1043
1044
87.2k
   if (nblock < 10000) {
1045
79.3k
      fallbackSort ( s->arr1, s->arr2, ftab, nblock, verb );
1046
79.3k
   } else {
1047
      /* Calculate the location for quadrant, remembering to get
1048
         the alignment right.  Assumes that &(block[0]) is at least
1049
         2-byte aligned -- this should be ok since block is really
1050
         the first section of arr2.
1051
      */
1052
7.90k
      i = nblock+BZ_N_OVERSHOOT;
1053
7.90k
      if (i & 1) i++;
1054
7.90k
      quadrant = (UInt16*)(&(block[i]));
1055
1056
      /* (wfact-1) / 3 puts the default-factor-30
1057
         transition point at very roughly the same place as 
1058
         with v0.1 and v0.9.0.  
1059
         Not that it particularly matters any more, since the
1060
         resulting compressed stream is now the same regardless
1061
         of whether or not we use the main sort or fallback sort.
1062
      */
1063
7.90k
      if (wfact < 1  ) wfact = 1;
1064
7.90k
      if (wfact > 100) wfact = 100;
1065
7.90k
      budgetInit = nblock * ((wfact-1) / 3);
1066
7.90k
      budget = budgetInit;
1067
1068
7.90k
      mainSort ( ptr, block, quadrant, ftab, nblock, verb, &budget );
1069
7.90k
      if (verb >= 3) 
1070
0
         VPrintf3 ( "      %d work, %d block, ratio %5.2f\n",
1071
7.90k
                    budgetInit - budget,
1072
7.90k
                    nblock, 
1073
7.90k
                    (float)(budgetInit - budget) /
1074
7.90k
                    (float)(nblock==0 ? 1 : nblock) ); 
1075
7.90k
      if (budget < 0) {
1076
5.36k
         if (verb >= 2) 
1077
0
            VPrintf0 ( "    too repetitive; using fallback"
1078
5.36k
                       " sorting algorithm\n" );
1079
5.36k
         fallbackSort ( s->arr1, s->arr2, ftab, nblock, verb );
1080
5.36k
      }
1081
7.90k
   }
1082
1083
87.2k
   s->origPtr = -1;
1084
80.5M
   for (i = 0; i < s->nblock; i++)
1085
80.5M
      if (ptr[i] == 0)
1086
87.2k
         { s->origPtr = i; break; };
1087
1088
87.2k
   AssertH( s->origPtr != -1, 1003 );
1089
87.2k
}
1090
1091
1092
/*-------------------------------------------------------------*/
1093
/*--- end                                       blocksort.c ---*/
1094
/*-------------------------------------------------------------*/