aboutsummaryrefslogtreecommitdiff
path: root/posix/regexec.c
blob: 9720879722f2f8b44637a432bed2b5a436702fb9 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
3698
3699
3700
3701
3702
3703
3704
3705
3706
3707
3708
3709
3710
3711
3712
3713
3714
3715
3716
3717
3718
3719
3720
3721
3722
3723
3724
3725
3726
3727
3728
3729
3730
3731
3732
3733
3734
3735
3736
3737
3738
3739
3740
3741
3742
3743
3744
3745
3746
3747
3748
3749
3750
3751
3752
3753
3754
3755
3756
3757
3758
3759
3760
3761
3762
3763
3764
3765
3766
3767
3768
3769
3770
3771
3772
3773
3774
3775
3776
3777
3778
3779
3780
3781
3782
3783
3784
3785
3786
3787
3788
3789
3790
3791
3792
3793
3794
3795
3796
3797
3798
3799
3800
3801
3802
3803
3804
3805
3806
3807
3808
3809
3810
3811
3812
3813
3814
3815
3816
3817
3818
3819
3820
3821
3822
3823
3824
3825
3826
3827
3828
3829
3830
3831
3832
3833
3834
3835
3836
3837
3838
3839
3840
3841
3842
3843
3844
3845
3846
3847
3848
3849
3850
3851
3852
3853
3854
3855
3856
3857
3858
3859
3860
3861
3862
3863
3864
3865
3866
3867
3868
3869
3870
3871
3872
3873
3874
3875
3876
3877
3878
3879
3880
3881
3882
3883
3884
3885
3886
3887
3888
3889
3890
3891
3892
3893
3894
3895
3896
3897
3898
3899
3900
3901
3902
3903
3904
3905
3906
3907
3908
3909
3910
3911
3912
3913
3914
3915
3916
3917
3918
3919
3920
3921
3922
3923
3924
3925
3926
3927
3928
3929
3930
3931
3932
3933
3934
3935
3936
3937
3938
3939
3940
3941
3942
3943
3944
3945
3946
3947
3948
3949
3950
3951
3952
3953
3954
3955
3956
3957
3958
3959
3960
3961
3962
3963
3964
3965
3966
3967
3968
3969
3970
3971
3972
3973
3974
3975
3976
3977
3978
3979
3980
3981
3982
3983
3984
3985
3986
3987
3988
3989
3990
3991
3992
3993
3994
3995
3996
3997
3998
3999
4000
4001
4002
4003
4004
4005
4006
4007
4008
4009
4010
4011
4012
4013
4014
4015
4016
4017
4018
4019
4020
4021
4022
4023
4024
4025
4026
4027
4028
4029
4030
4031
4032
4033
4034
4035
4036
4037
4038
4039
4040
4041
4042
4043
4044
4045
4046
4047
4048
4049
4050
4051
4052
4053
4054
4055
4056
4057
4058
4059
4060
4061
4062
4063
4064
4065
4066
4067
4068
4069
4070
4071
4072
4073
4074
4075
4076
4077
4078
4079
4080
4081
4082
4083
4084
4085
4086
4087
4088
4089
4090
4091
4092
4093
4094
4095
4096
4097
4098
4099
4100
4101
/* Extended regular expression matching and search library.
   Copyright (C) 2002, 2003 Free Software Foundation, Inc.
   This file is part of the GNU C Library.
   Contributed by Isamu Hasegawa <isamu@yamato.ibm.com>.

   The GNU C Library is free software; you can redistribute it and/or
   modify it under the terms of the GNU Lesser General Public
   License as published by the Free Software Foundation; either
   version 2.1 of the License, or (at your option) any later version.

   The GNU C Library is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   Lesser General Public License for more details.

   You should have received a copy of the GNU Lesser General Public
   License along with the GNU C Library; if not, write to the Free
   Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
   02111-1307 USA.  */

static reg_errcode_t match_ctx_init (re_match_context_t *cache, int eflags,
				     re_string_t *input, int n);
static void match_ctx_clean (re_match_context_t *mctx);
static void match_ctx_free (re_match_context_t *cache);
static void match_ctx_free_subtops (re_match_context_t *mctx);
static reg_errcode_t match_ctx_add_entry (re_match_context_t *cache, int node,
					  int str_idx, int from, int to);
static int search_cur_bkref_entry (re_match_context_t *mctx, int str_idx);
static void match_ctx_clear_flag (re_match_context_t *mctx);
static reg_errcode_t match_ctx_add_subtop (re_match_context_t *mctx, int node,
					   int str_idx);
static re_sub_match_last_t * match_ctx_add_sublast (re_sub_match_top_t *subtop,
						   int node, int str_idx);
static void sift_ctx_init (re_sift_context_t *sctx, re_dfastate_t **sifted_sts,
			   re_dfastate_t **limited_sts, int last_node,
			   int last_str_idx, int check_subexp);
static reg_errcode_t re_search_internal (const regex_t *preg,
					 const char *string, int length,
					 int start, int range, int stop,
					 size_t nmatch, regmatch_t pmatch[],
					 int eflags);
static int re_search_2_stub (struct re_pattern_buffer *bufp,
			     const char *string1, int length1,
			     const char *string2, int length2,
			     int start, int range, struct re_registers *regs,
			     int stop, int ret_len);
static int re_search_stub (struct re_pattern_buffer *bufp,
			   const char *string, int length, int start,
			   int range, int stop, struct re_registers *regs,
			   int ret_len);
static unsigned re_copy_regs (struct re_registers *regs, regmatch_t *pmatch,
			      int nregs, int regs_allocated);
static re_dfastate_t *acquire_init_state_context (reg_errcode_t *err,
						  const regex_t *preg,
						  const re_match_context_t *mctx,
						  int idx);
static reg_errcode_t prune_impossible_nodes (const regex_t *preg,
					     re_match_context_t *mctx);
static int check_matching (const regex_t *preg, re_match_context_t *mctx,
			   int fl_longest_match);
static int check_halt_node_context (const re_dfa_t *dfa, int node,
				    unsigned int context);
static int check_halt_state_context (const regex_t *preg,
				     const re_dfastate_t *state,
				     const re_match_context_t *mctx, int idx);
static void update_regs (re_dfa_t *dfa, regmatch_t *pmatch, int cur_node,
			 int cur_idx, int nmatch);
static int proceed_next_node (const regex_t *preg, int nregs, regmatch_t *regs,
			      const re_match_context_t *mctx,
			      int *pidx, int node, re_node_set *eps_via_nodes,
			      struct re_fail_stack_t *fs);
static reg_errcode_t push_fail_stack (struct re_fail_stack_t *fs,
				      int str_idx, int *dests, int nregs,
				      regmatch_t *regs,
				      re_node_set *eps_via_nodes);
static int pop_fail_stack (struct re_fail_stack_t *fs, int *pidx, int nregs,
			   regmatch_t *regs, re_node_set *eps_via_nodes);
static reg_errcode_t set_regs (const regex_t *preg,
			       const re_match_context_t *mctx,
			       size_t nmatch, regmatch_t *pmatch,
			       int fl_backtrack);
static reg_errcode_t free_fail_stack_return (struct re_fail_stack_t *fs);

#ifdef RE_ENABLE_I18N
static int sift_states_iter_mb (const regex_t *preg,
				const re_match_context_t *mctx,
				re_sift_context_t *sctx,
				int node_idx, int str_idx, int max_str_idx);
#endif /* RE_ENABLE_I18N */
static reg_errcode_t sift_states_backward (const regex_t *preg,
					   re_match_context_t *mctx,
					   re_sift_context_t *sctx);
static reg_errcode_t update_cur_sifted_state (const regex_t *preg,
					      re_match_context_t *mctx,
					      re_sift_context_t *sctx,
					      int str_idx,
					      re_node_set *dest_nodes);
static reg_errcode_t add_epsilon_src_nodes (re_dfa_t *dfa,
					    re_node_set *dest_nodes,
					    const re_node_set *candidates);
static reg_errcode_t sub_epsilon_src_nodes (re_dfa_t *dfa, int node,
					    re_node_set *dest_nodes,
					    const re_node_set *and_nodes);
static int check_dst_limits (re_dfa_t *dfa, re_node_set *limits,
			     re_match_context_t *mctx, int dst_node,
			     int dst_idx, int src_node, int src_idx);
static int check_dst_limits_calc_pos (re_dfa_t *dfa, re_match_context_t *mctx,
				      int limit, re_node_set *eclosures,
				      int subexp_idx, int node, int str_idx);
static reg_errcode_t check_subexp_limits (re_dfa_t *dfa,
					  re_node_set *dest_nodes,
					  const re_node_set *candidates,
					  re_node_set *limits,
					  struct re_backref_cache_entry *bkref_ents,
					  int str_idx);
static reg_errcode_t sift_states_bkref (const regex_t *preg,
					re_match_context_t *mctx,
					re_sift_context_t *sctx,
					int str_idx, re_node_set *dest_nodes);
static reg_errcode_t clean_state_log_if_need (re_match_context_t *mctx,
					      int next_state_log_idx);
static reg_errcode_t merge_state_array (re_dfa_t *dfa, re_dfastate_t **dst,
					re_dfastate_t **src, int num);
static re_dfastate_t *transit_state (reg_errcode_t *err, const regex_t *preg,
				     re_match_context_t *mctx,
				     re_dfastate_t *state);
static reg_errcode_t check_subexp_matching_top (re_dfa_t *dfa,
						re_match_context_t *mctx,
						re_node_set *cur_nodes,
						int str_idx);
#if 0
static re_dfastate_t *transit_state_sb (reg_errcode_t *err, const regex_t *preg,
					re_dfastate_t *pstate,
					re_match_context_t *mctx);
#endif
#ifdef RE_ENABLE_I18N
static reg_errcode_t transit_state_mb (const regex_t *preg,
				       re_dfastate_t *pstate,
				       re_match_context_t *mctx);
#endif /* RE_ENABLE_I18N */
static reg_errcode_t transit_state_bkref (const regex_t *preg,
					  const re_node_set *nodes,
					  re_match_context_t *mctx);
static reg_errcode_t get_subexp (const regex_t *preg, re_match_context_t *mctx,
				 int bkref_node, int bkref_str_idx);
static reg_errcode_t get_subexp_sub (const regex_t *preg,
				     re_match_context_t *mctx,
				     const re_sub_match_top_t *sub_top,
				     re_sub_match_last_t *sub_last,
				     int bkref_node, int bkref_str);
static int find_subexp_node (const re_dfa_t *dfa, const re_node_set *nodes,
			     int subexp_idx, int fl_open);
static reg_errcode_t check_arrival (const regex_t *preg,
				    re_match_context_t *mctx,
				    state_array_t *path, int top_node,
				    int top_str, int last_node, int last_str,
				    int fl_open);
static reg_errcode_t check_arrival_add_next_nodes (const regex_t *preg,
						   re_dfa_t *dfa,
						   re_match_context_t *mctx,
						   int str_idx,
						   re_node_set *cur_nodes,
						   re_node_set *next_nodes);
static reg_errcode_t check_arrival_expand_ecl (re_dfa_t *dfa,
					       re_node_set *cur_nodes,
					       int ex_subexp, int fl_open);
static reg_errcode_t check_arrival_expand_ecl_sub (re_dfa_t *dfa,
						   re_node_set *dst_nodes,
						   int target, int ex_subexp,
						   int fl_open);
static reg_errcode_t expand_bkref_cache (const regex_t *preg,
					 re_match_context_t *mctx,
					 re_node_set *cur_nodes, int cur_str,
					 int last_str, int subexp_num,
					 int fl_open);
static re_dfastate_t **build_trtable (const regex_t *dfa,
				      re_dfastate_t *state);
#ifdef RE_ENABLE_I18N
static int check_node_accept_bytes (const regex_t *preg, int node_idx,
				    const re_string_t *input, int idx);
# ifdef _LIBC
static unsigned int find_collation_sequence_value (const unsigned char *mbs,
						   size_t name_len);
# endif /* _LIBC */
#endif /* RE_ENABLE_I18N */
static int group_nodes_into_DFAstates (const regex_t *dfa,
				       const re_dfastate_t *state,
				       re_node_set *states_node,
				       bitset *states_ch);
static int check_node_accept (const regex_t *preg, const re_token_t *node,
			      const re_match_context_t *mctx, int idx);
static reg_errcode_t extend_buffers (re_match_context_t *mctx);

/* Entry point for POSIX code.  */

/* regexec searches for a given pattern, specified by PREG, in the
   string STRING.

   If NMATCH is zero or REG_NOSUB was set in the cflags argument to
   `regcomp', we ignore PMATCH.  Otherwise, we assume PMATCH has at
   least NMATCH elements, and we set them to the offsets of the
   corresponding matched substrings.

   EFLAGS specifies `execution flags' which affect matching: if
   REG_NOTBOL is set, then ^ does not match at the beginning of the
   string; if REG_NOTEOL is set, then $ does not match at the end.

   We return 0 if we find a match and REG_NOMATCH if not.  */

int
regexec (preg, string, nmatch, pmatch, eflags)
    const regex_t *__restrict preg;
    const char *__restrict string;
    size_t nmatch;
    regmatch_t pmatch[];
    int eflags;
{
  reg_errcode_t err;
  int length = strlen (string);
  if (preg->no_sub)
    err = re_search_internal (preg, string, length, 0, length, length, 0,
			      NULL, eflags);
  else
    err = re_search_internal (preg, string, length, 0, length, length, nmatch,
			      pmatch, eflags);
  return err != REG_NOERROR;
}
#ifdef _LIBC
weak_alias (__regexec, regexec)
#endif

/* Entry points for GNU code.  */

/* re_match, re_search, re_match_2, re_search_2

   The former two functions operate on STRING with length LENGTH,
   while the later two operate on concatenation of STRING1 and STRING2
   with lengths LENGTH1 and LENGTH2, respectively.

   re_match() matches the compiled pattern in BUFP against the string,
   starting at index START.

   re_search() first tries matching at index START, then it tries to match
   starting from index START + 1, and so on.  The last start position tried
   is START + RANGE.  (Thus RANGE = 0 forces re_search to operate the same
   way as re_match().)

   The parameter STOP of re_{match,search}_2 specifies that no match exceeding
   the first STOP characters of the concatenation of the strings should be
   concerned.

   If REGS is not NULL, and BUFP->no_sub is not set, the offsets of the match
   and all groups is stroed in REGS.  (For the "_2" variants, the offsets are
   computed relative to the concatenation, not relative to the individual
   strings.)

   On success, re_match* functions return the length of the match, re_search*
   return the position of the start of the match.  Return value -1 means no
   match was found and -2 indicates an internal error.  */

int
re_match (bufp, string, length, start, regs)
    struct re_pattern_buffer *bufp;
    const char *string;
    int length, start;
    struct re_registers *regs;
{
  return re_search_stub (bufp, string, length, start, 0, length, regs, 1);
}
#ifdef _LIBC
weak_alias (__re_match, re_match)
#endif

int
re_search (bufp, string, length, start, range, regs)
    struct re_pattern_buffer *bufp;
    const char *string;
    int length, start, range;
    struct re_registers *regs;
{
  return re_search_stub (bufp, string, length, start, range, length, regs, 0);
}
#ifdef _LIBC
weak_alias (__re_search, re_search)
#endif

int
re_match_2 (bufp, string1, length1, string2, length2, start, regs, stop)
    struct re_pattern_buffer *bufp;
    const char *string1, *string2;
    int length1, length2, start, stop;
    struct re_registers *regs;
{
  return re_search_2_stub (bufp, string1, length1, string2, length2,
			   start, 0, regs, stop, 1);
}
#ifdef _LIBC
weak_alias (__re_match_2, re_match_2)
#endif

int
re_search_2 (bufp, string1, length1, string2, length2, start, range, regs, stop)
    struct re_pattern_buffer *bufp;
    const char *string1, *string2;
    int length1, length2, start, range, stop;
    struct re_registers *regs;
{
  return re_search_2_stub (bufp, string1, length1, string2, length2,
			   start, range, regs, stop, 0);
}
#ifdef _LIBC
weak_alias (__re_search_2, re_search_2)
#endif

static int
re_search_2_stub (bufp, string1, length1, string2, length2, start, range, regs,
		  stop, ret_len)
    struct re_pattern_buffer *bufp;
    const char *string1, *string2;
    int length1, length2, start, range, stop, ret_len;
    struct re_registers *regs;
{
  const char *str;
  int rval;
  int len = length1 + length2;
  int free_str = 0;

  if (BE (length1 < 0 || length2 < 0 || stop < 0, 0))
    return -2;

  /* Concatenate the strings.  */
  if (length2 > 0)
    if (length1 > 0)
      {
	char *s = re_malloc (char, len);

	if (BE (s == NULL, 0))
	  return -2;
	memcpy (s, string1, length1);
	memcpy (s + length1, string2, length2);
	str = s;
	free_str = 1;
      }
    else
      str = string2;
  else
    str = string1;

  rval = re_search_stub (bufp, str, len, start, range, stop, regs,
			 ret_len);
  if (free_str)
    re_free ((char *) str);
  return rval;
}

/* The parameters have the same meaning as those of re_search.
   Additional parameters:
   If RET_LEN is nonzero the length of the match is returned (re_match style);
   otherwise the position of the match is returned.  */

static int
re_search_stub (bufp, string, length, start, range, stop, regs, ret_len)
    struct re_pattern_buffer *bufp;
    const char *string;
    int length, start, range, stop, ret_len;
    struct re_registers *regs;
{
  reg_errcode_t result;
  regmatch_t *pmatch;
  int nregs, rval;
  int eflags = 0;

  /* Check for out-of-range.  */
  if (BE (start < 0 || start > length, 0))
    return -1;
  if (BE (start + range > length, 0))
    range = length - start;
  else if (BE (start + range < 0, 0))
    range = -start;

  eflags |= (bufp->not_bol) ? REG_NOTBOL : 0;
  eflags |= (bufp->not_eol) ? REG_NOTEOL : 0;

  /* Compile fastmap if we haven't yet.  */
  if (range > 0 && bufp->fastmap != NULL && !bufp->fastmap_accurate)
    re_compile_fastmap (bufp);

  if (BE (bufp->no_sub, 0))
    regs = NULL;

  /* We need at least 1 register.  */
  if (regs == NULL)
    nregs = 1;
  else if (BE (bufp->regs_allocated == REGS_FIXED &&
	       regs->num_regs < bufp->re_nsub + 1, 0))
    {
      nregs = regs->num_regs;
      if (BE (nregs < 1, 0))
	{
	  /* Nothing can be copied to regs.  */
	  regs = NULL;
	  nregs = 1;
	}
    }
  else
    nregs = bufp->re_nsub + 1;
  pmatch = re_malloc (regmatch_t, nregs);
  if (BE (pmatch == NULL, 0))
    return -2;

  result = re_search_internal (bufp, string, length, start, range, stop,
			       nregs, pmatch, eflags);

  rval = 0;

  /* I hope we needn't fill ther regs with -1's when no match was found.  */
  if (result != REG_NOERROR)
    rval = -1;
  else if (regs != NULL)
    {
      /* If caller wants register contents data back, copy them.  */
      bufp->regs_allocated = re_copy_regs (regs, pmatch, nregs,
					   bufp->regs_allocated);
      if (BE (bufp->regs_allocated == REGS_UNALLOCATED, 0))
	rval = -2;
    }

  if (BE (rval == 0, 1))
    {
      if (ret_len)
	{
	  assert (pmatch[0].rm_so == start);
	  rval = pmatch[0].rm_eo - start;
	}
      else
	rval = pmatch[0].rm_so;
    }
  re_free (pmatch);
  return rval;
}

static unsigned
re_copy_regs (regs, pmatch, nregs, regs_allocated)
    struct re_registers *regs;
    regmatch_t *pmatch;
    int nregs, regs_allocated;
{
  int rval = REGS_REALLOCATE;
  int i;
  int need_regs = nregs + 1;
  /* We need one extra element beyond `num_regs' for the `-1' marker GNU code
     uses.  */

  /* Have the register data arrays been allocated?  */
  if (regs_allocated == REGS_UNALLOCATED)
    { /* No.  So allocate them with malloc.  */
      regs->start = re_malloc (regoff_t, need_regs);
      if (BE (regs->start == NULL, 0))
	return REGS_UNALLOCATED;
      regs->end = re_malloc (regoff_t, need_regs);
      if (BE (regs->end == NULL, 0))
	{
	  re_free (regs->start);
	  return REGS_UNALLOCATED;
	}
      regs->num_regs = need_regs;
    }
  else if (regs_allocated == REGS_REALLOCATE)
    { /* Yes.  If we need more elements than were already
	 allocated, reallocate them.  If we need fewer, just
	 leave it alone.  */
      if (BE (need_regs > regs->num_regs, 0))
	{
	  regs->start = re_realloc (regs->start, regoff_t, need_regs);
	  if (BE (regs->start == NULL, 0))
	    {
	      if (regs->end != NULL)
		re_free (regs->end);
	      return REGS_UNALLOCATED;
	    }
	  regs->end = re_realloc (regs->end, regoff_t, need_regs);
	  if (BE (regs->end == NULL, 0))
	    {
	      re_free (regs->start);
	      return REGS_UNALLOCATED;
	    }
	  regs->num_regs = need_regs;
	}
    }
  else
    {
      assert (regs_allocated == REGS_FIXED);
      /* This function may not be called with REGS_FIXED and nregs too big.  */
      assert (regs->num_regs >= nregs);
      rval = REGS_FIXED;
    }

  /* Copy the regs.  */
  for (i = 0; i < nregs; ++i)
    {
      regs->start[i] = pmatch[i].rm_so;
      regs->end[i] = pmatch[i].rm_eo;
    }
  for ( ; i < regs->num_regs; ++i)
    regs->start[i] = regs->end[i] = -1;

  return rval;
}

/* Set REGS to hold NUM_REGS registers, storing them in STARTS and
   ENDS.  Subsequent matches using PATTERN_BUFFER and REGS will use
   this memory for recording register information.  STARTS and ENDS
   must be allocated using the malloc library routine, and must each
   be at least NUM_REGS * sizeof (regoff_t) bytes long.

   If NUM_REGS == 0, then subsequent matches should allocate their own
   register data.

   Unless this function is called, the first search or match using
   PATTERN_BUFFER will allocate its own register data, without
   freeing the old data.  */

void
re_set_registers (bufp, regs, num_regs, starts, ends)
    struct re_pattern_buffer *bufp;
    struct re_registers *regs;
    unsigned num_regs;
    regoff_t *starts, *ends;
{
  if (num_regs)
    {
      bufp->regs_allocated = REGS_REALLOCATE;
      regs->num_regs = num_regs;
      regs->start = starts;
      regs->end = ends;
    }
  else
    {
      bufp->regs_allocated = REGS_UNALLOCATED;
      regs->num_regs = 0;
      regs->start = regs->end = (regoff_t *) 0;
    }
}
#ifdef _LIBC
weak_alias (__re_set_registers, re_set_registers)
#endif

/* Entry points compatible with 4.2 BSD regex library.  We don't define
   them unless specifically requested.  */

#if defined _REGEX_RE_COMP || defined _LIBC
int
# ifdef _LIBC
weak_function
# endif
re_exec (s)
     const char *s;
{
  return 0 == regexec (&re_comp_buf, s, 0, NULL, 0);
}
#endif /* _REGEX_RE_COMP */

static re_node_set empty_set;

/* Internal entry point.  */

/* Searches for a compiled pattern PREG in the string STRING, whose
   length is LENGTH.  NMATCH, PMATCH, and EFLAGS have the same
   mingings with regexec.  START, and RANGE have the same meanings
   with re_search.
   Return REG_NOERROR if we find a match, and REG_NOMATCH if not,
   otherwise return the error code.
   Note: We assume front end functions already check ranges.
   (START + RANGE >= 0 && START + RANGE <= LENGTH)  */

static reg_errcode_t
re_search_internal (preg, string, length, start, range, stop, nmatch, pmatch,
		    eflags)
    const regex_t *preg;
    const char *string;
    int length, start, range, stop, eflags;
    size_t nmatch;
    regmatch_t pmatch[];
{
  reg_errcode_t err;
  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
  re_string_t input;
  int left_lim, right_lim, incr;
  int fl_longest_match, match_first, match_last = -1;
  int fast_translate, sb;
  re_match_context_t mctx;
  char *fastmap = ((preg->fastmap != NULL && preg->fastmap_accurate
		    && range && !preg->can_be_null) ? preg->fastmap : NULL);

  /* Check if the DFA haven't been compiled.  */
  if (BE (preg->used == 0 || dfa->init_state == NULL
	  || dfa->init_state_word == NULL || dfa->init_state_nl == NULL
	  || dfa->init_state_begbuf == NULL, 0))
    return REG_NOMATCH;

  re_node_set_init_empty (&empty_set);
  memset (&mctx, '\0', sizeof (re_match_context_t));

  /* We must check the longest matching, if nmatch > 0.  */
  fl_longest_match = (nmatch != 0 || dfa->nbackref);

  err = re_string_allocate (&input, string, length, dfa->nodes_len + 1,
			    preg->translate, preg->syntax & RE_ICASE, dfa);
  if (BE (err != REG_NOERROR, 0))
    goto free_return;
  input.stop = stop;

  err = match_ctx_init (&mctx, eflags, &input, dfa->nbackref * 2);
  if (BE (err != REG_NOERROR, 0))
    goto free_return;

  /* We will log all the DFA states through which the dfa pass,
     if nmatch > 1, or this dfa has "multibyte node", which is a
     back-reference or a node which can accept multibyte character or
     multi character collating element.  */
  if (nmatch > 1 || dfa->has_mb_node)
    {
      mctx.state_log = re_malloc (re_dfastate_t *, input.bufs_len + 1);
      if (BE (mctx.state_log == NULL, 0))
	{
	  err = REG_ESPACE;
	  goto free_return;
	}
    }
  else
    mctx.state_log = NULL;

#ifdef DEBUG
  /* We assume front-end functions already check them.  */
  assert (start + range >= 0 && start + range <= length);
#endif

  match_first = start;
  input.tip_context = ((eflags & REG_NOTBOL) ? CONTEXT_BEGBUF
		       : CONTEXT_NEWLINE | CONTEXT_BEGBUF);

  /* Check incrementally whether of not the input string match.  */
  incr = (range < 0) ? -1 : 1;
  left_lim = (range < 0) ? start + range : start;
  right_lim = (range < 0) ? start : start + range;
  sb = dfa->mb_cur_max == 1;
  fast_translate = sb || !(preg->syntax & RE_ICASE || preg->translate);

  for (;;)
    {
      /* At first get the current byte from input string.  */
      if (fastmap)
	{
	  if (BE (fast_translate, 1))
	    {
	      unsigned RE_TRANSLATE_TYPE t
		= (unsigned RE_TRANSLATE_TYPE) preg->translate;
	      if (BE (range >= 0, 1))
		{
		  if (BE (t != NULL, 0))
		    {
		      while (BE (match_first < right_lim, 1)
			     && !fastmap[t[(unsigned char) string[match_first]]])
			++match_first;
		    }
		  else
		    {
		      while (BE (match_first < right_lim, 1)
			     && !fastmap[(unsigned char) string[match_first]])
			++match_first;
		    }
		  if (BE (match_first == right_lim, 0))
		    {
		      int ch = match_first >= length
			       ? 0 : (unsigned char) string[match_first];
		      if (!fastmap[t ? t[ch] : ch])
			break;
		    }
		}
	      else
		{
		  while (match_first >= left_lim)
		    {
		      int ch = match_first >= length
			       ? 0 : (unsigned char) string[match_first];
		      if (fastmap[t ? t[ch] : ch])
			break;
		      --match_first;
		    }
		  if (match_first < left_lim)
		    break;
		}
	    }
	  else
	    {
	      int ch;

	      do
		{
		  /* In this case, we can't determine easily the current byte,
		     since it might be a component byte of a multibyte
		     character.  Then we use the constructed buffer
		     instead.  */
		  /* If MATCH_FIRST is out of the valid range, reconstruct the
		     buffers.  */
		  if (input.raw_mbs_idx + input.valid_len <= match_first
		      || match_first < input.raw_mbs_idx)
		    {
		      err = re_string_reconstruct (&input, match_first, eflags,
						   preg->newline_anchor);
		      if (BE (err != REG_NOERROR, 0))
			goto free_return;
		    }
		  /* If MATCH_FIRST is out of the buffer, leave it as '\0'.
		     Note that MATCH_FIRST must not be smaller than 0.  */
		  ch = ((match_first >= length) ? 0
		       : re_string_byte_at (&input,
					    match_first - input.raw_mbs_idx));
		  if (fastmap[ch])
		    break;
		  match_first += incr;
		}
	      while (match_first >= left_lim && match_first <= right_lim);
	      if (! fastmap[ch])
		break;
	    }
	}

      /* Reconstruct the buffers so that the matcher can assume that
	 the matching starts from the beginning of the buffer.  */
      err = re_string_reconstruct (&input, match_first, eflags,
				   preg->newline_anchor);
      if (BE (err != REG_NOERROR, 0))
	goto free_return;
#ifdef RE_ENABLE_I18N
     /* Eliminate it when it is a component of a multibyte character
	 and isn't the head of a multibyte character.  */
      if (sb || re_string_first_byte (&input, 0))
#endif
	{
	  /* It seems to be appropriate one, then use the matcher.  */
	  /* We assume that the matching starts from 0.  */
	  mctx.state_log_top = mctx.nbkref_ents = mctx.max_mb_elem_len = 0;
	  match_last = check_matching (preg, &mctx, fl_longest_match);
	  if (match_last != -1)
	    {
	      if (BE (match_last == -2, 0))
		{
		  err = REG_ESPACE;
		  goto free_return;
		}
	      else
		{
		  mctx.match_last = match_last;
		  if ((!preg->no_sub && nmatch > 1) || dfa->nbackref)
		    {
		      re_dfastate_t *pstate = mctx.state_log[match_last];
		      mctx.last_node = check_halt_state_context (preg, pstate,
								 &mctx, match_last);
		    }
		  if ((!preg->no_sub && nmatch > 1 && dfa->has_plural_match)
		      || dfa->nbackref)
		    {
		      err = prune_impossible_nodes (preg, &mctx);
		      if (err == REG_NOERROR)
			break;
		      if (BE (err != REG_NOMATCH, 0))
			goto free_return;
		      match_last = -1;
		    }
		  else
		    break; /* We found a match.  */
		}
	    }
	  match_ctx_clean (&mctx);
	}
      /* Update counter.  */
      match_first += incr;
      if (match_first < left_lim || right_lim < match_first)
	break;
    }

  /* Set pmatch[] if we need.  */
  if (match_last != -1 && nmatch > 0)
    {
      int reg_idx;

      /* Initialize registers.  */
      for (reg_idx = 1; reg_idx < nmatch; ++reg_idx)
	pmatch[reg_idx].rm_so = pmatch[reg_idx].rm_eo = -1;

      /* Set the points where matching start/end.  */
      pmatch[0].rm_so = 0;
      pmatch[0].rm_eo = mctx.match_last;

      if (!preg->no_sub && nmatch > 1)
	{
	  err = set_regs (preg, &mctx, nmatch, pmatch,
			  dfa->has_plural_match && dfa->nbackref > 0);
	  if (BE (err != REG_NOERROR, 0))
	    goto free_return;
	}

      /* At last, add the offset to the each registers, since we slided
	 the buffers so that we could assume that the matching starts
	 from 0.  */
      for (reg_idx = 0; reg_idx < nmatch; ++reg_idx)
	if (pmatch[reg_idx].rm_so != -1)
	  {
	    pmatch[reg_idx].rm_so += match_first;
	    pmatch[reg_idx].rm_eo += match_first;
	  }
    }
  err = (match_last == -1) ? REG_NOMATCH : REG_NOERROR;
 free_return:
  re_free (mctx.state_log);
  if (dfa->nbackref)
    match_ctx_free (&mctx);
  re_string_destruct (&input);
  return err;
}

static reg_errcode_t
prune_impossible_nodes (preg, mctx)
     const regex_t *preg;
     re_match_context_t *mctx;
{
  int halt_node, match_last;
  reg_errcode_t ret;
  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
  re_dfastate_t **sifted_states;
  re_dfastate_t **lim_states = NULL;
  re_sift_context_t sctx;
#ifdef DEBUG
  assert (mctx->state_log != NULL);
#endif
  match_last = mctx->match_last;
  halt_node = mctx->last_node;
  sifted_states = re_malloc (re_dfastate_t *, match_last + 1);
  if (BE (sifted_states == NULL, 0))
    {
      ret = REG_ESPACE;
      goto free_return;
    }
  if (dfa->nbackref)
    {
      lim_states = re_malloc (re_dfastate_t *, match_last + 1);
      if (BE (lim_states == NULL, 0))
	{
	  ret = REG_ESPACE;
	  goto free_return;
	}
      while (1)
	{
	  memset (lim_states, '\0',
		  sizeof (re_dfastate_t *) * (match_last + 1));
	  match_ctx_clear_flag (mctx);
	  sift_ctx_init (&sctx, sifted_states, lim_states, halt_node,
			 match_last, 0);
	  ret = sift_states_backward (preg, mctx, &sctx);
	  re_node_set_free (&sctx.limits);
	  if (BE (ret != REG_NOERROR, 0))
	      goto free_return;
	  if (sifted_states[0] != NULL || lim_states[0] != NULL)
	    break;
	  do
	    {
	      --match_last;
	      if (match_last < 0)
		{
		  ret = REG_NOMATCH;
		  goto free_return;
		}
	    } while (mctx->state_log[match_last] == NULL
		     || !mctx->state_log[match_last]->halt);
	  halt_node = check_halt_state_context (preg,
						mctx->state_log[match_last],
						mctx, match_last);
	}
      ret = merge_state_array (dfa, sifted_states, lim_states,
			       match_last + 1);
      re_free (lim_states);
      lim_states = NULL;
      if (BE (ret != REG_NOERROR, 0))
	goto free_return;
    }
  else
    {
      sift_ctx_init (&sctx, sifted_states, lim_states, halt_node,
		     match_last, 0);
      ret = sift_states_backward (preg, mctx, &sctx);
      re_node_set_free (&sctx.limits);
      if (BE (ret != REG_NOERROR, 0))
	goto free_return;
    }
  re_free (mctx->state_log);
  mctx->state_log = sifted_states;
  sifted_states = NULL;
  mctx->last_node = halt_node;
  mctx->match_last = match_last;
  ret = REG_NOERROR;
 free_return:
  re_free (sifted_states);
  re_free (lim_states);
  return ret;
}

/* Acquire an initial state and return it.
   We must select appropriate initial state depending on the context,
   since initial states may have constraints like "\<", "^", etc..  */

static re_dfastate_t *
acquire_init_state_context (err, preg, mctx, idx)
     reg_errcode_t *err;
     const regex_t *preg;
     const re_match_context_t *mctx;
     int idx;
{
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;

  *err = REG_NOERROR;
  if (dfa->init_state->has_constraint)
    {
      unsigned int context;
      context = re_string_context_at (mctx->input, idx - 1, mctx->eflags,
				      preg->newline_anchor);
      if (IS_WORD_CONTEXT (context))
	return dfa->init_state_word;
      else if (IS_ORDINARY_CONTEXT (context))
	return dfa->init_state;
      else if (IS_BEGBUF_CONTEXT (context) && IS_NEWLINE_CONTEXT (context))
	return dfa->init_state_begbuf;
      else if (IS_NEWLINE_CONTEXT (context))
	return dfa->init_state_nl;
      else if (IS_BEGBUF_CONTEXT (context))
	{
	  /* It is relatively rare case, then calculate on demand.  */
	  return  re_acquire_state_context (err, dfa,
					    dfa->init_state->entrance_nodes,
					    context);
	}
      else
	/* Must not happen?  */
	return dfa->init_state;
    }
  else
    return dfa->init_state;
}

/* Check whether the regular expression match input string INPUT or not,
   and return the index where the matching end, return -1 if not match,
   or return -2 in case of an error.
   FL_LONGEST_MATCH means we want the POSIX longest matching.
   Note that the matcher assume that the maching starts from the current
   index of the buffer.  */

static int
check_matching (preg, mctx, fl_longest_match)
    const regex_t *preg;
    re_match_context_t *mctx;
    int fl_longest_match;
{
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  reg_errcode_t err;
  int match = 0;
  int match_last = -1;
  int cur_str_idx = re_string_cur_idx (mctx->input);
  re_dfastate_t *cur_state;

  cur_state = acquire_init_state_context (&err, preg, mctx, cur_str_idx);
  /* An initial state must not be NULL(invalid state).  */
  if (BE (cur_state == NULL, 0))
    return -2;
  if (mctx->state_log != NULL)
    mctx->state_log[cur_str_idx] = cur_state;

  /* Check OP_OPEN_SUBEXP in the initial state in case that we use them
     later.  E.g. Processing back references.  */
  if (dfa->nbackref)
    {
      err = check_subexp_matching_top (dfa, mctx, &cur_state->nodes, 0);
      if (BE (err != REG_NOERROR, 0))
	return err;
    }

  if (cur_state->has_backref)
    {
      err = transit_state_bkref (preg, &cur_state->nodes, mctx);
      if (BE (err != REG_NOERROR, 0))
	return err;
    }

  /* If the RE accepts NULL string.  */
  if (cur_state->halt)
    {
      if (!cur_state->has_constraint
	  || check_halt_state_context (preg, cur_state, mctx, cur_str_idx))
	{
	  if (!fl_longest_match)
	    return cur_str_idx;
	  else
	    {
	      match_last = cur_str_idx;
	      match = 1;
	    }
	}
    }

  while (!re_string_eoi (mctx->input))
    {
      cur_state = transit_state (&err, preg, mctx, cur_state);
      if (cur_state == NULL) /* Reached at the invalid state or an error.  */
	{
	  cur_str_idx = re_string_cur_idx (mctx->input);
	  if (BE (err != REG_NOERROR, 0))
	    return -2;
	  if (!fl_longest_match && match)
	    break;
	  else
	    {
	      if (mctx->state_log == NULL)
		break;
	      else
		{
		  int max = mctx->state_log_top;
		  for (; cur_str_idx <= max; ++cur_str_idx)
		    if (mctx->state_log[cur_str_idx] != NULL)
		      break;
		  if (cur_str_idx > max)
		    break;
		}
	    }
	}

      if (cur_state != NULL && cur_state->halt)
	{
	  /* Reached at a halt state.
	     Check the halt state can satisfy the current context.  */
	  if (!cur_state->has_constraint
	      || check_halt_state_context (preg, cur_state, mctx,
					   re_string_cur_idx (mctx->input)))
	    {
	      /* We found an appropriate halt state.  */
	      match_last = re_string_cur_idx (mctx->input);
	      match = 1;
	      if (!fl_longest_match)
		break;
	    }
	}
   }
  return match_last;
}

/* Check NODE match the current context.  */

static int check_halt_node_context (dfa, node, context)
    const re_dfa_t *dfa;
    int node;
    unsigned int context;
{
  re_token_type_t type = dfa->nodes[node].type;
  unsigned int constraint = dfa->nodes[node].constraint;
  if (type != END_OF_RE)
    return 0;
  if (!constraint)
    return 1;
  if (NOT_SATISFY_NEXT_CONSTRAINT (constraint, context))
    return 0;
  return 1;
}

/* Check the halt state STATE match the current context.
   Return 0 if not match, if the node, STATE has, is a halt node and
   match the context, return the node.  */

static int
check_halt_state_context (preg, state, mctx, idx)
    const regex_t *preg;
    const re_dfastate_t *state;
    const re_match_context_t *mctx;
    int idx;
{
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  int i;
  unsigned int context;
#ifdef DEBUG
  assert (state->halt);
#endif
  context = re_string_context_at (mctx->input, idx, mctx->eflags,
				  preg->newline_anchor);
  for (i = 0; i < state->nodes.nelem; ++i)
    if (check_halt_node_context (dfa, state->nodes.elems[i], context))
      return state->nodes.elems[i];
  return 0;
}

/* Compute the next node to which "NFA" transit from NODE("NFA" is a NFA
   corresponding to the DFA).
   Return the destination node, and update EPS_VIA_NODES, return -1 in case
   of errors.  */

static int
proceed_next_node (preg, nregs, regs, mctx, pidx, node, eps_via_nodes, fs)
    const regex_t *preg;
    regmatch_t *regs;
    const re_match_context_t *mctx;
    int nregs, *pidx, node;
    re_node_set *eps_via_nodes;
    struct re_fail_stack_t *fs;
{
  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
  int i, err, dest_node;
  dest_node = -1;
  if (IS_EPSILON_NODE (dfa->nodes[node].type))
    {
      re_node_set *cur_nodes = &mctx->state_log[*pidx]->nodes;
      int ndest, dest_nodes[2];
      err = re_node_set_insert (eps_via_nodes, node);
      if (BE (err < 0, 0))
	return -1;
      /* Pick up valid destinations.  */
      for (ndest = 0, i = 0; i < dfa->edests[node].nelem; ++i)
	{
	  int candidate = dfa->edests[node].elems[i];
	  if (!re_node_set_contains (cur_nodes, candidate))
	    continue;
	  dest_nodes[0] = (ndest == 0) ? candidate : dest_nodes[0];
	  dest_nodes[1] = (ndest == 1) ? candidate : dest_nodes[1];
	  ++ndest;
	}
      if (ndest <= 1)
	return ndest == 0 ? -1 : (ndest == 1 ? dest_nodes[0] : 0);
      /* In order to avoid infinite loop like "(a*)*".  */
      if (re_node_set_contains (eps_via_nodes, dest_nodes[0]))
	return dest_nodes[1];
      if (fs != NULL)
	push_fail_stack (fs, *pidx, dest_nodes, nregs, regs, eps_via_nodes);
      return dest_nodes[0];
    }
  else
    {
      int naccepted = 0;
      re_token_type_t type = dfa->nodes[node].type;

#ifdef RE_ENABLE_I18N
      if (ACCEPT_MB_NODE (type))
	naccepted = check_node_accept_bytes (preg, node, mctx->input, *pidx);
      else
#endif /* RE_ENABLE_I18N */
      if (type == OP_BACK_REF)
	{
	  int subexp_idx = dfa->nodes[node].opr.idx;
	  naccepted = regs[subexp_idx].rm_eo - regs[subexp_idx].rm_so;
	  if (fs != NULL)
	    {
	      if (regs[subexp_idx].rm_so == -1 || regs[subexp_idx].rm_eo == -1)
		return -1;
	      else if (naccepted)
		{
		  char *buf = (char *) re_string_get_buffer (mctx->input);
		  if (memcmp (buf + regs[subexp_idx].rm_so, buf + *pidx,
			      naccepted) != 0)
		    return -1;
		}
	    }

	  if (naccepted == 0)
	    {
	      err = re_node_set_insert (eps_via_nodes, node);
	      if (BE (err < 0, 0))
		return -2;
	      dest_node = dfa->edests[node].elems[0];
	      if (re_node_set_contains (&mctx->state_log[*pidx]->nodes,
					dest_node))
		return dest_node;
	    }
	}

      if (naccepted != 0
	  || check_node_accept (preg, dfa->nodes + node, mctx, *pidx))
	{
	  dest_node = dfa->nexts[node];
	  *pidx = (naccepted == 0) ? *pidx + 1 : *pidx + naccepted;
	  if (fs && (*pidx > mctx->match_last || mctx->state_log[*pidx] == NULL
		     || !re_node_set_contains (&mctx->state_log[*pidx]->nodes,
					       dest_node)))
	    return -1;
	  re_node_set_empty (eps_via_nodes);
	  return dest_node;
	}
    }
  return -1;
}

static reg_errcode_t
push_fail_stack (fs, str_idx, dests, nregs, regs, eps_via_nodes)
     struct re_fail_stack_t *fs;
     int str_idx, *dests, nregs;
     regmatch_t *regs;
     re_node_set *eps_via_nodes;
{
  reg_errcode_t err;
  int num = fs->num++;
  if (fs->num == fs->alloc)
    {
      struct re_fail_stack_ent_t *new_array;
      fs->alloc *= 2;
      new_array = realloc (fs->stack, (sizeof (struct re_fail_stack_ent_t)
				       * fs->alloc));
      if (new_array == NULL)
	return REG_ESPACE;
      fs->stack = new_array;
    }
  fs->stack[num].idx = str_idx;
  fs->stack[num].node = dests[1];
  fs->stack[num].regs = re_malloc (regmatch_t, nregs);
  memcpy (fs->stack[num].regs, regs, sizeof (regmatch_t) * nregs);
  err = re_node_set_init_copy (&fs->stack[num].eps_via_nodes, eps_via_nodes);
  return err;
}

static int
pop_fail_stack (fs, pidx, nregs, regs, eps_via_nodes)
     struct re_fail_stack_t *fs;
     int *pidx, nregs;
     regmatch_t *regs;
     re_node_set *eps_via_nodes;
{
  int num = --fs->num;
  assert (num >= 0);
 *pidx = fs->stack[num].idx;
  memcpy (regs, fs->stack[num].regs, sizeof (regmatch_t) * nregs);
  re_node_set_free (eps_via_nodes);
  re_free (fs->stack[num].regs);
  *eps_via_nodes = fs->stack[num].eps_via_nodes;
  return fs->stack[num].node;
}

/* Set the positions where the subexpressions are starts/ends to registers
   PMATCH.
   Note: We assume that pmatch[0] is already set, and
   pmatch[i].rm_so == pmatch[i].rm_eo == -1 for 0 < i < nmatch.  */

static reg_errcode_t
set_regs (preg, mctx, nmatch, pmatch, fl_backtrack)
     const regex_t *preg;
     const re_match_context_t *mctx;
     size_t nmatch;
     regmatch_t *pmatch;
     int fl_backtrack;
{
  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
  int idx, cur_node, real_nmatch;
  re_node_set eps_via_nodes;
  struct re_fail_stack_t *fs;
  struct re_fail_stack_t fs_body = {0, 2, NULL};
#ifdef DEBUG
  assert (nmatch > 1);
  assert (mctx->state_log != NULL);
#endif
  if (fl_backtrack)
    {
      fs = &fs_body;
      fs->stack = re_malloc (struct re_fail_stack_ent_t, fs->alloc);
    }
  else
    fs = NULL;
  cur_node = dfa->init_node;
  real_nmatch = (nmatch <= preg->re_nsub) ? nmatch : preg->re_nsub + 1;
  re_node_set_init_empty (&eps_via_nodes);
  for (idx = pmatch[0].rm_so; idx <= pmatch[0].rm_eo ;)
    {
      update_regs (dfa, pmatch, cur_node, idx, real_nmatch);
      if (idx == pmatch[0].rm_eo && cur_node == mctx->last_node)
	{
	  int reg_idx;
	  if (fs)
	    {
	      for (reg_idx = 0; reg_idx < nmatch; ++reg_idx)
		if (pmatch[reg_idx].rm_so > -1 && pmatch[reg_idx].rm_eo == -1)
		  break;
	      if (reg_idx == nmatch)
		{
		  re_node_set_free (&eps_via_nodes);
		  return free_fail_stack_return (fs);
		}
	      cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch,
					 &eps_via_nodes);
	    }
	  else
	    {
	      re_node_set_free (&eps_via_nodes);
	      return REG_NOERROR;
	    }
	}

      /* Proceed to next node.  */
      cur_node = proceed_next_node (preg, nmatch, pmatch, mctx, &idx, cur_node,
				    &eps_via_nodes, fs);

      if (BE (cur_node < 0, 0))
	{
	  if (cur_node == -2)
	    return REG_ESPACE;
	  if (fs)
	    cur_node = pop_fail_stack (fs, &idx, nmatch, pmatch,
				       &eps_via_nodes);
	  else
	    {
	      re_node_set_free (&eps_via_nodes);
	      return REG_NOMATCH;
	    }
	}
    }
  re_node_set_free (&eps_via_nodes);
  return free_fail_stack_return (fs);
}

static reg_errcode_t
free_fail_stack_return (fs)
     struct re_fail_stack_t *fs;
{
  if (fs)
    {
      int fs_idx;
      for (fs_idx = 0; fs_idx < fs->num; ++fs_idx)
	{
	  re_node_set_free (&fs->stack[fs_idx].eps_via_nodes);
	  re_free (fs->stack[fs_idx].regs);
	}
      re_free (fs->stack);
    }
  return REG_NOERROR;
}

static void
update_regs (dfa, pmatch, cur_node, cur_idx, nmatch)
     re_dfa_t *dfa;
     regmatch_t *pmatch;
     int cur_node, cur_idx, nmatch;
{
  int type = dfa->nodes[cur_node].type;
  int reg_num;
  if (type != OP_OPEN_SUBEXP && type != OP_CLOSE_SUBEXP)
    return;
  reg_num = dfa->nodes[cur_node].opr.idx + 1;
  if (reg_num >= nmatch)
    return;
  if (type == OP_OPEN_SUBEXP)
    {
      /* We are at the first node of this sub expression.  */
      pmatch[reg_num].rm_so = cur_idx;
      pmatch[reg_num].rm_eo = -1;
    }
  else if (type == OP_CLOSE_SUBEXP)
    /* We are at the first node of this sub expression.  */
    pmatch[reg_num].rm_eo = cur_idx;
}

#define NUMBER_OF_STATE 1

/* This function checks the STATE_LOG from the SCTX->last_str_idx to 0
   and sift the nodes in each states according to the following rules.
   Updated state_log will be wrote to STATE_LOG.

   Rules: We throw away the Node `a' in the STATE_LOG[STR_IDX] if...
     1. When STR_IDX == MATCH_LAST(the last index in the state_log):
	If `a' isn't the LAST_NODE and `a' can't epsilon transit to
	the LAST_NODE, we throw away the node `a'.
     2. When 0 <= STR_IDX < MATCH_LAST and `a' accepts
	string `s' and transit to `b':
	i. If 'b' isn't in the STATE_LOG[STR_IDX+strlen('s')], we throw
	   away the node `a'.
	ii. If 'b' is in the STATE_LOG[STR_IDX+strlen('s')] but 'b' is
	    throwed away, we throw away the node `a'.
     3. When 0 <= STR_IDX < MATCH_LAST and 'a' epsilon transit to 'b':
	i. If 'b' isn't in the STATE_LOG[STR_IDX], we throw away the
	   node `a'.
	ii. If 'b' is in the STATE_LOG[STR_IDX] but 'b' is throwed away,
	    we throw away the node `a'.  */

#define STATE_NODE_CONTAINS(state,node) \
  ((state) != NULL && re_node_set_contains (&(state)->nodes, node))

static reg_errcode_t
sift_states_backward (preg, mctx, sctx)
     const regex_t *preg;
     re_match_context_t *mctx;
     re_sift_context_t *sctx;
{
  reg_errcode_t err;
  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
  int null_cnt = 0;
  int str_idx = sctx->last_str_idx;
  re_node_set cur_dest;
  re_node_set *cur_src; /* Points the state_log[str_idx]->nodes  */

#ifdef DEBUG
  assert (mctx->state_log != NULL && mctx->state_log[str_idx] != NULL);
#endif
  cur_src = &mctx->state_log[str_idx]->nodes;

  /* Build sifted state_log[str_idx].  It has the nodes which can epsilon
     transit to the last_node and the last_node itself.  */
  err = re_node_set_init_1 (&cur_dest, sctx->last_node);
  if (BE (err != REG_NOERROR, 0))
    return err;
  err = update_cur_sifted_state (preg, mctx, sctx, str_idx, &cur_dest);
  if (BE (err != REG_NOERROR, 0))
    goto free_return;

  /* Then check each states in the state_log.  */
  while (str_idx > 0)
    {
      int i, ret;
      /* Update counters.  */
      null_cnt = (sctx->sifted_states[str_idx] == NULL) ? null_cnt + 1 : 0;
      if (null_cnt > mctx->max_mb_elem_len)
	{
	  memset (sctx->sifted_states, '\0',
		  sizeof (re_dfastate_t *) * str_idx);
	  re_node_set_free (&cur_dest);
	  return REG_NOERROR;
	}
      re_node_set_empty (&cur_dest);
      --str_idx;
      cur_src = ((mctx->state_log[str_idx] == NULL) ? &empty_set
		 : &mctx->state_log[str_idx]->nodes);

      /* Then build the next sifted state.
	 We build the next sifted state on `cur_dest', and update
	 `sifted_states[str_idx]' with `cur_dest'.
	 Note:
	 `cur_dest' is the sifted state from `state_log[str_idx + 1]'.
	 `cur_src' points the node_set of the old `state_log[str_idx]'.  */
      for (i = 0; i < cur_src->nelem; i++)
	{
	  int prev_node = cur_src->elems[i];
	  int naccepted = 0;
	  re_token_type_t type = dfa->nodes[prev_node].type;

	  if (IS_EPSILON_NODE(type))
	    continue;
#ifdef RE_ENABLE_I18N
	  /* If the node may accept `multi byte'.  */
	  if (ACCEPT_MB_NODE (type))
	    naccepted = sift_states_iter_mb (preg, mctx, sctx, prev_node,
					     str_idx, sctx->last_str_idx);

#endif /* RE_ENABLE_I18N */
	  /* We don't check backreferences here.
	     See update_cur_sifted_state().  */

	  if (!naccepted
	      && check_node_accept (preg, dfa->nodes + prev_node, mctx,
				    str_idx)
	      && STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + 1],
				      dfa->nexts[prev_node]))
	    naccepted = 1;

	  if (naccepted == 0)
	    continue;

	  if (sctx->limits.nelem)
	    {
	      int to_idx = str_idx + naccepted;
	      if (check_dst_limits (dfa, &sctx->limits, mctx,
				    dfa->nexts[prev_node], to_idx,
				    prev_node, str_idx))
		continue;
	    }
	  ret = re_node_set_insert (&cur_dest, prev_node);
	  if (BE (ret == -1, 0))
	    {
	      err = REG_ESPACE;
	      goto free_return;
	    }
	}

      /* Add all the nodes which satisfy the following conditions:
	 - It can epsilon transit to a node in CUR_DEST.
	 - It is in CUR_SRC.
	 And update state_log.  */
      err = update_cur_sifted_state (preg, mctx, sctx, str_idx, &cur_dest);
      if (BE (err != REG_NOERROR, 0))
	goto free_return;
    }
  err = REG_NOERROR;
 free_return:
  re_node_set_free (&cur_dest);
  return err;
}

/* Helper functions.  */

static reg_errcode_t
clean_state_log_if_need (mctx, next_state_log_idx)
    re_match_context_t *mctx;
    int next_state_log_idx;
{
  int top = mctx->state_log_top;

  if (next_state_log_idx >= mctx->input->bufs_len
      || (next_state_log_idx >= mctx->input->valid_len
	  && mctx->input->valid_len < mctx->input->len))
    {
      reg_errcode_t err;
      err = extend_buffers (mctx);
      if (BE (err != REG_NOERROR, 0))
	return err;
    }

  if (top < next_state_log_idx)
    {
      memset (mctx->state_log + top + 1, '\0',
	      sizeof (re_dfastate_t *) * (next_state_log_idx - top));
      mctx->state_log_top = next_state_log_idx;
    }
  return REG_NOERROR;
}

static reg_errcode_t
merge_state_array (dfa, dst, src, num)
     re_dfa_t *dfa;
     re_dfastate_t **dst;
     re_dfastate_t **src;
     int num;
{
  int st_idx;
  reg_errcode_t err;
  for (st_idx = 0; st_idx < num; ++st_idx)
    {
      if (dst[st_idx] == NULL)
	dst[st_idx] = src[st_idx];
      else if (src[st_idx] != NULL)
	{
	  re_node_set merged_set;
	  err = re_node_set_init_union (&merged_set, &dst[st_idx]->nodes,
					&src[st_idx]->nodes);
	  if (BE (err != REG_NOERROR, 0))
	    return err;
	  dst[st_idx] = re_acquire_state (&err, dfa, &merged_set);
	  re_node_set_free (&merged_set);
	  if (BE (err != REG_NOERROR, 0))
	    return err;
	}
    }
  return REG_NOERROR;
}

static reg_errcode_t
update_cur_sifted_state (preg, mctx, sctx, str_idx, dest_nodes)
     const regex_t *preg;
     re_match_context_t *mctx;
     re_sift_context_t *sctx;
     int str_idx;
     re_node_set *dest_nodes;
{
  reg_errcode_t err;
  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
  const re_node_set *candidates;
  candidates = ((mctx->state_log[str_idx] == NULL) ? &empty_set
		: &mctx->state_log[str_idx]->nodes);

  /* At first, add the nodes which can epsilon transit to a node in
     DEST_NODE.  */
  if (dest_nodes->nelem)
    {
      err = add_epsilon_src_nodes (dfa, dest_nodes, candidates);
      if (BE (err != REG_NOERROR, 0))
	return err;
    }

  /* Then, check the limitations in the current sift_context.  */
  if (dest_nodes->nelem && sctx->limits.nelem)
    {
      err = check_subexp_limits (dfa, dest_nodes, candidates, &sctx->limits,
				 mctx->bkref_ents, str_idx);
      if (BE (err != REG_NOERROR, 0))
	return err;
    }

  /* Update state_log.  */
  sctx->sifted_states[str_idx] = re_acquire_state (&err, dfa, dest_nodes);
  if (BE (sctx->sifted_states[str_idx] == NULL && err != REG_NOERROR, 0))
    return err;

  if ((mctx->state_log[str_idx] != NULL
       && mctx->state_log[str_idx]->has_backref))
    {
      err = sift_states_bkref (preg, mctx, sctx, str_idx, dest_nodes);
      if (BE (err != REG_NOERROR, 0))
	return err;
    }
  return REG_NOERROR;
}

static reg_errcode_t
add_epsilon_src_nodes (dfa, dest_nodes, candidates)
     re_dfa_t *dfa;
     re_node_set *dest_nodes;
     const re_node_set *candidates;
{
  reg_errcode_t err;
  int src_idx;
  re_node_set src_copy;

  err = re_node_set_init_copy (&src_copy, dest_nodes);
  if (BE (err != REG_NOERROR, 0))
    return err;
  for (src_idx = 0; src_idx < src_copy.nelem; ++src_idx)
    {
      err = re_node_set_add_intersect (dest_nodes, candidates,
				       dfa->inveclosures
				       + src_copy.elems[src_idx]);
      if (BE (err != REG_NOERROR, 0))
	{
	  re_node_set_free (&src_copy);
	  return err;
	}
    }
  re_node_set_free (&src_copy);
  return REG_NOERROR;
}

static reg_errcode_t
sub_epsilon_src_nodes (dfa, node, dest_nodes, candidates)
     re_dfa_t *dfa;
     int node;
     re_node_set *dest_nodes;
     const re_node_set *candidates;
{
    int ecl_idx;
    reg_errcode_t err;
    re_node_set *inv_eclosure = dfa->inveclosures + node;
    re_node_set except_nodes;
    re_node_set_init_empty (&except_nodes);
    for (ecl_idx = 0; ecl_idx < inv_eclosure->nelem; ++ecl_idx)
      {
	int cur_node = inv_eclosure->elems[ecl_idx];
	if (cur_node == node)
	  continue;
	if (IS_EPSILON_NODE (dfa->nodes[cur_node].type))
	  {
	    int edst1 = dfa->edests[cur_node].elems[0];
	    int edst2 = ((dfa->edests[cur_node].nelem > 1)
			 ? dfa->edests[cur_node].elems[1] : -1);
	    if ((!re_node_set_contains (inv_eclosure, edst1)
		 && re_node_set_contains (dest_nodes, edst1))
		|| (edst2 > 0
		    && !re_node_set_contains (inv_eclosure, edst2)
		    && re_node_set_contains (dest_nodes, edst2)))
	      {
		err = re_node_set_add_intersect (&except_nodes, candidates,
						 dfa->inveclosures + cur_node);
		if (BE (err != REG_NOERROR, 0))
		  {
		    re_node_set_free (&except_nodes);
		    return err;
		  }
	      }
	  }
      }
    for (ecl_idx = 0; ecl_idx < inv_eclosure->nelem; ++ecl_idx)
      {
	int cur_node = inv_eclosure->elems[ecl_idx];
	if (!re_node_set_contains (&except_nodes, cur_node))
	  {
	    int idx = re_node_set_contains (dest_nodes, cur_node) - 1;
	    re_node_set_remove_at (dest_nodes, idx);
	  }
      }
    re_node_set_free (&except_nodes);
    return REG_NOERROR;
}

static int
check_dst_limits (dfa, limits, mctx, dst_node, dst_idx, src_node, src_idx)
     re_dfa_t *dfa;
     re_node_set *limits;
     re_match_context_t *mctx;
     int dst_node, dst_idx, src_node, src_idx;
{
  int lim_idx, src_pos, dst_pos;

  for (lim_idx = 0; lim_idx < limits->nelem; ++lim_idx)
    {
      int subexp_idx;
      struct re_backref_cache_entry *ent;
      ent = mctx->bkref_ents + limits->elems[lim_idx];
      subexp_idx = dfa->nodes[ent->node].opr.idx - 1;

      dst_pos = check_dst_limits_calc_pos (dfa, mctx, limits->elems[lim_idx],
					   dfa->eclosures + dst_node,
					   subexp_idx, dst_node, dst_idx);
      src_pos = check_dst_limits_calc_pos (dfa, mctx, limits->elems[lim_idx],
					   dfa->eclosures + src_node,
					   subexp_idx, src_node, src_idx);

      /* In case of:
	 <src> <dst> ( <subexp> )
	 ( <subexp> ) <src> <dst>
	 ( <subexp1> <src> <subexp2> <dst> <subexp3> )  */
      if (src_pos == dst_pos)
	continue; /* This is unrelated limitation.  */
      else
	return 1;
    }
  return 0;
}

static int
check_dst_limits_calc_pos (dfa, mctx, limit, eclosures, subexp_idx, from_node,
			   str_idx)
     re_dfa_t *dfa;
     re_match_context_t *mctx;
     re_node_set *eclosures;
     int limit, subexp_idx, from_node, str_idx;
{
  struct re_backref_cache_entry *lim = mctx->bkref_ents + limit;
      int node_idx;

  /* If we are outside the range of the subexpression, return -1 or 1.  */
  if (str_idx < lim->subexp_from)
    return -1;

  if (lim->subexp_to < str_idx)
    return 1;

  /* If we are within the subexpression, return 0.  */
  if (str_idx != lim->subexp_from && str_idx != lim->subexp_to)
    return 0;

  /* Else, we are on the boundary: examine the nodes on the epsilon
     closure.  */
      for (node_idx = 0; node_idx < eclosures->nelem; ++node_idx)
	{
	  int node = eclosures->elems[node_idx];
      switch (dfa->nodes[node].type)
	{
	case OP_BACK_REF:
	    {
	      int bi = search_cur_bkref_entry (mctx, str_idx);
	      for (; bi < mctx->nbkref_ents; ++bi)
		{
		  struct re_backref_cache_entry *ent = mctx->bkref_ents + bi;
		int dst, cpos;

		/* If this backreference goes beyond the point we're
		   examining, don't go any further.  */
		  if (ent->str_idx > str_idx)
		    break;

		if (ent->node != node || ent->subexp_from != ent->subexp_to)
		  continue;

		/* Recurse trying to reach the OP_OPEN_SUBEXP and
		   OP_CLOSE_SUBEXP cases below.  But, if the
		   destination node is the same node as the source
		   node, don't recurse because it would cause an
		   infinite loop: a regex that exhibits this behavior
		   is ()\1*\1*  */
		      dst = dfa->edests[node].elems[0];
		if (dst == from_node)
		  {
		    if (str_idx == lim->subexp_from)
		      return -1;
		    else /* if (str_idx == lim->subexp_to) */
		      return 0;
		  }

		      cpos = check_dst_limits_calc_pos (dfa, mctx, limit,
							dfa->eclosures + dst,
							subexp_idx, dst,
							str_idx);

		if (cpos == -1 && str_idx == lim->subexp_from)
		  return -1;

		if (cpos == 0 /* && str_idx == lim->lim->subexp_to */)
		  return 0;
	    }
	      break;
	    }

	case OP_OPEN_SUBEXP:
	  if (str_idx == lim->subexp_from && subexp_idx == dfa->nodes[node].opr.idx)
	    return -1;
	  break;

	case OP_CLOSE_SUBEXP:
	  if (str_idx == lim->subexp_to && subexp_idx == dfa->nodes[node].opr.idx)
	    return 0;
	  break;

	default:
	    break;
	}
    }

  if (str_idx == lim->subexp_to)
    return 1;
  else
    return 0;
}

/* Check the limitations of sub expressions LIMITS, and remove the nodes
   which are against limitations from DEST_NODES. */

static reg_errcode_t
check_subexp_limits (dfa, dest_nodes, candidates, limits, bkref_ents, str_idx)
     re_dfa_t *dfa;
     re_node_set *dest_nodes;
     const re_node_set *candidates;
     re_node_set *limits;
     struct re_backref_cache_entry *bkref_ents;
     int str_idx;
{
  reg_errcode_t err;
  int node_idx, lim_idx;

  for (lim_idx = 0; lim_idx < limits->nelem; ++lim_idx)
    {
      int subexp_idx;
      struct re_backref_cache_entry *ent;
      ent = bkref_ents + limits->elems[lim_idx];

      if (str_idx <= ent->subexp_from || ent->str_idx < str_idx)
	continue; /* This is unrelated limitation.  */

      subexp_idx = dfa->nodes[ent->node].opr.idx - 1;
      if (ent->subexp_to == str_idx)
	{
	  int ops_node = -1;
	  int cls_node = -1;
	  for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx)
	    {
	      int node = dest_nodes->elems[node_idx];
	      re_token_type_t type= dfa->nodes[node].type;
	      if (type == OP_OPEN_SUBEXP
		  && subexp_idx == dfa->nodes[node].opr.idx)
		ops_node = node;
	      else if (type == OP_CLOSE_SUBEXP
		       && subexp_idx == dfa->nodes[node].opr.idx)
		cls_node = node;
	    }

	  /* Check the limitation of the open subexpression.  */
	  /* Note that (ent->subexp_to = str_idx != ent->subexp_from).  */
	  if (ops_node >= 0)
	    {
	      err = sub_epsilon_src_nodes(dfa, ops_node, dest_nodes,
					  candidates);
	      if (BE (err != REG_NOERROR, 0))
		return err;
	    }
	  /* Check the limitation of the close subexpression.  */
	  for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx)
	    {
	      int node = dest_nodes->elems[node_idx];
	      if (!re_node_set_contains (dfa->inveclosures + node, cls_node)
		  && !re_node_set_contains (dfa->eclosures + node, cls_node))
		{
		  /* It is against this limitation.
		     Remove it form the current sifted state.  */
		  err = sub_epsilon_src_nodes(dfa, node, dest_nodes,
					      candidates);
		  if (BE (err != REG_NOERROR, 0))
		    return err;
		  --node_idx;
		}
	    }
	}
      else /* (ent->subexp_to != str_idx)  */
	{
	  for (node_idx = 0; node_idx < dest_nodes->nelem; ++node_idx)
	    {
	      int node = dest_nodes->elems[node_idx];
	      re_token_type_t type= dfa->nodes[node].type;
	      if (type == OP_CLOSE_SUBEXP || type == OP_OPEN_SUBEXP)
		{
		  if (subexp_idx != dfa->nodes[node].opr.idx)
		    continue;
		  if ((type == OP_CLOSE_SUBEXP && ent->subexp_to != str_idx)
		      || (type == OP_OPEN_SUBEXP))
		    {
		      /* It is against this limitation.
			 Remove it form the current sifted state.  */
		      err = sub_epsilon_src_nodes(dfa, node, dest_nodes,
						  candidates);
		      if (BE (err != REG_NOERROR, 0))
			return err;
		    }
		}
	    }
	}
    }
  return REG_NOERROR;
}

static reg_errcode_t
sift_states_bkref (preg, mctx, sctx, str_idx, dest_nodes)
     const regex_t *preg;
     re_match_context_t *mctx;
     re_sift_context_t *sctx;
     int str_idx;
     re_node_set *dest_nodes;
{
  reg_errcode_t err;
  re_dfa_t *dfa = (re_dfa_t *)preg->buffer;
  int node_idx, node;
  re_sift_context_t local_sctx;
  const re_node_set *candidates;
  candidates = ((mctx->state_log[str_idx] == NULL) ? &empty_set
		: &mctx->state_log[str_idx]->nodes);
  local_sctx.sifted_states = NULL; /* Mark that it hasn't been initialized.  */

  for (node_idx = 0; node_idx < candidates->nelem; ++node_idx)
    {
      int cur_bkref_idx = re_string_cur_idx (mctx->input);
      re_token_type_t type;
      node = candidates->elems[node_idx];
      type = dfa->nodes[node].type;
      if (node == sctx->cur_bkref && str_idx == cur_bkref_idx)
	continue;
      /* Avoid infinite loop for the REs like "()\1+".  */
      if (node == sctx->last_node && str_idx == sctx->last_str_idx)
	continue;
      if (type == OP_BACK_REF)
	{
	  int enabled_idx = search_cur_bkref_entry (mctx, str_idx);
	  for (; enabled_idx < mctx->nbkref_ents; ++enabled_idx)
	    {
	      int disabled_idx, subexp_len, to_idx, dst_node;
	      struct re_backref_cache_entry *entry;
	      entry = mctx->bkref_ents + enabled_idx;
	      if (entry->str_idx > str_idx)
		break;
	      if (entry->node != node)
		  continue;
	      subexp_len = entry->subexp_to - entry->subexp_from;
	      to_idx = str_idx + subexp_len;
	      dst_node = (subexp_len ? dfa->nexts[node]
			  : dfa->edests[node].elems[0]);

	      if (to_idx > sctx->last_str_idx
		  || sctx->sifted_states[to_idx] == NULL
		  || !STATE_NODE_CONTAINS (sctx->sifted_states[to_idx],
					   dst_node)
		  || check_dst_limits (dfa, &sctx->limits, mctx, node,
				       str_idx, dst_node, to_idx))
		continue;
		{
		  re_dfastate_t *cur_state;
		  entry->flag = 0;
		  for (disabled_idx = enabled_idx + 1;
		       disabled_idx < mctx->nbkref_ents; ++disabled_idx)
		    {
		      struct re_backref_cache_entry *entry2;
		      entry2 = mctx->bkref_ents + disabled_idx;
		      if (entry2->str_idx > str_idx)
			break;
		      entry2->flag = (entry2->node == node) ? 1 : entry2->flag;
		    }

		  if (local_sctx.sifted_states == NULL)
		    {
		      local_sctx = *sctx;
		      err = re_node_set_init_copy (&local_sctx.limits,
						   &sctx->limits);
		      if (BE (err != REG_NOERROR, 0))
			goto free_return;
		    }
		  local_sctx.last_node = node;
		  local_sctx.last_str_idx = str_idx;
		  err = re_node_set_insert (&local_sctx.limits, enabled_idx);
		  if (BE (err < 0, 0))
		    {
		      err = REG_ESPACE;
		      goto free_return;
		    }
		  cur_state = local_sctx.sifted_states[str_idx];
		  err = sift_states_backward (preg, mctx, &local_sctx);
		  if (BE (err != REG_NOERROR, 0))
		    goto free_return;
		  if (sctx->limited_states != NULL)
		    {
		      err = merge_state_array (dfa, sctx->limited_states,
					       local_sctx.sifted_states,
					       str_idx + 1);
		      if (BE (err != REG_NOERROR, 0))
			goto free_return;
		    }
		  local_sctx.sifted_states[str_idx] = cur_state;
		  re_node_set_remove (&local_sctx.limits, enabled_idx);
		  /* We must not use the variable entry here, since
		     mctx->bkref_ents might be realloced.  */
		  mctx->bkref_ents[enabled_idx].flag = 1;
		}
	    }
	  enabled_idx = search_cur_bkref_entry (mctx, str_idx);
	  for (; enabled_idx < mctx->nbkref_ents; ++enabled_idx)
	    {
	      struct re_backref_cache_entry *entry;
	      entry = mctx->bkref_ents + enabled_idx;
	      if (entry->str_idx > str_idx)
		break;
	      if (entry->node == node)
		entry->flag = 0;
	    }
	}
    }
  err = REG_NOERROR;
 free_return:
  if (local_sctx.sifted_states != NULL)
    {
      re_node_set_free (&local_sctx.limits);
    }

  return err;
}


#ifdef RE_ENABLE_I18N
static int
sift_states_iter_mb (preg, mctx, sctx, node_idx, str_idx, max_str_idx)
    const regex_t *preg;
    const re_match_context_t *mctx;
    re_sift_context_t *sctx;
    int node_idx, str_idx, max_str_idx;
{
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  int naccepted;
  /* Check the node can accept `multi byte'.  */
  naccepted = check_node_accept_bytes (preg, node_idx, mctx->input, str_idx);
  if (naccepted > 0 && str_idx + naccepted <= max_str_idx &&
      !STATE_NODE_CONTAINS (sctx->sifted_states[str_idx + naccepted],
			    dfa->nexts[node_idx]))
    /* The node can't accept the `multi byte', or the
       destination was already throwed away, then the node
       could't accept the current input `multi byte'.   */
    naccepted = 0;
  /* Otherwise, it is sure that the node could accept
     `naccepted' bytes input.  */
  return naccepted;
}
#endif /* RE_ENABLE_I18N */


/* Functions for state transition.  */

/* Return the next state to which the current state STATE will transit by
   accepting the current input byte, and update STATE_LOG if necessary.
   If STATE can accept a multibyte char/collating element/back reference
   update the destination of STATE_LOG.  */

static re_dfastate_t *
transit_state (err, preg, mctx, state)
     reg_errcode_t *err;
     const regex_t *preg;
     re_match_context_t *mctx;
     re_dfastate_t *state;
{
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  re_dfastate_t **trtable, *next_state;
  unsigned char ch;
  int cur_idx;

  if (re_string_cur_idx (mctx->input) + 1 >= mctx->input->bufs_len
      || (re_string_cur_idx (mctx->input) + 1 >= mctx->input->valid_len
	  && mctx->input->valid_len < mctx->input->len))
    {
      *err = extend_buffers (mctx);
      if (BE (*err != REG_NOERROR, 0))
	return NULL;
    }

  *err = REG_NOERROR;
  if (state == NULL)
    {
      next_state = state;
      re_string_skip_bytes (mctx->input, 1);
    }
  else
    {
#ifdef RE_ENABLE_I18N
      /* If the current state can accept multibyte.  */
      if (state->accept_mb)
	{
	  *err = transit_state_mb (preg, state, mctx);
	  if (BE (*err != REG_NOERROR, 0))
	    return NULL;
	}
#endif /* RE_ENABLE_I18N */

      /* Then decide the next state with the single byte.  */
      if (1)
	{
	  /* Use transition table  */
	  ch = re_string_fetch_byte (mctx->input);
	  trtable = state->trtable;
	  if (trtable == NULL)
	    {
	      trtable = build_trtable (preg, state);
	      if (trtable == NULL)
		{
		  *err = REG_ESPACE;
		  return NULL;
		}
	    }
	  if (BE (state->word_trtable, 0))
	    {
	      unsigned int context;
	      context
		= re_string_context_at (mctx->input,
					re_string_cur_idx (mctx->input) - 1,
					mctx->eflags, preg->newline_anchor);
	      if (IS_WORD_CONTEXT (context))
		next_state = trtable[ch + SBC_MAX];
	      else
		next_state = trtable[ch];
	    }
	  else
	    next_state = trtable[ch];
	}
#if 0
      else
	{
	  /* don't use transition table  */
	  next_state = transit_state_sb (err, preg, state, mctx);
	  if (BE (next_state == NULL && err != REG_NOERROR, 0))
	    return NULL;
	}
#endif
    }

  cur_idx = re_string_cur_idx (mctx->input);
  /* Update the state_log if we need.  */
  if (mctx->state_log != NULL)
    {
      if (cur_idx > mctx->state_log_top)
	{
	  mctx->state_log[cur_idx] = next_state;
	  mctx->state_log_top = cur_idx;
	}
      else if (mctx->state_log[cur_idx] == 0)
	{
	  mctx->state_log[cur_idx] = next_state;
	}
      else
	{
	  re_dfastate_t *pstate;
	  unsigned int context;
	  re_node_set next_nodes, *log_nodes, *table_nodes = NULL;
	  /* If (state_log[cur_idx] != 0), it implies that cur_idx is
	     the destination of a multibyte char/collating element/
	     back reference.  Then the next state is the union set of
	     these destinations and the results of the transition table.  */
	  pstate = mctx->state_log[cur_idx];
	  log_nodes = pstate->entrance_nodes;
	  if (next_state != NULL)
	    {
	      table_nodes = next_state->entrance_nodes;
	      *err = re_node_set_init_union (&next_nodes, table_nodes,
					     log_nodes);
	      if (BE (*err != REG_NOERROR, 0))
		return NULL;
	    }
	  else
	    next_nodes = *log_nodes;
	  /* Note: We already add the nodes of the initial state,
		   then we don't need to add them here.  */

	  context = re_string_context_at (mctx->input,
					  re_string_cur_idx (mctx->input) - 1,
					  mctx->eflags, preg->newline_anchor);
	  next_state = mctx->state_log[cur_idx]
	    = re_acquire_state_context (err, dfa, &next_nodes, context);
	  /* We don't need to check errors here, since the return value of
	     this function is next_state and ERR is already set.  */

	  if (table_nodes != NULL)
	    re_node_set_free (&next_nodes);
	}
    }

  /* Check OP_OPEN_SUBEXP in the current state in case that we use them
     later.  We must check them here, since the back references in the
     next state might use them.  */
  if (dfa->nbackref && next_state/* && fl_process_bkref */)
    {
      *err = check_subexp_matching_top (dfa, mctx, &next_state->nodes,
					cur_idx);
      if (BE (*err != REG_NOERROR, 0))
	return NULL;
    }

  /* If the next state has back references.  */
  if (next_state != NULL && next_state->has_backref)
    {
      *err = transit_state_bkref (preg, &next_state->nodes, mctx);
      if (BE (*err != REG_NOERROR, 0))
	return NULL;
      next_state = mctx->state_log[cur_idx];
    }
  return next_state;
}

/* Helper functions for transit_state.  */

/* From the node set CUR_NODES, pick up the nodes whose types are
   OP_OPEN_SUBEXP and which have corresponding back references in the regular
   expression. And register them to use them later for evaluating the
   correspoding back references.  */

static reg_errcode_t
check_subexp_matching_top (dfa, mctx, cur_nodes, str_idx)
     re_dfa_t *dfa;
     re_match_context_t *mctx;
     re_node_set *cur_nodes;
     int str_idx;
{
  int node_idx;
  reg_errcode_t err;

  /* TODO: This isn't efficient.
	   Because there might be more than one nodes whose types are
	   OP_OPEN_SUBEXP and whose index is SUBEXP_IDX, we must check all
	   nodes.
	   E.g. RE: (a){2}  */
  for (node_idx = 0; node_idx < cur_nodes->nelem; ++node_idx)
    {
      int node = cur_nodes->elems[node_idx];
      if (dfa->nodes[node].type == OP_OPEN_SUBEXP
	  && dfa->nodes[node].opr.idx < (8 * sizeof (dfa->used_bkref_map))
	  && dfa->used_bkref_map & (1 << dfa->nodes[node].opr.idx))
	{
	  err = match_ctx_add_subtop (mctx, node, str_idx);
	  if (BE (err != REG_NOERROR, 0))
	    return err;
	}
    }
  return REG_NOERROR;
}

#if 0
/* Return the next state to which the current state STATE will transit by
   accepting the current input byte.  */

static re_dfastate_t *
transit_state_sb (err, preg, state, mctx)
     reg_errcode_t *err;
     const regex_t *preg;
     re_dfastate_t *state;
     re_match_context_t *mctx;
{
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  re_node_set next_nodes;
  re_dfastate_t *next_state;
  int node_cnt, cur_str_idx = re_string_cur_idx (mctx->input);
  unsigned int context;

  *err = re_node_set_alloc (&next_nodes, state->nodes.nelem + 1);
  if (BE (*err != REG_NOERROR, 0))
    return NULL;
  for (node_cnt = 0; node_cnt < state->nodes.nelem; ++node_cnt)
    {
      int cur_node = state->nodes.elems[node_cnt];
      if (check_node_accept (preg, dfa->nodes + cur_node, mctx, cur_str_idx))
	{
	  *err = re_node_set_merge (&next_nodes,
				    dfa->eclosures + dfa->nexts[cur_node]);
	  if (BE (*err != REG_NOERROR, 0))
	    {
	      re_node_set_free (&next_nodes);
	      return NULL;
	    }
	}
    }
  context = re_string_context_at (mctx->input, cur_str_idx, mctx->eflags,
				  preg->newline_anchor);
  next_state = re_acquire_state_context (err, dfa, &next_nodes, context);
  /* We don't need to check errors here, since the return value of
     this function is next_state and ERR is already set.  */

  re_node_set_free (&next_nodes);
  re_string_skip_bytes (mctx->input, 1);
  return next_state;
}
#endif

#ifdef RE_ENABLE_I18N
static reg_errcode_t
transit_state_mb (preg, pstate, mctx)
    const regex_t *preg;
    re_dfastate_t *pstate;
    re_match_context_t *mctx;
{
  reg_errcode_t err;
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  int i;

  for (i = 0; i < pstate->nodes.nelem; ++i)
    {
      re_node_set dest_nodes, *new_nodes;
      int cur_node_idx = pstate->nodes.elems[i];
      int naccepted = 0, dest_idx;
      unsigned int context;
      re_dfastate_t *dest_state;

      if (dfa->nodes[cur_node_idx].constraint)
	{
	  context = re_string_context_at (mctx->input,
					  re_string_cur_idx (mctx->input),
					  mctx->eflags, preg->newline_anchor);
	  if (NOT_SATISFY_NEXT_CONSTRAINT (dfa->nodes[cur_node_idx].constraint,
					   context))
	    continue;
	}

      /* How many bytes the node can accept?  */
      if (ACCEPT_MB_NODE (dfa->nodes[cur_node_idx].type))
	naccepted = check_node_accept_bytes (preg, cur_node_idx, mctx->input,
					     re_string_cur_idx (mctx->input));
      if (naccepted == 0)
	continue;

      /* The node can accepts `naccepted' bytes.  */
      dest_idx = re_string_cur_idx (mctx->input) + naccepted;
      mctx->max_mb_elem_len = ((mctx->max_mb_elem_len < naccepted) ? naccepted
			       : mctx->max_mb_elem_len);
      err = clean_state_log_if_need (mctx, dest_idx);
      if (BE (err != REG_NOERROR, 0))
	return err;
#ifdef DEBUG
      assert (dfa->nexts[cur_node_idx] != -1);
#endif
      /* `cur_node_idx' may point the entity of the OP_CONTEXT_NODE,
	 then we use pstate->nodes.elems[i] instead.  */
      new_nodes = dfa->eclosures + dfa->nexts[pstate->nodes.elems[i]];

      dest_state = mctx->state_log[dest_idx];
      if (dest_state == NULL)
	dest_nodes = *new_nodes;
      else
	{
	  err = re_node_set_init_union (&dest_nodes,
					dest_state->entrance_nodes, new_nodes);
	  if (BE (err != REG_NOERROR, 0))
	    return err;
	}
      context = re_string_context_at (mctx->input, dest_idx - 1, mctx->eflags,
				      preg->newline_anchor);
      mctx->state_log[dest_idx]
	= re_acquire_state_context (&err, dfa, &dest_nodes, context);
      if (dest_state != NULL)
	re_node_set_free (&dest_nodes);
      if (BE (mctx->state_log[dest_idx] == NULL && err != REG_NOERROR, 0))
	return err;
    }
  return REG_NOERROR;
}
#endif /* RE_ENABLE_I18N */

static reg_errcode_t
transit_state_bkref (preg, nodes, mctx)
    const regex_t *preg;
    const re_node_set *nodes;
    re_match_context_t *mctx;
{
  reg_errcode_t err;
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  int i;
  int cur_str_idx = re_string_cur_idx (mctx->input);

  for (i = 0; i < nodes->nelem; ++i)
    {
      int dest_str_idx, prev_nelem, bkc_idx;
      int node_idx = nodes->elems[i];
      unsigned int context;
      const re_token_t *node = dfa->nodes + node_idx;
      re_node_set *new_dest_nodes;

      /* Check whether `node' is a backreference or not.  */
      if (node->type != OP_BACK_REF)
	continue;

      if (node->constraint)
	{
	  context = re_string_context_at (mctx->input, cur_str_idx,
					  mctx->eflags, preg->newline_anchor);
	  if (NOT_SATISFY_NEXT_CONSTRAINT (node->constraint, context))
	    continue;
	}

      /* `node' is a backreference.
	 Check the substring which the substring matched.  */
      bkc_idx = mctx->nbkref_ents;
      err = get_subexp (preg, mctx, node_idx, cur_str_idx);
      if (BE (err != REG_NOERROR, 0))
	goto free_return;

      /* And add the epsilon closures (which is `new_dest_nodes') of
	 the backreference to appropriate state_log.  */
#ifdef DEBUG
      assert (dfa->nexts[node_idx] != -1);
#endif
      for (; bkc_idx < mctx->nbkref_ents; ++bkc_idx)
	{
	  int subexp_len;
	  re_dfastate_t *dest_state;
	  struct re_backref_cache_entry *bkref_ent;
	  bkref_ent = mctx->bkref_ents + bkc_idx;
	  if (bkref_ent->node != node_idx || bkref_ent->str_idx != cur_str_idx)
	    continue;
	  subexp_len = bkref_ent->subexp_to - bkref_ent->subexp_from;
	  new_dest_nodes = (subexp_len == 0
			    ? dfa->eclosures + dfa->edests[node_idx].elems[0]
			    : dfa->eclosures + dfa->nexts[node_idx]);
	  dest_str_idx = (cur_str_idx + bkref_ent->subexp_to
			  - bkref_ent->subexp_from);
	  context = re_string_context_at (mctx->input, dest_str_idx - 1,
					  mctx->eflags, preg->newline_anchor);
	  dest_state = mctx->state_log[dest_str_idx];
	  prev_nelem = ((mctx->state_log[cur_str_idx] == NULL) ? 0
			: mctx->state_log[cur_str_idx]->nodes.nelem);
	  /* Add `new_dest_node' to state_log.  */
	  if (dest_state == NULL)
	    {
	      mctx->state_log[dest_str_idx]
		= re_acquire_state_context (&err, dfa, new_dest_nodes,
					    context);
	      if (BE (mctx->state_log[dest_str_idx] == NULL
		      && err != REG_NOERROR, 0))
		goto free_return;
	    }
	  else
	    {
	      re_node_set dest_nodes;
	      err = re_node_set_init_union (&dest_nodes,
					    dest_state->entrance_nodes,
					    new_dest_nodes);
	      if (BE (err != REG_NOERROR, 0))
		{
		  re_node_set_free (&dest_nodes);
		  goto free_return;
		}
	      mctx->state_log[dest_str_idx]
		= re_acquire_state_context (&err, dfa, &dest_nodes, context);
	      re_node_set_free (&dest_nodes);
	      if (BE (mctx->state_log[dest_str_idx] == NULL
		      && err != REG_NOERROR, 0))
		goto free_return;
	    }
	  /* We need to check recursively if the backreference can epsilon
	     transit.  */
	  if (subexp_len == 0
	      && mctx->state_log[cur_str_idx]->nodes.nelem > prev_nelem)
	    {
	      err = check_subexp_matching_top (dfa, mctx, new_dest_nodes,
					       cur_str_idx);
	      if (BE (err != REG_NOERROR, 0))
		goto free_return;
	      err = transit_state_bkref (preg, new_dest_nodes, mctx);
	      if (BE (err != REG_NOERROR, 0))
		goto free_return;
	    }
	}
    }
  err = REG_NOERROR;
 free_return:
  return err;
}

/* Enumerate all the candidates which the backreference BKREF_NODE can match
   at BKREF_STR_IDX, and register them by match_ctx_add_entry().
   Note that we might collect inappropriate candidates here.
   However, the cost of checking them strictly here is too high, then we
   delay these checking for prune_impossible_nodes().  */

static reg_errcode_t
get_subexp (preg, mctx, bkref_node, bkref_str_idx)
     const regex_t *preg;
     re_match_context_t *mctx;
     int bkref_node, bkref_str_idx;
{
  int subexp_num, sub_top_idx;
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  const char *buf = (const char *) re_string_get_buffer (mctx->input);
  /* Return if we have already checked BKREF_NODE at BKREF_STR_IDX.  */
  int cache_idx = search_cur_bkref_entry (mctx, bkref_str_idx);
  for (; cache_idx < mctx->nbkref_ents; ++cache_idx)
    {
      const struct re_backref_cache_entry *entry
	= &mctx->bkref_ents[cache_idx];
      if (entry->str_idx > bkref_str_idx)
	break;
      if (entry->node == bkref_node)
	return REG_NOERROR; /* We already checked it.  */
    }
  subexp_num = dfa->nodes[bkref_node].opr.idx - 1;

  /* For each sub expression  */
  for (sub_top_idx = 0; sub_top_idx < mctx->nsub_tops; ++sub_top_idx)
    {
      reg_errcode_t err;
      re_sub_match_top_t *sub_top = mctx->sub_tops[sub_top_idx];
      re_sub_match_last_t *sub_last;
      int sub_last_idx, sl_str;
      const char *bkref_str;

      if (dfa->nodes[sub_top->node].opr.idx != subexp_num)
	continue; /* It isn't related.  */

      sl_str = sub_top->str_idx;
      bkref_str = buf + bkref_str_idx;
      /* At first, check the last node of sub expressions we already
	 evaluated.  */
      for (sub_last_idx = 0; sub_last_idx < sub_top->nlasts; ++sub_last_idx)
	{
	  int sl_str_diff;
	  sub_last = sub_top->lasts[sub_last_idx];
	  sl_str_diff = sub_last->str_idx - sl_str;
	  /* The matched string by the sub expression match with the substring
	     at the back reference?  */
	  if (sl_str_diff > 0
	      && memcmp (bkref_str, buf + sl_str, sl_str_diff) != 0)
	    break; /* We don't need to search this sub expression any more.  */
	  bkref_str += sl_str_diff;
	  sl_str += sl_str_diff;
	  err = get_subexp_sub (preg, mctx, sub_top, sub_last, bkref_node,
				bkref_str_idx);

	  /* Reload buf and bkref_str, since the preceding call might
	     have reallocated the buffer.  */
	  buf = (const char *) re_string_get_buffer (mctx->input);
	  bkref_str = buf + bkref_str_idx;

	  if (err == REG_NOMATCH)
	    continue;
	  if (BE (err != REG_NOERROR, 0))
	    return err;
	}
      if (sub_last_idx < sub_top->nlasts)
	continue;
      if (sub_last_idx > 0)
	++sl_str;
      /* Then, search for the other last nodes of the sub expression.  */
      for (; sl_str <= bkref_str_idx; ++sl_str)
	{
	  int cls_node, sl_str_off;
	  const re_node_set *nodes;
	  sl_str_off = sl_str - sub_top->str_idx;
	  /* The matched string by the sub expression match with the substring
	     at the back reference?  */
	  if (sl_str_off > 0 && *bkref_str++ != buf[sl_str - 1])
	    break; /* We don't need to search this sub expression any more.  */
	  if (mctx->state_log[sl_str] == NULL)
	    continue;
	  /* Does this state have a ')' of the sub expression?  */
	  nodes = &mctx->state_log[sl_str]->nodes;
	  cls_node = find_subexp_node (dfa, nodes, subexp_num, 0);
	  if (cls_node == -1)
	    continue; /* No.  */
	  if (sub_top->path == NULL)
	    {
	      sub_top->path = calloc (sizeof (state_array_t),
				      sl_str - sub_top->str_idx + 1);
	      if (sub_top->path == NULL)
		return REG_ESPACE;
	    }
	  /* Can the OP_OPEN_SUBEXP node arrive the OP_CLOSE_SUBEXP node
	     in the current context?  */
	  err = check_arrival (preg, mctx, sub_top->path, sub_top->node,
			       sub_top->str_idx, cls_node, sl_str, 0);
	  if (err == REG_NOMATCH)
	      continue;
	  if (BE (err != REG_NOERROR, 0))
	      return err;
	  sub_last = match_ctx_add_sublast (sub_top, cls_node, sl_str);
	  if (BE (sub_last == NULL, 0))
	    return REG_ESPACE;
	  err = get_subexp_sub (preg, mctx, sub_top, sub_last, bkref_node,
				bkref_str_idx);
	  if (err == REG_NOMATCH)
	    continue;
	}
    }
  return REG_NOERROR;
}

/* Helper functions for get_subexp().  */

/* Check SUB_LAST can arrive to the back reference BKREF_NODE at BKREF_STR.
   If it can arrive, register the sub expression expressed with SUB_TOP
   and SUB_LAST.  */

static reg_errcode_t
get_subexp_sub (preg, mctx, sub_top, sub_last, bkref_node, bkref_str)
     const regex_t *preg;
     re_match_context_t *mctx;
     const re_sub_match_top_t *sub_top;
     re_sub_match_last_t *sub_last;
     int bkref_node, bkref_str;
{
  reg_errcode_t err;
  int to_idx;
  /* Can the subexpression arrive the back reference?  */
  err = check_arrival (preg, mctx, &sub_last->path, sub_last->node,
		       sub_last->str_idx, bkref_node, bkref_str, 1);
  if (err != REG_NOERROR)
    return err;
  err = match_ctx_add_entry (mctx, bkref_node, bkref_str, sub_top->str_idx,
			     sub_last->str_idx);
  if (BE (err != REG_NOERROR, 0))
    return err;
  to_idx = bkref_str + sub_last->str_idx - sub_top->str_idx;
  clean_state_log_if_need (mctx, to_idx);
  return REG_NOERROR;
}

/* Find the first node which is '(' or ')' and whose index is SUBEXP_IDX.
   Search '(' if FL_OPEN, or search ')' otherwise.
   TODO: This function isn't efficient...
	 Because there might be more than one nodes whose types are
	 OP_OPEN_SUBEXP and whose index is SUBEXP_IDX, we must check all
	 nodes.
	 E.g. RE: (a){2}  */

static int
find_subexp_node (dfa, nodes, subexp_idx, fl_open)
     const re_dfa_t *dfa;
     const re_node_set *nodes;
     int subexp_idx, fl_open;
{
  int cls_idx;
  for (cls_idx = 0; cls_idx < nodes->nelem; ++cls_idx)
    {
      int cls_node = nodes->elems[cls_idx];
      const re_token_t *node = dfa->nodes + cls_node;
      if (((fl_open && node->type == OP_OPEN_SUBEXP)
	  || (!fl_open && node->type == OP_CLOSE_SUBEXP))
	  && node->opr.idx == subexp_idx)
	return cls_node;
    }
  return -1;
}

/* Check whether the node TOP_NODE at TOP_STR can arrive to the node
   LAST_NODE at LAST_STR.  We record the path onto PATH since it will be
   heavily reused.
   Return REG_NOERROR if it can arrive, or REG_NOMATCH otherwise.  */

static reg_errcode_t
check_arrival (preg, mctx, path, top_node, top_str, last_node, last_str,
	       fl_open)
     const regex_t *preg;
     re_match_context_t *mctx;
     state_array_t *path;
     int top_node, top_str, last_node, last_str, fl_open;
{
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  reg_errcode_t err;
  int subexp_num, backup_cur_idx, str_idx, null_cnt;
  re_dfastate_t *cur_state = NULL;
  re_node_set *cur_nodes, next_nodes;
  re_dfastate_t **backup_state_log;
  unsigned int context;

  subexp_num = dfa->nodes[top_node].opr.idx;
  /* Extend the buffer if we need.  */
  if (BE (path->alloc < last_str + mctx->max_mb_elem_len + 1, 0))
    {
      re_dfastate_t **new_array;
      int old_alloc = path->alloc;
      path->alloc += last_str + mctx->max_mb_elem_len + 1;
      new_array = re_realloc (path->array, re_dfastate_t *, path->alloc);
      if (new_array == NULL)
	{
	  path->alloc = old_alloc;
	  return REG_ESPACE;
	}
      path->array = new_array;
      memset (new_array + old_alloc, '\0',
	      sizeof (re_dfastate_t *) * (path->alloc - old_alloc));
    }

  str_idx = path->next_idx == 0 ? top_str : path->next_idx;

  /* Temporary modify MCTX.  */
  backup_state_log = mctx->state_log;
  backup_cur_idx = mctx->input->cur_idx;
  mctx->state_log = path->array;
  mctx->input->cur_idx = str_idx;

  /* Setup initial node set.  */
  context = re_string_context_at (mctx->input, str_idx - 1, mctx->eflags,
				  preg->newline_anchor);
  if (str_idx == top_str)
    {
      err = re_node_set_init_1 (&next_nodes, top_node);
      if (BE (err != REG_NOERROR, 0))
	return err;
      err = check_arrival_expand_ecl (dfa, &next_nodes, subexp_num, fl_open);
      if (BE (err != REG_NOERROR, 0))
	{
	  re_node_set_free (&next_nodes);
	  return err;
	}
    }
  else
    {
      cur_state = mctx->state_log[str_idx];
      if (cur_state && cur_state->has_backref)
	{
	  err = re_node_set_init_copy (&next_nodes, &cur_state->nodes);
	  if (BE ( err != REG_NOERROR, 0))
	    return err;
	}
      else
	re_node_set_init_empty (&next_nodes);
    }
  if (str_idx == top_str || (cur_state && cur_state->has_backref))
    {
      if (next_nodes.nelem)
	{
	  err = expand_bkref_cache (preg, mctx, &next_nodes, str_idx, last_str,
				    subexp_num, fl_open);
	  if (BE ( err != REG_NOERROR, 0))
	    {
	      re_node_set_free (&next_nodes);
	      return err;
	    }
	}
      cur_state = re_acquire_state_context (&err, dfa, &next_nodes, context);
      if (BE (cur_state == NULL && err != REG_NOERROR, 0))
	{
	  re_node_set_free (&next_nodes);
	  return err;
	}
      mctx->state_log[str_idx] = cur_state;
    }

  for (null_cnt = 0; str_idx < last_str && null_cnt <= mctx->max_mb_elem_len;)
    {
      re_node_set_empty (&next_nodes);
      if (mctx->state_log[str_idx + 1])
	{
	  err = re_node_set_merge (&next_nodes,
				   &mctx->state_log[str_idx + 1]->nodes);
	  if (BE (err != REG_NOERROR, 0))
	    {
	      re_node_set_free (&next_nodes);
	      return err;
	    }
	}
      if (cur_state)
	{
	  err = check_arrival_add_next_nodes(preg, dfa, mctx, str_idx,
					     &cur_state->nodes, &next_nodes);
	  if (BE (err != REG_NOERROR, 0))
	    {
	      re_node_set_free (&next_nodes);
	      return err;
	    }
	}
      ++str_idx;
      if (next_nodes.nelem)
	{
	  err = check_arrival_expand_ecl (dfa, &next_nodes, subexp_num,
					  fl_open);
	  if (BE (err != REG_NOERROR, 0))
	    {
	      re_node_set_free (&next_nodes);
	      return err;
	    }
	  err = expand_bkref_cache (preg, mctx, &next_nodes, str_idx, last_str,
				    subexp_num, fl_open);
	  if (BE ( err != REG_NOERROR, 0))
	    {
	      re_node_set_free (&next_nodes);
	      return err;
	    }
	}
      context = re_string_context_at (mctx->input, str_idx - 1, mctx->eflags,
				      preg->newline_anchor);
      cur_state = re_acquire_state_context (&err, dfa, &next_nodes, context);
      if (BE (cur_state == NULL && err != REG_NOERROR, 0))
	{
	  re_node_set_free (&next_nodes);
	  return err;
	}
      mctx->state_log[str_idx] = cur_state;
      null_cnt = cur_state == NULL ? null_cnt + 1 : 0;
    }
  re_node_set_free (&next_nodes);
  cur_nodes = (mctx->state_log[last_str] == NULL ? NULL
	       : &mctx->state_log[last_str]->nodes);
  path->next_idx = str_idx;

  /* Fix MCTX.  */
  mctx->state_log = backup_state_log;
  mctx->input->cur_idx = backup_cur_idx;

  if (cur_nodes == NULL)
    return REG_NOMATCH;
  /* Then check the current node set has the node LAST_NODE.  */
  return (re_node_set_contains (cur_nodes, last_node)
	  || re_node_set_contains (cur_nodes, last_node) ? REG_NOERROR
	  : REG_NOMATCH);
}

/* Helper functions for check_arrival.  */

/* Calculate the destination nodes of CUR_NODES at STR_IDX, and append them
   to NEXT_NODES.
   TODO: This function is similar to the functions transit_state*(),
	 however this function has many additional works.
	 Can't we unify them?  */

static reg_errcode_t
check_arrival_add_next_nodes (preg, dfa, mctx, str_idx, cur_nodes, next_nodes)
     const regex_t *preg;
     re_dfa_t *dfa;
     re_match_context_t *mctx;
     int str_idx;
     re_node_set *cur_nodes, *next_nodes;
{
  int cur_idx;
  reg_errcode_t err;
  re_node_set union_set;
  re_node_set_init_empty (&union_set);
  for (cur_idx = 0; cur_idx < cur_nodes->nelem; ++cur_idx)
    {
      int naccepted = 0;
      int cur_node = cur_nodes->elems[cur_idx];
      re_token_type_t type = dfa->nodes[cur_node].type;
      if (IS_EPSILON_NODE(type))
	continue;
#ifdef RE_ENABLE_I18N
      /* If the node may accept `multi byte'.  */
      if (ACCEPT_MB_NODE (type))
	{
	  naccepted = check_node_accept_bytes (preg, cur_node, mctx->input,
					       str_idx);
	  if (naccepted > 1)
	    {
	      re_dfastate_t *dest_state;
	      int next_node = dfa->nexts[cur_node];
	      int next_idx = str_idx + naccepted;
	      dest_state = mctx->state_log[next_idx];
	      re_node_set_empty (&union_set);
	      if (dest_state)
		{
		  err = re_node_set_merge (&union_set, &dest_state->nodes);
		  if (BE (err != REG_NOERROR, 0))
		    {
		      re_node_set_free (&union_set);
		      return err;
		    }
		  err = re_node_set_insert (&union_set, next_node);
		  if (BE (err < 0, 0))
		    {
		      re_node_set_free (&union_set);
		      return REG_ESPACE;
		    }
		}
	      else
		{
		  err = re_node_set_insert (&union_set, next_node);
		  if (BE (err < 0, 0))
		    {
		      re_node_set_free (&union_set);
		      return REG_ESPACE;
		    }
		}
	      mctx->state_log[next_idx] = re_acquire_state (&err, dfa,
							    &union_set);
	      if (BE (mctx->state_log[next_idx] == NULL
		      && err != REG_NOERROR, 0))
		{
		  re_node_set_free (&union_set);
		  return err;
		}
	    }
	}
#endif /* RE_ENABLE_I18N */
      if (naccepted
	  || check_node_accept (preg, dfa->nodes + cur_node, mctx,
				str_idx))
	{
	  err = re_node_set_insert (next_nodes, dfa->nexts[cur_node]);
	  if (BE (err < 0, 0))
	    {
	      re_node_set_free (&union_set);
	      return REG_ESPACE;
	    }
	}
    }
  re_node_set_free (&union_set);
  return REG_NOERROR;
}

/* For all the nodes in CUR_NODES, add the epsilon closures of them to
   CUR_NODES, however exclude the nodes which are:
    - inside the sub expression whose number is EX_SUBEXP, if FL_OPEN.
    - out of the sub expression whose number is EX_SUBEXP, if !FL_OPEN.
*/

static reg_errcode_t
check_arrival_expand_ecl (dfa, cur_nodes, ex_subexp, fl_open)
     re_dfa_t *dfa;
     re_node_set *cur_nodes;
     int ex_subexp, fl_open;
{
  reg_errcode_t err;
  int idx, outside_node;
  re_node_set new_nodes;
#ifdef DEBUG
  assert (cur_nodes->nelem);
#endif
  err = re_node_set_alloc (&new_nodes, cur_nodes->nelem);
  if (BE (err != REG_NOERROR, 0))
    return err;
  /* Create a new node set NEW_NODES with the nodes which are epsilon
     closures of the node in CUR_NODES.  */

  for (idx = 0; idx < cur_nodes->nelem; ++idx)
    {
      int cur_node = cur_nodes->elems[idx];
      re_node_set *eclosure = dfa->eclosures + cur_node;
      outside_node = find_subexp_node (dfa, eclosure, ex_subexp, fl_open);
      if (outside_node == -1)
	{
	  /* There are no problematic nodes, just merge them.  */
	  err = re_node_set_merge (&new_nodes, eclosure);
	  if (BE (err != REG_NOERROR, 0))
	    {
	      re_node_set_free (&new_nodes);
	      return err;
	    }
	}
      else
	{
	  /* There are problematic nodes, re-calculate incrementally.  */
	  err = check_arrival_expand_ecl_sub (dfa, &new_nodes, cur_node,
					      ex_subexp, fl_open);
	  if (BE (err != REG_NOERROR, 0))
	    {
	      re_node_set_free (&new_nodes);
	      return err;
	    }
	}
    }
  re_node_set_free (cur_nodes);
  *cur_nodes = new_nodes;
  return REG_NOERROR;
}

/* Helper function for check_arrival_expand_ecl.
   Check incrementally the epsilon closure of TARGET, and if it isn't
   problematic append it to DST_NODES.  */

static reg_errcode_t
check_arrival_expand_ecl_sub (dfa, dst_nodes, target, ex_subexp, fl_open)
     re_dfa_t *dfa;
     int target, ex_subexp, fl_open;
     re_node_set *dst_nodes;
{
  int cur_node, type;
  for (cur_node = target; !re_node_set_contains (dst_nodes, cur_node);)
    {
      int err;
      type = dfa->nodes[cur_node].type;

      if (((type == OP_OPEN_SUBEXP && fl_open)
	   || (type == OP_CLOSE_SUBEXP && !fl_open))
	  && dfa->nodes[cur_node].opr.idx == ex_subexp)
	{
	  if (!fl_open)
	    {
	      err = re_node_set_insert (dst_nodes, cur_node);
	      if (BE (err == -1, 0))
		return REG_ESPACE;
	    }
	  break;
	}
      err = re_node_set_insert (dst_nodes, cur_node);
      if (BE (err == -1, 0))
	return REG_ESPACE;
      if (dfa->edests[cur_node].nelem == 0)
	break;
      if (dfa->edests[cur_node].nelem == 2)
	{
	  err = check_arrival_expand_ecl_sub (dfa, dst_nodes,
					      dfa->edests[cur_node].elems[1],
					      ex_subexp, fl_open);
	  if (BE (err != REG_NOERROR, 0))
	    return err;
	}
      cur_node = dfa->edests[cur_node].elems[0];
    }
  return REG_NOERROR;
}


/* For all the back references in the current state, calculate the
   destination of the back references by the appropriate entry
   in MCTX->BKREF_ENTS.  */

static reg_errcode_t
expand_bkref_cache (preg, mctx, cur_nodes, cur_str, last_str, subexp_num,
		    fl_open)
     const regex_t *preg;
     re_match_context_t *mctx;
     int cur_str, last_str, subexp_num, fl_open;
     re_node_set *cur_nodes;
{
  reg_errcode_t err;
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  int cache_idx, cache_idx_start;
  /* The current state.  */

  cache_idx_start = search_cur_bkref_entry (mctx, cur_str);
  for (cache_idx = cache_idx_start; cache_idx < mctx->nbkref_ents; ++cache_idx)
    {
      int to_idx, next_node;
      struct re_backref_cache_entry *ent = mctx->bkref_ents + cache_idx;
      if (ent->str_idx > cur_str)
	break;
      /* Is this entry ENT is appropriate?  */
      if (!re_node_set_contains (cur_nodes, ent->node))
	continue; /* No.  */

      to_idx = cur_str + ent->subexp_to - ent->subexp_from;
      /* Calculate the destination of the back reference, and append it
	 to MCTX->STATE_LOG.  */
      if (to_idx == cur_str)
	{
	  /* The backreference did epsilon transit, we must re-check all the
	     node in the current state.  */
	  re_node_set new_dests;
	  reg_errcode_t err2, err3;
	  next_node = dfa->edests[ent->node].elems[0];
	  if (re_node_set_contains (cur_nodes, next_node))
	    continue;
	  err = re_node_set_init_1 (&new_dests, next_node);
	  err2 = check_arrival_expand_ecl (dfa, &new_dests, subexp_num,
					   fl_open);
	  err3 = re_node_set_merge (cur_nodes, &new_dests);
	  re_node_set_free (&new_dests);
	  if (BE (err != REG_NOERROR || err2 != REG_NOERROR
		  || err3 != REG_NOERROR, 0))
	    {
	      err = (err != REG_NOERROR ? err
		     : (err2 != REG_NOERROR ? err2 : err3));
	      return err;
	    }
	  /* TODO: It is still inefficient...  */
	  cache_idx = cache_idx_start - 1;
	  continue;
	}
      else
	{
	  re_node_set union_set;
	  next_node = dfa->nexts[ent->node];
	  if (mctx->state_log[to_idx])
	    {
	      int ret;
	      if (re_node_set_contains (&mctx->state_log[to_idx]->nodes,
					next_node))
		continue;
	      err = re_node_set_init_copy (&union_set,
					   &mctx->state_log[to_idx]->nodes);
	      ret = re_node_set_insert (&union_set, next_node);
	      if (BE (err != REG_NOERROR || ret < 0, 0))
		{
		  re_node_set_free (&union_set);
		  err = err != REG_NOERROR ? err : REG_ESPACE;
		  return err;
		}
	    }
	  else
	    {
	      err = re_node_set_init_1 (&union_set, next_node);
	      if (BE (err != REG_NOERROR, 0))
		return err;
	    }
	  mctx->state_log[to_idx] = re_acquire_state (&err, dfa, &union_set);
	  re_node_set_free (&union_set);
	  if (BE (mctx->state_log[to_idx] == NULL
		  && err != REG_NOERROR, 0))
	    return err;
	}
    }
  return REG_NOERROR;
}

/* Build transition table for the state.
   Return the new table if succeeded, otherwise return NULL.  */

static re_dfastate_t **
build_trtable (preg, state)
    const regex_t *preg;
    re_dfastate_t *state;
{
  reg_errcode_t err;
  re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  int i, j, k, ch;
  int dests_node_malloced = 0, dest_states_malloced = 0;
  int ndests; /* Number of the destination states from `state'.  */
  re_dfastate_t **trtable;
  re_dfastate_t **dest_states = NULL, **dest_states_word, **dest_states_nl;
  re_node_set follows, *dests_node;
  bitset *dests_ch;
  bitset acceptable;

  /* We build DFA states which corresponds to the destination nodes
     from `state'.  `dests_node[i]' represents the nodes which i-th
     destination state contains, and `dests_ch[i]' represents the
     characters which i-th destination state accepts.  */
#ifdef _LIBC
  if (__libc_use_alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX))
    dests_node = (re_node_set *)
		 alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX);
  else
#endif
    {
      dests_node = (re_node_set *)
		   malloc ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX);
      if (BE (dests_node == NULL, 0))
	return NULL;
      dests_node_malloced = 1;
    }
  dests_ch = (bitset *) (dests_node + SBC_MAX);

  /* Initialize transiton table.  */
  trtable = (re_dfastate_t **) calloc (sizeof (re_dfastate_t *), SBC_MAX);
  state->word_trtable = 0;
  if (BE (trtable == NULL, 0))
    {
      if (dests_node_malloced)
	free (dests_node);
      return NULL;
    }

  /* At first, group all nodes belonging to `state' into several
     destinations.  */
  ndests = group_nodes_into_DFAstates (preg, state, dests_node, dests_ch);
  if (BE (ndests <= 0, 0))
    {
      if (dests_node_malloced)
	free (dests_node);
      /* Return NULL in case of an error, trtable otherwise.  */
      if (ndests == 0)
	{
	  state->trtable = trtable;
	  return trtable;
	}
      free (trtable);
      return NULL;
    }

  err = re_node_set_alloc (&follows, ndests + 1);
  if (BE (err != REG_NOERROR, 0))
    goto out_free;

#ifdef _LIBC
  if (__libc_use_alloca ((sizeof (re_node_set) + sizeof (bitset)) * SBC_MAX
			 + ndests * 3 * sizeof (re_dfastate_t *)))
    dest_states = (re_dfastate_t **)
		  alloca (ndests * 3 * sizeof (re_dfastate_t *));
  else
#endif
    {
      dest_states = (re_dfastate_t **)
		    malloc (ndests * 3 * sizeof (re_dfastate_t *));
      if (BE (dest_states == NULL, 0))
	{
out_free:
	  if (dest_states_malloced)
	    free (dest_states);
	  re_node_set_free (&follows);
	  for (i = 0; i < ndests; ++i)
	    re_node_set_free (dests_node + i);
	  free (trtable);
	  if (dests_node_malloced)
	    free (dests_node);
	  return NULL;
	}
      dest_states_malloced = 1;
    }
  dest_states_word = dest_states + ndests;
  dest_states_nl = dest_states_word + ndests;
  bitset_empty (acceptable);

  /* Then build the states for all destinations.  */
  for (i = 0; i < ndests; ++i)
    {
      int next_node;
      re_node_set_empty (&follows);
      /* Merge the follows of this destination states.  */
      for (j = 0; j < dests_node[i].nelem; ++j)
	{
	  next_node = dfa->nexts[dests_node[i].elems[j]];
	  if (next_node != -1)
	    {
	      err = re_node_set_merge (&follows, dfa->eclosures + next_node);
	      if (BE (err != REG_NOERROR, 0))
		goto out_free;
	    }
	}
      dest_states[i] = re_acquire_state_context (&err, dfa, &follows, 0);
      if (BE (dest_states[i] == NULL && err != REG_NOERROR, 0))
	goto out_free;
      /* If the new state has context constraint,
	 build appropriate states for these contexts.  */
      if (dest_states[i]->has_constraint)
	{
	  dest_states_word[i] = re_acquire_state_context (&err, dfa, &follows,
							  CONTEXT_WORD);
	  if (BE (dest_states_word[i] == NULL && err != REG_NOERROR, 0))
	    goto out_free;
	  dest_states_nl[i] = re_acquire_state_context (&err, dfa, &follows,
							CONTEXT_NEWLINE);
	  if (BE (dest_states_nl[i] == NULL && err != REG_NOERROR, 0))
	    goto out_free;
	}
      else
	{
	  dest_states_word[i] = dest_states[i];
	  dest_states_nl[i] = dest_states[i];
	}
      bitset_merge (acceptable, dests_ch[i]);
    }

  /* Update the transition table.  */
  /* For all characters ch...:  */
  for (i = 0, ch = 0; i < BITSET_UINTS; ++i)
    for (j = 0; j < UINT_BITS; ++j, ++ch)
      if ((acceptable[i] >> j) & 1)
	{
	  for (k = 0; k < ndests; ++k)
	    if ((dests_ch[k][i] >> j) & 1)
	      {
		/* k-th destination accepts the word character ch.  */
		if (state->word_trtable)
		  {
		    trtable[ch] = dest_states[k];
		    trtable[ch + SBC_MAX] = dest_states_word[k];
		  }
		else if (dfa->mb_cur_max > 1
			 && dest_states[k] != dest_states_word[k])
		  {
		    re_dfastate_t **new_trtable;

		    new_trtable = (re_dfastate_t **)
				  realloc (trtable,
					   sizeof (re_dfastate_t *)
					   * 2 * SBC_MAX);
		    if (BE (new_trtable == NULL, 0))
		      goto out_free;
		    memcpy (new_trtable + SBC_MAX, new_trtable,
			    sizeof (re_dfastate_t *) * SBC_MAX);
		    trtable = new_trtable;
		    state->word_trtable = 1;
		    trtable[ch] = dest_states[k];
		    trtable[ch + SBC_MAX] = dest_states_word[k];
		  }
		else if (IS_WORD_CHAR (ch))
		  trtable[ch] = dest_states_word[k];
		else
		  trtable[ch] = dest_states[k];
		/* There must be only one destination which accepts
		   character ch.  See group_nodes_into_DFAstates.  */
		break;
	      }
	}
  /* new line */
  if (bitset_contain (acceptable, NEWLINE_CHAR))
    {
      /* The current state accepts newline character.  */
      for (k = 0; k < ndests; ++k)
	if (bitset_contain (dests_ch[k], NEWLINE_CHAR))
	  {
	    /* k-th destination accepts newline character.  */
	    trtable[NEWLINE_CHAR] = dest_states_nl[k];
	    if (state->word_trtable)
	      trtable[NEWLINE_CHAR + SBC_MAX] = dest_states_nl[k];
	    /* There must be only one destination which accepts
	       newline.  See group_nodes_into_DFAstates.  */
	    break;
	  }
    }

  if (dest_states_malloced)
    free (dest_states);

  re_node_set_free (&follows);
  for (i = 0; i < ndests; ++i)
    re_node_set_free (dests_node + i);

  if (dests_node_malloced)
    free (dests_node);

  state->trtable = trtable;
  return trtable;
}

/* Group all nodes belonging to STATE into several destinations.
   Then for all destinations, set the nodes belonging to the destination
   to DESTS_NODE[i] and set the characters accepted by the destination
   to DEST_CH[i].  This function return the number of destinations.  */

static int
group_nodes_into_DFAstates (preg, state, dests_node, dests_ch)
    const regex_t *preg;
    const re_dfastate_t *state;
    re_node_set *dests_node;
    bitset *dests_ch;
{
  reg_errcode_t err;
  const re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  int i, j, k;
  int ndests; /* Number of the destinations from `state'.  */
  bitset accepts; /* Characters a node can accept.  */
  const re_node_set *cur_nodes = &state->nodes;
  bitset_empty (accepts);
  ndests = 0;

  /* For all the nodes belonging to `state',  */
  for (i = 0; i < cur_nodes->nelem; ++i)
    {
      re_token_t *node = &dfa->nodes[cur_nodes->elems[i]];
      re_token_type_t type = node->type;
      unsigned int constraint = node->constraint;

      /* Enumerate all single byte character this node can accept.  */
      if (type == CHARACTER)
	bitset_set (accepts, node->opr.c);
      else if (type == SIMPLE_BRACKET)
	{
	  bitset_merge (accepts, node->opr.sbcset);
	}
      else if (type == OP_PERIOD)
	{
#ifdef RE_ENABLE_I18N
	  if (dfa->mb_cur_max > 1)
	    bitset_merge (accepts, dfa->sb_char);
	  else
#endif	  
	    bitset_set_all (accepts);
	  if (!(preg->syntax & RE_DOT_NEWLINE))
	    bitset_clear (accepts, '\n');
	  if (preg->syntax & RE_DOT_NOT_NULL)
	    bitset_clear (accepts, '\0');
	}
      else if (type == OP_UTF8_PERIOD)
        {
	  memset (accepts, 255, sizeof (unsigned int) * BITSET_UINTS / 2);
	  if (!(preg->syntax & RE_DOT_NEWLINE))
	    bitset_clear (accepts, '\n');
	  if (preg->syntax & RE_DOT_NOT_NULL)
	    bitset_clear (accepts, '\0');
        }
      else
	continue;

      /* Check the `accepts' and sift the characters which are not
	 match it the context.  */
      if (constraint)
	{
	  if (constraint & NEXT_NEWLINE_CONSTRAINT)
	    {
	      int accepts_newline = bitset_contain (accepts, NEWLINE_CHAR);
	      bitset_empty (accepts);
	      if (accepts_newline)
		bitset_set (accepts, NEWLINE_CHAR);
	      else
		continue;
	    }
	  if (constraint & NEXT_ENDBUF_CONSTRAINT)
	    {
	      bitset_empty (accepts);
	      continue;
	    }

	  if (constraint & NEXT_WORD_CONSTRAINT)
	    {
#ifdef RE_ENABLE_I18N
	      if (dfa->mb_cur_max > 1)
		for (j = 0; j < BITSET_UINTS; ++j)
		  accepts[j] &= (dfa->word_char[j] | ~dfa->sb_char[j]);
	      else
#endif
		for (j = 0; j < BITSET_UINTS; ++j)
		  accepts[j] &= dfa->word_char[j];
	    }
	  if (constraint & NEXT_NOTWORD_CONSTRAINT)
	    {
#ifdef RE_ENABLE_I18N
	      if (dfa->mb_cur_max > 1)
		for (j = 0; j < BITSET_UINTS; ++j)
		  accepts[j] &= ~(dfa->word_char[j] & dfa->sb_char[j]);
	      else
#endif
		for (j = 0; j < BITSET_UINTS; ++j)
		  accepts[j] &= ~dfa->word_char[j];
	    }
	}

      /* Then divide `accepts' into DFA states, or create a new
	 state.  */
      for (j = 0; j < ndests; ++j)
	{
	  bitset intersec; /* Intersection sets, see below.  */
	  bitset remains;
	  /* Flags, see below.  */
	  int has_intersec, not_subset, not_consumed;

	  /* Optimization, skip if this state doesn't accept the character.  */
	  if (type == CHARACTER && !bitset_contain (dests_ch[j], node->opr.c))
	    continue;

	  /* Enumerate the intersection set of this state and `accepts'.  */
	  has_intersec = 0;
	  for (k = 0; k < BITSET_UINTS; ++k)
	    has_intersec |= intersec[k] = accepts[k] & dests_ch[j][k];
	  /* And skip if the intersection set is empty.  */
	  if (!has_intersec)
	    continue;

	  /* Then check if this state is a subset of `accepts'.  */
	  not_subset = not_consumed = 0;
	  for (k = 0; k < BITSET_UINTS; ++k)
	    {
	      not_subset |= remains[k] = ~accepts[k] & dests_ch[j][k];
	      not_consumed |= accepts[k] = accepts[k] & ~dests_ch[j][k];
	    }

	  /* If this state isn't a subset of `accepts', create a
	     new group state, which has the `remains'. */
	  if (not_subset)
	    {
	      bitset_copy (dests_ch[ndests], remains);
	      bitset_copy (dests_ch[j], intersec);
	      err = re_node_set_init_copy (dests_node + ndests, &dests_node[j]);
	      if (BE (err != REG_NOERROR, 0))
		goto error_return;
	      ++ndests;
	    }

	  /* Put the position in the current group. */
	  err = re_node_set_insert (&dests_node[j], cur_nodes->elems[i]);
	  if (BE (err < 0, 0))
	    goto error_return;

	  /* If all characters are consumed, go to next node. */
	  if (!not_consumed)
	    break;
	}
      /* Some characters remain, create a new group. */
      if (j == ndests)
	{
	  bitset_copy (dests_ch[ndests], accepts);
	  err = re_node_set_init_1 (dests_node + ndests, cur_nodes->elems[i]);
	  if (BE (err != REG_NOERROR, 0))
	    goto error_return;
	  ++ndests;
	  bitset_empty (accepts);
	}
    }
  return ndests;
 error_return:
  for (j = 0; j < ndests; ++j)
    re_node_set_free (dests_node + j);
  return -1;
}

#ifdef RE_ENABLE_I18N
/* Check how many bytes the node `dfa->nodes[node_idx]' accepts.
   Return the number of the bytes the node accepts.
   STR_IDX is the current index of the input string.

   This function handles the nodes which can accept one character, or
   one collating element like '.', '[a-z]', opposite to the other nodes
   can only accept one byte.  */

static int
check_node_accept_bytes (preg, node_idx, input, str_idx)
    const regex_t *preg;
    int node_idx, str_idx;
    const re_string_t *input;
{
  const re_dfa_t *dfa = (re_dfa_t *) preg->buffer;
  const re_token_t *node = dfa->nodes + node_idx;
  int char_len, elem_len;
  int i;

  if (BE (node->type == OP_UTF8_PERIOD, 0))
    {
      unsigned char c = re_string_byte_at (input, str_idx), d;
      if (BE (c < 0xc2, 1))
	return 0;

      if (str_idx + 2 > input->len)
	return 0;

      d = re_string_byte_at (input, str_idx + 1);
      if (c < 0xe0)
	return (d < 0x80 || d > 0xbf) ? 0 : 2;
      else if (c < 0xf0)
	{
	  char_len = 3;
	  if (c == 0xe0 && d < 0xa0)
	    return 0;
	}
      else if (c < 0xf8)
	{
	  char_len = 4;
	  if (c == 0xf0 && d < 0x90)
	    return 0;
	}
      else if (c < 0xfc)
	{
	  char_len = 5;
	  if (c == 0xf8 && d < 0x88)
	    return 0;
	}
      else if (c < 0xfe)
	{
	  char_len = 6;
	  if (c == 0xfc && d < 0x84)
	    return 0;
	}
      else
	return 0;

      if (str_idx + char_len > input->len)
	return 0;

      for (i = 1; i < char_len; ++i)
	{
	  d = re_string_byte_at (input, str_idx + i);
	  if (d < 0x80 || d > 0xbf)
	    return 0;
	}
      return char_len;
    }

  char_len = re_string_char_size_at (input, str_idx);
  if (node->type == OP_PERIOD)
    {
      if (char_len <= 1)
        return 0;
      /* FIXME: I don't think this if is needed, as both '\n'
	 and '\0' are char_len == 1.  */
      /* '.' accepts any one character except the following two cases.  */
      if ((!(preg->syntax & RE_DOT_NEWLINE) &&
	   re_string_byte_at (input, str_idx) == '\n') ||
	  ((preg->syntax & RE_DOT_NOT_NULL) &&
	   re_string_byte_at (input, str_idx) == '\0'))
	return 0;
      return char_len;
    }

  elem_len = re_string_elem_size_at (input, str_idx);
  if (elem_len <= 1 && char_len <= 1)
    return 0;

  if (node->type == COMPLEX_BRACKET)
    {
      const re_charset_t *cset = node->opr.mbcset;
# ifdef _LIBC
      const unsigned char *pin = ((char *) re_string_get_buffer (input)
				  + str_idx);
      int j;
      uint32_t nrules;
# endif /* _LIBC */
      int match_len = 0;
      wchar_t wc = ((cset->nranges || cset->nchar_classes || cset->nmbchars)
		    ? re_string_wchar_at (input, str_idx) : 0);

      /* match with multibyte character?  */
      for (i = 0; i < cset->nmbchars; ++i)
	if (wc == cset->mbchars[i])
	  {
	    match_len = char_len;
	    goto check_node_accept_bytes_match;
	  }
      /* match with character_class?  */
      for (i = 0; i < cset->nchar_classes; ++i)
	{
	  wctype_t wt = cset->char_classes[i];
	  if (__iswctype (wc, wt))
	    {
	      match_len = char_len;
	      goto check_node_accept_bytes_match;
	    }
	}

# ifdef _LIBC
      nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
      if (nrules != 0)
	{
	  unsigned int in_collseq = 0;
	  const int32_t *table, *indirect;
	  const unsigned char *weights, *extra;
	  const char *collseqwc;
	  int32_t idx;
	  /* This #include defines a local function!  */
#  include <locale/weight.h>

	  /* match with collating_symbol?  */
	  if (cset->ncoll_syms)
	    extra = (const unsigned char *)
	      _NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB);
	  for (i = 0; i < cset->ncoll_syms; ++i)
	    {
	      const unsigned char *coll_sym = extra + cset->coll_syms[i];
	      /* Compare the length of input collating element and
		 the length of current collating element.  */
	      if (*coll_sym != elem_len)
		continue;
	      /* Compare each bytes.  */
	      for (j = 0; j < *coll_sym; j++)
		if (pin[j] != coll_sym[1 + j])
		  break;
	      if (j == *coll_sym)
		{
		  /* Match if every bytes is equal.  */
		  match_len = j;
		  goto check_node_accept_bytes_match;
		}
	    }

	  if (cset->nranges)
	    {
	      if (elem_len <= char_len)
		{
		  collseqwc = _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQWC);
		  in_collseq = __collseq_table_lookup (collseqwc, wc);
		}
	      else
		in_collseq = find_collation_sequence_value (pin, elem_len);
	    }
	  /* match with range expression?  */
	  for (i = 0; i < cset->nranges; ++i)
	    if (cset->range_starts[i] <= in_collseq
		&& in_collseq <= cset->range_ends[i])
	      {
		match_len = elem_len;
		goto check_node_accept_bytes_match;
	      }

	  /* match with equivalence_class?  */
	  if (cset->nequiv_classes)
	    {
	      const unsigned char *cp = pin;
	      table = (const int32_t *)
		_NL_CURRENT (LC_COLLATE, _NL_COLLATE_TABLEMB);
	      weights = (const unsigned char *)
		_NL_CURRENT (LC_COLLATE, _NL_COLLATE_WEIGHTMB);
	      extra = (const unsigned char *)
		_NL_CURRENT (LC_COLLATE, _NL_COLLATE_EXTRAMB);
	      indirect = (const int32_t *)
		_NL_CURRENT (LC_COLLATE, _NL_COLLATE_INDIRECTMB);
	      idx = findidx (&cp);
	      if (idx > 0)
		for (i = 0; i < cset->nequiv_classes; ++i)
		  {
		    int32_t equiv_class_idx = cset->equiv_classes[i];
		    size_t weight_len = weights[idx];
		    if (weight_len == weights[equiv_class_idx])
		      {
			int cnt = 0;
			while (cnt <= weight_len
			       && (weights[equiv_class_idx + 1 + cnt]
				   == weights[idx + 1 + cnt]))
			  ++cnt;
			if (cnt > weight_len)
			  {
			    match_len = elem_len;
			    goto check_node_accept_bytes_match;
			  }
		      }
		  }
	    }
	}
      else
# endif /* _LIBC */
	{
	  /* match with range expression?  */
#if __GNUC__ >= 2
	  wchar_t cmp_buf[] = {L'\0', L'\0', wc, L'\0', L'\0', L'\0'};
#else
	  wchar_t cmp_buf[] = {L'\0', L'\0', L'\0', L'\0', L'\0', L'\0'};
	  cmp_buf[2] = wc;
#endif
	  for (i = 0; i < cset->nranges; ++i)
	    {
	      cmp_buf[0] = cset->range_starts[i];
	      cmp_buf[4] = cset->range_ends[i];
	      if (wcscoll (cmp_buf, cmp_buf + 2) <= 0
		  && wcscoll (cmp_buf + 2, cmp_buf + 4) <= 0)
		{
		  match_len = char_len;
		  goto check_node_accept_bytes_match;
		}
	    }
	}
    check_node_accept_bytes_match:
      if (!cset->non_match)
	return match_len;
      else
	{
	  if (match_len > 0)
	    return 0;
	  else
	    return (elem_len > char_len) ? elem_len : char_len;
	}
    }
  return 0;
}

# ifdef _LIBC
static unsigned int
find_collation_sequence_value (mbs, mbs_len)
    const unsigned char *mbs;
    size_t mbs_len;
{
  uint32_t nrules = _NL_CURRENT_WORD (LC_COLLATE, _NL_COLLATE_NRULES);
  if (nrules == 0)
    {
      if (mbs_len == 1)
	{
	  /* No valid character.  Match it as a single byte character.  */
	  const unsigned char *collseq = (const unsigned char *)
	    _NL_CURRENT (LC_COLLATE, _NL_COLLATE_COLLSEQMB);
	  return collseq[mbs[0]];
	}
      return UINT_MAX;
    }
  else
    {
      int32_t idx;
      const unsigned char *extra = (const unsigned char *)
	_NL_CURRENT (LC_COLLATE, _NL_COLLATE_SYMB_EXTRAMB);

      for (idx = 0; ;)
	{
	  int mbs_cnt, found = 0;
	  int32_t elem_mbs_len;
	  /* Skip the name of collating element name.  */
	  idx = idx + extra[idx] + 1;
	  elem_mbs_len = extra[idx++];
	  if (mbs_len == elem_mbs_len)
	    {
	      for (mbs_cnt = 0; mbs_cnt < elem_mbs_len; ++mbs_cnt)
		if (extra[idx + mbs_cnt] != mbs[mbs_cnt])
		  break;
	      if (mbs_cnt == elem_mbs_len)
		/* Found the entry.  */
		found = 1;
	    }
	  /* Skip the byte sequence of the collating element.  */
	  idx += elem_mbs_len;
	  /* Adjust for the alignment.  */
	  idx = (idx + 3) & ~3;
	  /* Skip the collation sequence value.  */
	  idx += sizeof (uint32_t);
	  /* Skip the wide char sequence of the collating element.  */
	  idx = idx + sizeof (uint32_t) * (extra[idx] + 1);
	  /* If we found the entry, return the sequence value.  */
	  if (found)
	    return *(uint32_t *) (extra + idx);
	  /* Skip the collation sequence value.  */
	  idx += sizeof (uint32_t);
	}
    }
}
# endif /* _LIBC */
#endif /* RE_ENABLE_I18N */

/* Check whether the node accepts the byte which is IDX-th
   byte of the INPUT.  */

static int
check_node_accept (preg, node, mctx, idx)
    const regex_t *preg;
    const re_token_t *node;
    const re_match_context_t *mctx;
    int idx;
{
  unsigned char ch;
  if (node->constraint)
    {
      /* The node has constraints.  Check whether the current context
	 satisfies the constraints.  */
      unsigned int context = re_string_context_at (mctx->input, idx,
						   mctx->eflags,
						   preg->newline_anchor);
      if (NOT_SATISFY_NEXT_CONSTRAINT (node->constraint, context))
	return 0;
    }
  ch = re_string_byte_at (mctx->input, idx);
  switch (node->type)
    {
    case CHARACTER:
      return node->opr.c == ch;
    case SIMPLE_BRACKET:
      return bitset_contain (node->opr.sbcset, ch);
    case OP_UTF8_PERIOD:
      if (ch >= 0x80)
        return 0;
      /* FALLTHROUGH */
    case OP_PERIOD:
      return !((ch == '\n' && !(preg->syntax & RE_DOT_NEWLINE))
	       || (ch == '\0' && (preg->syntax & RE_DOT_NOT_NULL)));
    default:
      return 0;
    }
}

/* Extend the buffers, if the buffers have run out.  */

static reg_errcode_t
extend_buffers (mctx)
     re_match_context_t *mctx;
{
  reg_errcode_t ret;
  re_string_t *pstr = mctx->input;

  /* Double the lengthes of the buffers.  */
  ret = re_string_realloc_buffers (pstr, pstr->bufs_len * 2);
  if (BE (ret != REG_NOERROR, 0))
    return ret;

  if (mctx->state_log != NULL)
    {
      /* And double the length of state_log.  */
      /* XXX We have no indication of the size of this buffer.  If this
	 allocation fail we have no indication that the state_log array
	 does not have the right size.  */
      re_dfastate_t **new_array = re_realloc (mctx->state_log, re_dfastate_t *,
					      pstr->bufs_len + 1);
      if (BE (new_array == NULL, 0))
	return REG_ESPACE;
      mctx->state_log = new_array;
    }

  /* Then reconstruct the buffers.  */
  if (pstr->icase)
    {
#ifdef RE_ENABLE_I18N
      if (pstr->mb_cur_max > 1)
	build_wcs_upper_buffer (pstr);
      else
#endif /* RE_ENABLE_I18N  */
	build_upper_buffer (pstr);
    }
  else
    {
#ifdef RE_ENABLE_I18N
      if (pstr->mb_cur_max > 1)
	build_wcs_buffer (pstr);
      else
#endif /* RE_ENABLE_I18N  */
	{
	  if (pstr->trans != NULL)
	    re_string_translate_buffer (pstr);
	}
    }
  return REG_NOERROR;
}


/* Functions for matching context.  */

/* Initialize MCTX.  */

static reg_errcode_t
match_ctx_init (mctx, eflags, input, n)
    re_match_context_t *mctx;
    int eflags, n;
    re_string_t *input;
{
  mctx->eflags = eflags;
  mctx->input = input;
  mctx->match_last = -1;
  if (n > 0)
    {
      mctx->bkref_ents = re_malloc (struct re_backref_cache_entry, n);
      mctx->sub_tops = re_malloc (re_sub_match_top_t *, n);
      if (BE (mctx->bkref_ents == NULL || mctx->sub_tops == NULL, 0))
	return REG_ESPACE;
    }
  else
    mctx->bkref_ents = NULL;
  mctx->nbkref_ents = 0;
  mctx->abkref_ents = n;
  mctx->max_mb_elem_len = 1;
  mctx->nsub_tops = 0;
  mctx->asub_tops = n;
  return REG_NOERROR;
}

/* Clean the entries which depend on the current input in MCTX.
   This function must be invoked when the matcher changes the start index
   of the input, or changes the input string.  */

static void
match_ctx_clean (mctx)
    re_match_context_t *mctx;
{
  match_ctx_free_subtops (mctx);
  mctx->nsub_tops = 0;
  mctx->nbkref_ents = 0;
}

/* Free all the memory associated with MCTX.  */

static void
match_ctx_free (mctx)
    re_match_context_t *mctx;
{
  match_ctx_free_subtops (mctx);
  re_free (mctx->sub_tops);
  re_free (mctx->bkref_ents);
}

/* Free all the memory associated with MCTX->SUB_TOPS.  */

static void
match_ctx_free_subtops (mctx)
     re_match_context_t *mctx;
{
  int st_idx;
  for (st_idx = 0; st_idx < mctx->nsub_tops; ++st_idx)
    {
      int sl_idx;
      re_sub_match_top_t *top = mctx->sub_tops[st_idx];
      for (sl_idx = 0; sl_idx < top->nlasts; ++sl_idx)
	{
	  re_sub_match_last_t *last = top->lasts[sl_idx];
	  re_free (last->path.array);
	  re_free (last);
	}
      re_free (top->lasts);
      if (top->path)
	{
	  re_free (top->path->array);
	  re_free (top->path);
	}
      free (top);
    }
}

/* Add a new backreference entry to MCTX.
   Note that we assume that caller never call this function with duplicate
   entry, and call with STR_IDX which isn't smaller than any existing entry.
*/

static reg_errcode_t
match_ctx_add_entry (mctx, node, str_idx, from, to)
     re_match_context_t *mctx;
     int node, str_idx, from, to;
{
  if (mctx->nbkref_ents >= mctx->abkref_ents)
    {
      struct re_backref_cache_entry* new_entry;
      new_entry = re_realloc (mctx->bkref_ents, struct re_backref_cache_entry,
			      mctx->abkref_ents * 2);
      if (BE (new_entry == NULL, 0))
	{
	  re_free (mctx->bkref_ents);
	  return REG_ESPACE;
	}
      mctx->bkref_ents = new_entry;
      memset (mctx->bkref_ents + mctx->nbkref_ents, '\0',
	      sizeof (struct re_backref_cache_entry) * mctx->abkref_ents);
      mctx->abkref_ents *= 2;
    }
  mctx->bkref_ents[mctx->nbkref_ents].node = node;
  mctx->bkref_ents[mctx->nbkref_ents].str_idx = str_idx;
  mctx->bkref_ents[mctx->nbkref_ents].subexp_from = from;
  mctx->bkref_ents[mctx->nbkref_ents].subexp_to = to;
  mctx->bkref_ents[mctx->nbkref_ents++].flag = 0;
  if (mctx->max_mb_elem_len < to - from)
    mctx->max_mb_elem_len = to - from;
  return REG_NOERROR;
}

/* Search for the first entry which has the same str_idx.
   Note that MCTX->BKREF_ENTS is already sorted by MCTX->STR_IDX.  */

static int
search_cur_bkref_entry (mctx, str_idx)
     re_match_context_t *mctx;
     int str_idx;
{
  int left, right, mid;
  right = mctx->nbkref_ents;
  for (left = 0; left < right;)
    {
      mid = (left + right) / 2;
      if (mctx->bkref_ents[mid].str_idx < str_idx)
	left = mid + 1;
      else
	right = mid;
    }
  return left;
}

static void
match_ctx_clear_flag (mctx)
     re_match_context_t *mctx;
{
  int i;
  for (i = 0; i < mctx->nbkref_ents; ++i)
    mctx->bkref_ents[i].flag = 0;
}

/* Register the node NODE, whose type is OP_OPEN_SUBEXP, and which matches
   at STR_IDX.  */

static reg_errcode_t
match_ctx_add_subtop (mctx, node, str_idx)
     re_match_context_t *mctx;
     int node, str_idx;
{
#ifdef DEBUG
  assert (mctx->sub_tops != NULL);
  assert (mctx->asub_tops > 0);
#endif
  if (BE (mctx->nsub_tops == mctx->asub_tops, 0))
    {
      int new_asub_tops = mctx->asub_tops * 2;
      re_sub_match_top_t **new_array = re_realloc (mctx->sub_tops,
						   re_sub_match_top_t *,
						   new_asub_tops);
      if (BE (new_array == NULL, 0))
	return REG_ESPACE;
      mctx->sub_tops = new_array;
      mctx->asub_tops = new_asub_tops;
    }
  mctx->sub_tops[mctx->nsub_tops] = calloc (1, sizeof (re_sub_match_top_t));
  if (BE (mctx->sub_tops[mctx->nsub_tops] == NULL, 0))
    return REG_ESPACE;
  mctx->sub_tops[mctx->nsub_tops]->node = node;
  mctx->sub_tops[mctx->nsub_tops++]->str_idx = str_idx;
  return REG_NOERROR;
}

/* Register the node NODE, whose type is OP_CLOSE_SUBEXP, and which matches
   at STR_IDX, whose corresponding OP_OPEN_SUBEXP is SUB_TOP.  */

static re_sub_match_last_t *
match_ctx_add_sublast (subtop, node, str_idx)
     re_sub_match_top_t *subtop;
     int node, str_idx;
{
  re_sub_match_last_t *new_entry;
  if (BE (subtop->nlasts == subtop->alasts, 0))
    {
      int new_alasts = 2 * subtop->alasts + 1;
      re_sub_match_last_t **new_array = re_realloc (subtop->lasts,
						    re_sub_match_last_t *,
						    new_alasts);
      if (BE (new_array == NULL, 0))
	return NULL;
      subtop->lasts = new_array;
      subtop->alasts = new_alasts;
    }
  new_entry = calloc (1, sizeof (re_sub_match_last_t));
  if (BE (new_entry != NULL, 1))
    {
      subtop->lasts[subtop->nlasts] = new_entry;
      new_entry->node = node;
      new_entry->str_idx = str_idx;
      ++subtop->nlasts;
    }
  return new_entry;
}

static void
sift_ctx_init (sctx, sifted_sts, limited_sts, last_node, last_str_idx,
	       check_subexp)
    re_sift_context_t *sctx;
    re_dfastate_t **sifted_sts, **limited_sts;
    int last_node, last_str_idx, check_subexp;
{
  sctx->sifted_states = sifted_sts;
  sctx->limited_states = limited_sts;
  sctx->last_node = last_node;
  sctx->last_str_idx = last_str_idx;
  sctx->check_subexp = check_subexp;
  sctx->cur_bkref = -1;
  sctx->cls_subexp_idx = -1;
  re_node_set_init_empty (&sctx->limits);
}