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
4102
4103
4104
4105
4106
4107
4108
4109
4110
4111
4112
4113
4114
4115
4116
4117
4118
4119
4120
4121
4122
4123
4124
4125
4126
4127
4128
4129
4130
4131
4132
4133
4134
4135
4136
4137
4138
4139
4140
4141
4142
4143
4144
4145
4146
4147
4148
4149
4150
4151
4152
4153
4154
4155
4156
4157
4158
4159
4160
4161
4162
4163
4164
4165
4166
4167
4168
4169
4170
4171
4172
4173
4174
4175
4176
4177
4178
4179
4180
4181
4182
4183
4184
4185
4186
4187
4188
4189
4190
4191
4192
4193
4194
4195
4196
4197
4198
4199
4200
4201
4202
4203
4204
4205
4206
4207
4208
4209
4210
4211
4212
4213
4214
4215
4216
4217
4218
4219
4220
4221
4222
4223
4224
4225
4226
4227
4228
4229
4230
4231
4232
4233
4234
4235
4236
4237
4238
4239
4240
4241
4242
4243
4244
4245
4246
4247
4248
4249
4250
4251
4252
4253
4254
4255
4256
4257
4258
4259
4260
4261
4262
4263
4264
4265
4266
4267
4268
4269
4270
4271
4272
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
4290
4291
4292
4293
4294
4295
4296
4297
4298
4299
4300
4301
4302
4303
4304
4305
4306
4307
4308
4309
4310
4311
4312
4313
4314
4315
4316
4317
4318
4319
4320
4321
4322
4323
4324
4325
4326
4327
4328
4329
4330
4331
4332
4333
4334
4335
4336
4337
4338
4339
4340
4341
4342
4343
4344
4345
4346
4347
4348
4349
4350
4351
4352
4353
4354
4355
4356
4357
4358
4359
4360
4361
4362
4363
4364
4365
4366
4367
4368
4369
4370
4371
4372
4373
4374
4375
4376
4377
4378
4379
4380
4381
4382
4383
4384
4385
4386
4387
4388
4389
4390
4391
4392
4393
4394
4395
4396
4397
4398
4399
4400
4401
4402
4403
4404
4405
4406
4407
4408
4409
4410
4411
4412
4413
4414
4415
4416
4417
4418
4419
4420
4421
4422
4423
4424
4425
4426
4427
4428
4429
4430
4431
4432
4433
4434
4435
4436
4437
4438
4439
4440
4441
4442
4443
4444
4445
4446
4447
4448
4449
4450
4451
4452
4453
4454
4455
4456
4457
4458
4459
4460
4461
4462
4463
4464
4465
4466
4467
4468
4469
4470
4471
4472
4473
4474
4475
4476
4477
4478
4479
4480
4481
4482
4483
4484
4485
4486
4487
4488
4489
4490
4491
4492
4493
4494
4495
4496
4497
4498
4499
4500
4501
4502
4503
4504
4505
4506
4507
4508
4509
4510
4511
4512
4513
4514
4515
4516
4517
4518
4519
4520
4521
4522
4523
4524
4525
4526
4527
4528
4529
4530
4531
4532
4533
4534
4535
4536
4537
4538
4539
4540
4541
4542
4543
4544
4545
4546
4547
4548
4549
4550
4551
4552
4553
4554
4555
4556
4557
4558
4559
4560
4561
4562
4563
4564
4565
4566
4567
4568
4569
4570
4571
4572
4573
4574
4575
4576
4577
4578
4579
4580
4581
4582
4583
4584
4585
4586
4587
4588
4589
4590
4591
4592
4593
4594
4595
4596
4597
4598
4599
4600
4601
4602
4603
4604
4605
4606
4607
4608
4609
4610
4611
4612
4613
4614
4615
4616
4617
4618
4619
4620
4621
4622
4623
4624
4625
4626
4627
4628
4629
4630
4631
4632
4633
4634
4635
4636
4637
4638
4639
4640
4641
4642
4643
4644
4645
4646
4647
4648
4649
4650
4651
4652
4653
4654
4655
4656
4657
4658
4659
4660
4661
4662
4663
4664
4665
4666
4667
4668
4669
4670
4671
4672
4673
4674
4675
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
4686
4687
4688
4689
4690
4691
4692
4693
4694
4695
4696
4697
4698
4699
4700
4701
4702
4703
4704
4705
4706
4707
4708
4709
4710
4711
4712
4713
4714
4715
4716
4717
4718
4719
4720
4721
4722
4723
4724
4725
4726
4727
4728
4729
4730
4731
4732
4733
4734
4735
4736
4737
4738
4739
4740
4741
4742
4743
4744
4745
4746
4747
4748
4749
4750
4751
4752
4753
4754
4755
4756
4757
4758
4759
4760
4761
4762
4763
4764
4765
4766
4767
4768
4769
4770
4771
4772
4773
4774
4775
4776
4777
4778
4779
4780
4781
4782
4783
4784
4785
4786
4787
4788
4789
4790
4791
4792
4793
4794
4795
4796
4797
4798
4799
4800
4801
4802
4803
4804
4805
4806
4807
4808
4809
4810
4811
4812
4813
4814
4815
4816
4817
4818
4819
4820
4821
4822
4823
4824
4825
4826
4827
4828
4829
4830
4831
4832
4833
4834
4835
4836
4837
4838
4839
4840
4841
4842
4843
4844
4845
4846
4847
4848
4849
4850
4851
4852
4853
4854
4855
4856
4857
4858
4859
4860
4861
4862
4863
4864
4865
4866
4867
4868
4869
4870
4871
4872
4873
4874
4875
4876
4877
4878
4879
4880
4881
4882
4883
4884
4885
4886
4887
4888
4889
4890
4891
4892
4893
4894
4895
4896
4897
4898
4899
4900
4901
4902
4903
4904
4905
4906
4907
4908
4909
4910
4911
4912
4913
4914
4915
4916
4917
4918
4919
4920
4921
4922
4923
4924
4925
4926
4927
4928
4929
4930
4931
4932
4933
4934
4935
4936
4937
4938
4939
4940
4941
4942
4943
4944
4945
4946
4947
4948
4949
4950
4951
4952
4953
4954
4955
4956
4957
4958
4959
4960
4961
4962
4963
4964
4965
4966
4967
4968
4969
4970
4971
4972
4973
4974
4975
4976
4977
4978
4979
4980
4981
4982
4983
4984
4985
4986
4987
4988
4989
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
5012
5013
5014
5015
5016
5017
5018
5019
5020
5021
5022
5023
5024
5025
5026
5027
5028
5029
5030
5031
5032
5033
5034
5035
5036
5037
5038
5039
5040
5041
5042
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
5058
5059
5060
5061
5062
5063
5064
5065
5066
5067
5068
5069
5070
5071
5072
5073
5074
5075
5076
5077
5078
5079
5080
5081
5082
5083
5084
5085
5086
5087
5088
5089
5090
5091
5092
5093
5094
5095
5096
5097
5098
5099
5100
5101
5102
5103
5104
5105
5106
5107
5108
5109
5110
5111
5112
5113
5114
5115
5116
5117
5118
5119
5120
5121
5122
5123
5124
5125
5126
5127
5128
5129
5130
5131
5132
5133
5134
5135
5136
5137
5138
5139
5140
5141
5142
5143
5144
5145
5146
5147
5148
5149
5150
5151
5152
5153
5154
5155
5156
5157
5158
5159
5160
5161
5162
5163
5164
5165
5166
5167
5168
5169
5170
5171
5172
5173
5174
5175
5176
5177
5178
5179
5180
5181
5182
5183
5184
5185
5186
5187
5188
5189
5190
5191
5192
5193
5194
5195
5196
5197
5198
5199
5200
5201
5202
5203
5204
5205
5206
5207
5208
5209
5210
5211
5212
5213
5214
5215
5216
5217
5218
5219
5220
5221
5222
5223
5224
5225
5226
5227
5228
5229
5230
5231
5232
5233
5234
5235
5236
5237
5238
5239
5240
5241
5242
5243
5244
5245
5246
5247
5248
5249
5250
5251
5252
5253
5254
5255
5256
5257
5258
5259
5260
5261
5262
5263
5264
5265
5266
5267
5268
5269
5270
5271
5272
5273
5274
5275
5276
5277
5278
5279
5280
5281
|
/*************************************************
* Perl-Compatible Regular Expressions *
*************************************************/
/* PCRE is a library of functions to support regular expressions whose syntax
and semantics are as close as possible to those of the Perl 5 language.
Written by Philip Hazel
Copyright (c) 1997-2006 University of Cambridge
-----------------------------------------------------------------------------
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:
* Redistributions of source code must retain the above copyright notice,
this list of conditions and the following disclaimer.
* Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
* Neither the name of the University of Cambridge nor the names of its
contributors may be used to endorse or promote products derived from
this software without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.
-----------------------------------------------------------------------------
*/
/* This module contains the external function pcre_compile(), along with
supporting internal functions that are not used by other modules. */
#define NLBLOCK cd /* The block containing newline information */
#include "pcre_internal.h"
/* When DEBUG is defined, we need the pcre_printint() function, which is also
used by pcretest. DEBUG is not defined when building a production library. */
#ifdef DEBUG
#include "pcre_printint.src"
#endif
/*************************************************
* Code parameters and static tables *
*************************************************/
/* Maximum number of items on the nested bracket stacks at compile time. This
applies to the nesting of all kinds of parentheses. It does not limit
un-nested, non-capturing parentheses. This number can be made bigger if
necessary - it is used to dimension one int and one unsigned char vector at
compile time. */
#define BRASTACK_SIZE 200
/* Table for handling escaped characters in the range '0'-'z'. Positive returns
are simple data values; negative values are for special things like \d and so
on. Zero means further processing is needed (for things like \x), or the escape
is invalid. */
#if !EBCDIC /* This is the "normal" table for ASCII systems */
static const short int escapes[] = {
0, 0, 0, 0, 0, 0, 0, 0, /* 0 - 7 */
0, 0, ':', ';', '<', '=', '>', '?', /* 8 - ? */
'@', -ESC_A, -ESC_B, -ESC_C, -ESC_D, -ESC_E, 0, -ESC_G, /* @ - G */
0, 0, 0, 0, 0, 0, 0, 0, /* H - O */
-ESC_P, -ESC_Q, 0, -ESC_S, 0, 0, 0, -ESC_W, /* P - W */
-ESC_X, 0, -ESC_Z, '[', '\\', ']', '^', '_', /* X - _ */
'`', 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0, /* ` - g */
0, 0, 0, 0, 0, 0, ESC_n, 0, /* h - o */
-ESC_p, 0, ESC_r, -ESC_s, ESC_tee, 0, 0, -ESC_w, /* p - w */
0, 0, -ESC_z /* x - z */
};
#else /* This is the "abnormal" table for EBCDIC systems */
static const short int escapes[] = {
/* 48 */ 0, 0, 0, '.', '<', '(', '+', '|',
/* 50 */ '&', 0, 0, 0, 0, 0, 0, 0,
/* 58 */ 0, 0, '!', '$', '*', ')', ';', '~',
/* 60 */ '-', '/', 0, 0, 0, 0, 0, 0,
/* 68 */ 0, 0, '|', ',', '%', '_', '>', '?',
/* 70 */ 0, 0, 0, 0, 0, 0, 0, 0,
/* 78 */ 0, '`', ':', '#', '@', '\'', '=', '"',
/* 80 */ 0, 7, -ESC_b, 0, -ESC_d, ESC_e, ESC_f, 0,
/* 88 */ 0, 0, 0, '{', 0, 0, 0, 0,
/* 90 */ 0, 0, 0, 'l', 0, ESC_n, 0, -ESC_p,
/* 98 */ 0, ESC_r, 0, '}', 0, 0, 0, 0,
/* A0 */ 0, '~', -ESC_s, ESC_tee, 0, 0, -ESC_w, 0,
/* A8 */ 0,-ESC_z, 0, 0, 0, '[', 0, 0,
/* B0 */ 0, 0, 0, 0, 0, 0, 0, 0,
/* B8 */ 0, 0, 0, 0, 0, ']', '=', '-',
/* C0 */ '{',-ESC_A, -ESC_B, -ESC_C, -ESC_D,-ESC_E, 0, -ESC_G,
/* C8 */ 0, 0, 0, 0, 0, 0, 0, 0,
/* D0 */ '}', 0, 0, 0, 0, 0, 0, -ESC_P,
/* D8 */-ESC_Q, 0, 0, 0, 0, 0, 0, 0,
/* E0 */ '\\', 0, -ESC_S, 0, 0, 0, -ESC_W, -ESC_X,
/* E8 */ 0,-ESC_Z, 0, 0, 0, 0, 0, 0,
/* F0 */ 0, 0, 0, 0, 0, 0, 0, 0,
/* F8 */ 0, 0, 0, 0, 0, 0, 0, 0
};
#endif
/* Tables of names of POSIX character classes and their lengths. The list is
terminated by a zero length entry. The first three must be alpha, lower, upper,
as this is assumed for handling case independence. */
static const char *const posix_names[] = {
"alpha", "lower", "upper",
"alnum", "ascii", "blank", "cntrl", "digit", "graph",
"print", "punct", "space", "word", "xdigit" };
static const uschar posix_name_lengths[] = {
5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 4, 6, 0 };
/* Table of class bit maps for each POSIX class. Each class is formed from a
base map, with an optional addition or removal of another map. Then, for some
classes, there is some additional tweaking: for [:blank:] the vertical space
characters are removed, and for [:alpha:] and [:alnum:] the underscore
character is removed. The triples in the table consist of the base map offset,
second map offset or -1 if no second map, and a non-negative value for map
addition or a negative value for map subtraction (if there are two maps). The
absolute value of the third field has these meanings: 0 => no tweaking, 1 =>
remove vertical space characters, 2 => remove underscore. */
static const int posix_class_maps[] = {
cbit_word, cbit_digit, -2, /* alpha */
cbit_lower, -1, 0, /* lower */
cbit_upper, -1, 0, /* upper */
cbit_word, -1, 2, /* alnum - word without underscore */
cbit_print, cbit_cntrl, 0, /* ascii */
cbit_space, -1, 1, /* blank - a GNU extension */
cbit_cntrl, -1, 0, /* cntrl */
cbit_digit, -1, 0, /* digit */
cbit_graph, -1, 0, /* graph */
cbit_print, -1, 0, /* print */
cbit_punct, -1, 0, /* punct */
cbit_space, -1, 0, /* space */
cbit_word, -1, 0, /* word - a Perl extension */
cbit_xdigit,-1, 0 /* xdigit */
};
/* The texts of compile-time error messages. These are "char *" because they
are passed to the outside world. */
static const char *error_texts[] = {
"no error",
"\\ at end of pattern",
"\\c at end of pattern",
"unrecognized character follows \\",
"numbers out of order in {} quantifier",
/* 5 */
"number too big in {} quantifier",
"missing terminating ] for character class",
"invalid escape sequence in character class",
"range out of order in character class",
"nothing to repeat",
/* 10 */
"operand of unlimited repeat could match the empty string",
"internal error: unexpected repeat",
"unrecognized character after (?",
"POSIX named classes are supported only within a class",
"missing )",
/* 15 */
"reference to non-existent subpattern",
"erroffset passed as NULL",
"unknown option bit(s) set",
"missing ) after comment",
"parentheses nested too deeply",
/* 20 */
"regular expression too large",
"failed to get memory",
"unmatched parentheses",
"internal error: code overflow",
"unrecognized character after (?<",
/* 25 */
"lookbehind assertion is not fixed length",
"malformed number or name after (?(",
"conditional group contains more than two branches",
"assertion expected after (?(",
"(?R or (?digits must be followed by )",
/* 30 */
"unknown POSIX class name",
"POSIX collating elements are not supported",
"this version of PCRE is not compiled with PCRE_UTF8 support",
"spare error",
"character value in \\x{...} sequence is too large",
/* 35 */
"invalid condition (?(0)",
"\\C not allowed in lookbehind assertion",
"PCRE does not support \\L, \\l, \\N, \\U, or \\u",
"number after (?C is > 255",
"closing ) for (?C expected",
/* 40 */
"recursive call could loop indefinitely",
"unrecognized character after (?P",
"syntax error after (?P",
"two named subpatterns have the same name",
"invalid UTF-8 string",
/* 45 */
"support for \\P, \\p, and \\X has not been compiled",
"malformed \\P or \\p sequence",
"unknown property name after \\P or \\p",
"subpattern name is too long (maximum 32 characters)",
"too many named subpatterns (maximum 10,000)",
/* 50 */
"repeated subpattern is too long",
"octal value is greater than \\377 (not in UTF-8 mode)"
};
/* Table to identify digits and hex digits. This is used when compiling
patterns. Note that the tables in chartables are dependent on the locale, and
may mark arbitrary characters as digits - but the PCRE compiling code expects
to handle only 0-9, a-z, and A-Z as digits when compiling. That is why we have
a private table here. It costs 256 bytes, but it is a lot faster than doing
character value tests (at least in some simple cases I timed), and in some
applications one wants PCRE to compile efficiently as well as match
efficiently.
For convenience, we use the same bit definitions as in chartables:
0x04 decimal digit
0x08 hexadecimal digit
Then we can use ctype_digit and ctype_xdigit in the code. */
#if !EBCDIC /* This is the "normal" case, for ASCII systems */
static const unsigned char digitab[] =
{
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - ' */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ( - / */
0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 */
0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00, /* 8 - ? */
0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* @ - G */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H - O */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* P - W */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* X - _ */
0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* ` - g */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h - o */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* p - w */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* x -127 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 128-135 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 136-143 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144-151 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 152-159 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160-167 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 168-175 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 176-183 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 192-199 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 200-207 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 208-215 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 216-223 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 224-231 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 232-239 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 240-247 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00};/* 248-255 */
#else /* This is the "abnormal" case, for EBCDIC systems */
static const unsigned char digitab[] =
{
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 0- 7 0 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 8- 15 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 16- 23 10 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 32- 39 20 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 30 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 40 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 72- | */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 50 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 88- ¬ */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 60 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 104- ? */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 70 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */
0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* 128- g 80 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 144- p 90 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 160- x A0 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 B0 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
0x00,0x08,0x08,0x08,0x08,0x08,0x08,0x00, /* { - G C0 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* } - P D0 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* \ - X E0 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */
0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c,0x0c, /* 0 - 7 F0 */
0x0c,0x0c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */
static const unsigned char ebcdic_chartab[] = { /* chartable partial dup */
0x80,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 0- 7 */
0x00,0x00,0x00,0x00,0x01,0x01,0x00,0x00, /* 8- 15 */
0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 16- 23 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 24- 31 */
0x00,0x00,0x00,0x00,0x00,0x01,0x00,0x00, /* 32- 39 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 40- 47 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 48- 55 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 56- 63 */
0x01,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - 71 */
0x00,0x00,0x00,0x80,0x00,0x80,0x80,0x80, /* 72- | */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* & - 87 */
0x00,0x00,0x00,0x80,0x80,0x80,0x00,0x00, /* 88- ¬ */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* - -103 */
0x00,0x00,0x00,0x00,0x00,0x10,0x00,0x80, /* 104- ? */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 112-119 */
0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* 120- " */
0x00,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* 128- g */
0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* h -143 */
0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* 144- p */
0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* q -159 */
0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* 160- x */
0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* y -175 */
0x80,0x00,0x00,0x00,0x00,0x00,0x00,0x00, /* ^ -183 */
0x00,0x00,0x80,0x00,0x00,0x00,0x00,0x00, /* 184-191 */
0x80,0x1a,0x1a,0x1a,0x1a,0x1a,0x1a,0x12, /* { - G */
0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* H -207 */
0x00,0x12,0x12,0x12,0x12,0x12,0x12,0x12, /* } - P */
0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Q -223 */
0x00,0x00,0x12,0x12,0x12,0x12,0x12,0x12, /* \ - X */
0x12,0x12,0x00,0x00,0x00,0x00,0x00,0x00, /* Y -239 */
0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c,0x1c, /* 0 - 7 */
0x1c,0x1c,0x00,0x00,0x00,0x00,0x00,0x00};/* 8 -255 */
#endif
/* Definition to allow mutual recursion */
static BOOL
compile_regex(int, int, int *, uschar **, const uschar **, int *, BOOL, int,
int *, int *, branch_chain *, compile_data *);
/*************************************************
* Handle escapes *
*************************************************/
/* This function is called when a \ has been encountered. It either returns a
positive value for a simple escape such as \n, or a negative value which
encodes one of the more complicated things such as \d. When UTF-8 is enabled,
a positive value greater than 255 may be returned. On entry, ptr is pointing at
the \. On exit, it is on the final character of the escape sequence.
Arguments:
ptrptr points to the pattern position pointer
errorcodeptr points to the errorcode variable
bracount number of previous extracting brackets
options the options bits
isclass TRUE if inside a character class
Returns: zero or positive => a data character
negative => a special escape sequence
on error, errorptr is set
*/
static int
check_escape(const uschar **ptrptr, int *errorcodeptr, int bracount,
int options, BOOL isclass)
{
BOOL utf8 = (options & PCRE_UTF8) != 0;
const uschar *ptr = *ptrptr + 1;
int c, i;
GETCHARINCTEST(c, ptr); /* Get character value, increment pointer */
ptr--; /* Set pointer back to the last byte */
/* If backslash is at the end of the pattern, it's an error. */
if (c == 0) *errorcodeptr = ERR1;
/* Non-alphamerics are literals. For digits or letters, do an initial lookup in
a table. A non-zero result is something that can be returned immediately.
Otherwise further processing may be required. */
#if !EBCDIC /* ASCII coding */
else if (c < '0' || c > 'z') {} /* Not alphameric */
else if ((i = escapes[c - '0']) != 0) c = i;
#else /* EBCDIC coding */
else if (c < 'a' || (ebcdic_chartab[c] & 0x0E) == 0) {} /* Not alphameric */
else if ((i = escapes[c - 0x48]) != 0) c = i;
#endif
/* Escapes that need further processing, or are illegal. */
else
{
const uschar *oldptr;
switch (c)
{
/* A number of Perl escapes are not handled by PCRE. We give an explicit
error. */
case 'l':
case 'L':
case 'N':
case 'u':
case 'U':
*errorcodeptr = ERR37;
break;
/* The handling of escape sequences consisting of a string of digits
starting with one that is not zero is not straightforward. By experiment,
the way Perl works seems to be as follows:
Outside a character class, the digits are read as a decimal number. If the
number is less than 10, or if there are that many previous extracting
left brackets, then it is a back reference. Otherwise, up to three octal
digits are read to form an escaped byte. Thus \123 is likely to be octal
123 (cf \0123, which is octal 012 followed by the literal 3). If the octal
value is greater than 377, the least significant 8 bits are taken. Inside a
character class, \ followed by a digit is always an octal number. */
case '1': case '2': case '3': case '4': case '5':
case '6': case '7': case '8': case '9':
if (!isclass)
{
oldptr = ptr;
c -= '0';
while ((digitab[ptr[1]] & ctype_digit) != 0)
c = c * 10 + *(++ptr) - '0';
if (c < 10 || c <= bracount)
{
c = -(ESC_REF + c);
break;
}
ptr = oldptr; /* Put the pointer back and fall through */
}
/* Handle an octal number following \. If the first digit is 8 or 9, Perl
generates a binary zero byte and treats the digit as a following literal.
Thus we have to pull back the pointer by one. */
if ((c = *ptr) >= '8')
{
ptr--;
c = 0;
break;
}
/* \0 always starts an octal number, but we may drop through to here with a
larger first octal digit. The original code used just to take the least
significant 8 bits of octal numbers (I think this is what early Perls used
to do). Nowadays we allow for larger numbers in UTF-8 mode, but no more
than 3 octal digits. */
case '0':
c -= '0';
while(i++ < 2 && ptr[1] >= '0' && ptr[1] <= '7')
c = c * 8 + *(++ptr) - '0';
if (!utf8 && c > 255) *errorcodeptr = ERR51;
break;
/* \x is complicated. \x{ddd} is a character number which can be greater
than 0xff in utf8 mode, but only if the ddd are hex digits. If not, { is
treated as a data character. */
case 'x':
if (ptr[1] == '{')
{
const uschar *pt = ptr + 2;
int count = 0;
c = 0;
while ((digitab[*pt] & ctype_xdigit) != 0)
{
register int cc = *pt++;
if (c == 0 && cc == '0') continue; /* Leading zeroes */
count++;
#if !EBCDIC /* ASCII coding */
if (cc >= 'a') cc -= 32; /* Convert to upper case */
c = (c << 4) + cc - ((cc < 'A')? '0' : ('A' - 10));
#else /* EBCDIC coding */
if (cc >= 'a' && cc <= 'z') cc += 64; /* Convert to upper case */
c = (c << 4) + cc - ((cc >= '0')? '0' : ('A' - 10));
#endif
}
if (*pt == '}')
{
if (c < 0 || count > (utf8? 8 : 2)) *errorcodeptr = ERR34;
ptr = pt;
break;
}
/* If the sequence of hex digits does not end with '}', then we don't
recognize this construct; fall through to the normal \x handling. */
}
/* Read just a single-byte hex-defined char */
c = 0;
while (i++ < 2 && (digitab[ptr[1]] & ctype_xdigit) != 0)
{
int cc; /* Some compilers don't like ++ */
cc = *(++ptr); /* in initializers */
#if !EBCDIC /* ASCII coding */
if (cc >= 'a') cc -= 32; /* Convert to upper case */
c = c * 16 + cc - ((cc < 'A')? '0' : ('A' - 10));
#else /* EBCDIC coding */
if (cc <= 'z') cc += 64; /* Convert to upper case */
c = c * 16 + cc - ((cc >= '0')? '0' : ('A' - 10));
#endif
}
break;
/* Other special escapes not starting with a digit are straightforward */
case 'c':
c = *(++ptr);
if (c == 0)
{
*errorcodeptr = ERR2;
return 0;
}
/* A letter is upper-cased; then the 0x40 bit is flipped. This coding
is ASCII-specific, but then the whole concept of \cx is ASCII-specific.
(However, an EBCDIC equivalent has now been added.) */
#if !EBCDIC /* ASCII coding */
if (c >= 'a' && c <= 'z') c -= 32;
c ^= 0x40;
#else /* EBCDIC coding */
if (c >= 'a' && c <= 'z') c += 64;
c ^= 0xC0;
#endif
break;
/* PCRE_EXTRA enables extensions to Perl in the matter of escapes. Any
other alphameric following \ is an error if PCRE_EXTRA was set; otherwise,
for Perl compatibility, it is a literal. This code looks a bit odd, but
there used to be some cases other than the default, and there may be again
in future, so I haven't "optimized" it. */
default:
if ((options & PCRE_EXTRA) != 0) switch(c)
{
default:
*errorcodeptr = ERR3;
break;
}
break;
}
}
*ptrptr = ptr;
return c;
}
#ifdef SUPPORT_UCP
/*************************************************
* Handle \P and \p *
*************************************************/
/* This function is called after \P or \p has been encountered, provided that
PCRE is compiled with support for Unicode properties. On entry, ptrptr is
pointing at the P or p. On exit, it is pointing at the final character of the
escape sequence.
Argument:
ptrptr points to the pattern position pointer
negptr points to a boolean that is set TRUE for negation else FALSE
dptr points to an int that is set to the detailed property value
errorcodeptr points to the error code variable
Returns: type value from ucp_type_table, or -1 for an invalid type
*/
static int
get_ucp(const uschar **ptrptr, BOOL *negptr, int *dptr, int *errorcodeptr)
{
int c, i, bot, top;
const uschar *ptr = *ptrptr;
char name[32];
c = *(++ptr);
if (c == 0) goto ERROR_RETURN;
*negptr = FALSE;
/* \P or \p can be followed by a name in {}, optionally preceded by ^ for
negation. */
if (c == '{')
{
if (ptr[1] == '^')
{
*negptr = TRUE;
ptr++;
}
for (i = 0; i < sizeof(name) - 1; i++)
{
c = *(++ptr);
if (c == 0) goto ERROR_RETURN;
if (c == '}') break;
name[i] = c;
}
if (c !='}') goto ERROR_RETURN;
name[i] = 0;
}
/* Otherwise there is just one following character */
else
{
name[0] = c;
name[1] = 0;
}
*ptrptr = ptr;
/* Search for a recognized property name using binary chop */
bot = 0;
top = _pcre_utt_size;
while (bot < top)
{
i = (bot + top) >> 1;
c = strcmp(name, _pcre_utt[i].name);
if (c == 0)
{
*dptr = _pcre_utt[i].value;
return _pcre_utt[i].type;
}
if (c > 0) bot = i + 1; else top = i;
}
*errorcodeptr = ERR47;
*ptrptr = ptr;
return -1;
ERROR_RETURN:
*errorcodeptr = ERR46;
*ptrptr = ptr;
return -1;
}
#endif
/*************************************************
* Check for counted repeat *
*************************************************/
/* This function is called when a '{' is encountered in a place where it might
start a quantifier. It looks ahead to see if it really is a quantifier or not.
It is only a quantifier if it is one of the forms {ddd} {ddd,} or {ddd,ddd}
where the ddds are digits.
Arguments:
p pointer to the first char after '{'
Returns: TRUE or FALSE
*/
static BOOL
is_counted_repeat(const uschar *p)
{
if ((digitab[*p++] & ctype_digit) == 0) return FALSE;
while ((digitab[*p] & ctype_digit) != 0) p++;
if (*p == '}') return TRUE;
if (*p++ != ',') return FALSE;
if (*p == '}') return TRUE;
if ((digitab[*p++] & ctype_digit) == 0) return FALSE;
while ((digitab[*p] & ctype_digit) != 0) p++;
return (*p == '}');
}
/*************************************************
* Read repeat counts *
*************************************************/
/* Read an item of the form {n,m} and return the values. This is called only
after is_counted_repeat() has confirmed that a repeat-count quantifier exists,
so the syntax is guaranteed to be correct, but we need to check the values.
Arguments:
p pointer to first char after '{'
minp pointer to int for min
maxp pointer to int for max
returned as -1 if no max
errorcodeptr points to error code variable
Returns: pointer to '}' on success;
current ptr on error, with errorcodeptr set non-zero
*/
static const uschar *
read_repeat_counts(const uschar *p, int *minp, int *maxp, int *errorcodeptr)
{
int min = 0;
int max = -1;
/* Read the minimum value and do a paranoid check: a negative value indicates
an integer overflow. */
while ((digitab[*p] & ctype_digit) != 0) min = min * 10 + *p++ - '0';
if (min < 0 || min > 65535)
{
*errorcodeptr = ERR5;
return p;
}
/* Read the maximum value if there is one, and again do a paranoid on its size.
Also, max must not be less than min. */
if (*p == '}') max = min; else
{
if (*(++p) != '}')
{
max = 0;
while((digitab[*p] & ctype_digit) != 0) max = max * 10 + *p++ - '0';
if (max < 0 || max > 65535)
{
*errorcodeptr = ERR5;
return p;
}
if (max < min)
{
*errorcodeptr = ERR4;
return p;
}
}
}
/* Fill in the required variables, and pass back the pointer to the terminating
'}'. */
*minp = min;
*maxp = max;
return p;
}
/*************************************************
* Find forward referenced named subpattern *
*************************************************/
/* This function scans along a pattern looking for capturing subpatterns, and
counting them. If it finds a named pattern that matches the name it is given,
it returns its number. This is used for forward references to named
subpatterns. We know that if (?P< is encountered, the name will be terminated
by '>' because that is checked in the first pass.
Arguments:
pointer current position in the pattern
count current count of capturing parens
name name to seek
namelen name length
Returns: the number of the named subpattern, or -1 if not found
*/
static int
find_named_parens(const uschar *ptr, int count, const uschar *name, int namelen)
{
const uschar *thisname;
for (; *ptr != 0; ptr++)
{
if (*ptr == '\\' && ptr[1] != 0) { ptr++; continue; }
if (*ptr != '(') continue;
if (ptr[1] != '?') { count++; continue; }
if (ptr[2] == '(') { ptr += 2; continue; }
if (ptr[2] != 'P' || ptr[3] != '<') continue;
count++;
ptr += 4;
thisname = ptr;
while (*ptr != '>') ptr++;
if (namelen == ptr - thisname && strncmp(name, thisname, namelen) == 0)
return count;
}
return -1;
}
/*************************************************
* Find first significant op code *
*************************************************/
/* This is called by several functions that scan a compiled expression looking
for a fixed first character, or an anchoring op code etc. It skips over things
that do not influence this. For some calls, a change of option is important.
For some calls, it makes sense to skip negative forward and all backward
assertions, and also the \b assertion; for others it does not.
Arguments:
code pointer to the start of the group
options pointer to external options
optbit the option bit whose changing is significant, or
zero if none are
skipassert TRUE if certain assertions are to be skipped
Returns: pointer to the first significant opcode
*/
static const uschar*
first_significant_code(const uschar *code, int *options, int optbit,
BOOL skipassert)
{
for (;;)
{
switch ((int)*code)
{
case OP_OPT:
if (optbit > 0 && ((int)code[1] & optbit) != (*options & optbit))
*options = (int)code[1];
code += 2;
break;
case OP_ASSERT_NOT:
case OP_ASSERTBACK:
case OP_ASSERTBACK_NOT:
if (!skipassert) return code;
do code += GET(code, 1); while (*code == OP_ALT);
code += _pcre_OP_lengths[*code];
break;
case OP_WORD_BOUNDARY:
case OP_NOT_WORD_BOUNDARY:
if (!skipassert) return code;
/* Fall through */
case OP_CALLOUT:
case OP_CREF:
case OP_BRANUMBER:
code += _pcre_OP_lengths[*code];
break;
default:
return code;
}
}
/* Control never reaches here */
}
/*************************************************
* Find the fixed length of a pattern *
*************************************************/
/* Scan a pattern and compute the fixed length of subject that will match it,
if the length is fixed. This is needed for dealing with backward assertions.
In UTF8 mode, the result is in characters rather than bytes.
Arguments:
code points to the start of the pattern (the bracket)
options the compiling options
Returns: the fixed length, or -1 if there is no fixed length,
or -2 if \C was encountered
*/
static int
find_fixedlength(uschar *code, int options)
{
int length = -1;
register int branchlength = 0;
register uschar *cc = code + 1 + LINK_SIZE;
/* Scan along the opcodes for this branch. If we get to the end of the
branch, check the length against that of the other branches. */
for (;;)
{
int d;
register int op = *cc;
if (op >= OP_BRA) op = OP_BRA;
switch (op)
{
case OP_BRA:
case OP_ONCE:
case OP_COND:
d = find_fixedlength(cc, options);
if (d < 0) return d;
branchlength += d;
do cc += GET(cc, 1); while (*cc == OP_ALT);
cc += 1 + LINK_SIZE;
break;
/* Reached end of a branch; if it's a ket it is the end of a nested
call. If it's ALT it is an alternation in a nested call. If it is
END it's the end of the outer call. All can be handled by the same code. */
case OP_ALT:
case OP_KET:
case OP_KETRMAX:
case OP_KETRMIN:
case OP_END:
if (length < 0) length = branchlength;
else if (length != branchlength) return -1;
if (*cc != OP_ALT) return length;
cc += 1 + LINK_SIZE;
branchlength = 0;
break;
/* Skip over assertive subpatterns */
case OP_ASSERT:
case OP_ASSERT_NOT:
case OP_ASSERTBACK:
case OP_ASSERTBACK_NOT:
do cc += GET(cc, 1); while (*cc == OP_ALT);
/* Fall through */
/* Skip over things that don't match chars */
case OP_REVERSE:
case OP_BRANUMBER:
case OP_CREF:
case OP_OPT:
case OP_CALLOUT:
case OP_SOD:
case OP_SOM:
case OP_EOD:
case OP_EODN:
case OP_CIRC:
case OP_DOLL:
case OP_NOT_WORD_BOUNDARY:
case OP_WORD_BOUNDARY:
cc += _pcre_OP_lengths[*cc];
break;
/* Handle literal characters */
case OP_CHAR:
case OP_CHARNC:
case OP_NOT:
branchlength++;
cc += 2;
#ifdef SUPPORT_UTF8
if ((options & PCRE_UTF8) != 0)
{
while ((*cc & 0xc0) == 0x80) cc++;
}
#endif
break;
/* Handle exact repetitions. The count is already in characters, but we
need to skip over a multibyte character in UTF8 mode. */
case OP_EXACT:
branchlength += GET2(cc,1);
cc += 4;
#ifdef SUPPORT_UTF8
if ((options & PCRE_UTF8) != 0)
{
while((*cc & 0x80) == 0x80) cc++;
}
#endif
break;
case OP_TYPEEXACT:
branchlength += GET2(cc,1);
cc += 4;
break;
/* Handle single-char matchers */
case OP_PROP:
case OP_NOTPROP:
cc += 2;
/* Fall through */
case OP_NOT_DIGIT:
case OP_DIGIT:
case OP_NOT_WHITESPACE:
case OP_WHITESPACE:
case OP_NOT_WORDCHAR:
case OP_WORDCHAR:
case OP_ANY:
branchlength++;
cc++;
break;
/* The single-byte matcher isn't allowed */
case OP_ANYBYTE:
return -2;
/* Check a class for variable quantification */
#ifdef SUPPORT_UTF8
case OP_XCLASS:
cc += GET(cc, 1) - 33;
/* Fall through */
#endif
case OP_CLASS:
case OP_NCLASS:
cc += 33;
switch (*cc)
{
case OP_CRSTAR:
case OP_CRMINSTAR:
case OP_CRQUERY:
case OP_CRMINQUERY:
return -1;
case OP_CRRANGE:
case OP_CRMINRANGE:
if (GET2(cc,1) != GET2(cc,3)) return -1;
branchlength += GET2(cc,1);
cc += 5;
break;
default:
branchlength++;
}
break;
/* Anything else is variable length */
default:
return -1;
}
}
/* Control never gets here */
}
/*************************************************
* Scan compiled regex for numbered bracket *
*************************************************/
/* This little function scans through a compiled pattern until it finds a
capturing bracket with the given number.
Arguments:
code points to start of expression
utf8 TRUE in UTF-8 mode
number the required bracket number
Returns: pointer to the opcode for the bracket, or NULL if not found
*/
static const uschar *
find_bracket(const uschar *code, BOOL utf8, int number)
{
for (;;)
{
register int c = *code;
if (c == OP_END) return NULL;
/* XCLASS is used for classes that cannot be represented just by a bit
map. This includes negated single high-valued characters. The length in
the table is zero; the actual length is stored in the compiled code. */
if (c == OP_XCLASS) code += GET(code, 1);
/* Handle bracketed group */
else if (c > OP_BRA)
{
int n = c - OP_BRA;
if (n > EXTRACT_BASIC_MAX) n = GET2(code, 2+LINK_SIZE);
if (n == number) return (uschar *)code;
code += _pcre_OP_lengths[OP_BRA];
}
/* Otherwise, we get the item's length from the table. In UTF-8 mode, opcodes
that are followed by a character may be followed by a multi-byte character.
The length in the table is a minimum, so we have to scan along to skip the
extra bytes. All opcodes are less than 128, so we can use relatively
efficient code. */
else
{
code += _pcre_OP_lengths[c];
if (utf8) switch(c)
{
case OP_CHAR:
case OP_CHARNC:
case OP_EXACT:
case OP_UPTO:
case OP_MINUPTO:
case OP_STAR:
case OP_MINSTAR:
case OP_PLUS:
case OP_MINPLUS:
case OP_QUERY:
case OP_MINQUERY:
while ((*code & 0xc0) == 0x80) code++;
break;
}
}
}
}
/*************************************************
* Scan compiled regex for recursion reference *
*************************************************/
/* This little function scans through a compiled pattern until it finds an
instance of OP_RECURSE.
Arguments:
code points to start of expression
utf8 TRUE in UTF-8 mode
Returns: pointer to the opcode for OP_RECURSE, or NULL if not found
*/
static const uschar *
find_recurse(const uschar *code, BOOL utf8)
{
for (;;)
{
register int c = *code;
if (c == OP_END) return NULL;
if (c == OP_RECURSE) return code;
/* XCLASS is used for classes that cannot be represented just by a bit
map. This includes negated single high-valued characters. The length in
the table is zero; the actual length is stored in the compiled code. */
if (c == OP_XCLASS) code += GET(code, 1);
/* All bracketed groups have the same length. */
else if (c > OP_BRA)
{
code += _pcre_OP_lengths[OP_BRA];
}
/* Otherwise, we get the item's length from the table. In UTF-8 mode, opcodes
that are followed by a character may be followed by a multi-byte character.
The length in the table is a minimum, so we have to scan along to skip the
extra bytes. All opcodes are less than 128, so we can use relatively
efficient code. */
else
{
code += _pcre_OP_lengths[c];
if (utf8) switch(c)
{
case OP_CHAR:
case OP_CHARNC:
case OP_EXACT:
case OP_UPTO:
case OP_MINUPTO:
case OP_STAR:
case OP_MINSTAR:
case OP_PLUS:
case OP_MINPLUS:
case OP_QUERY:
case OP_MINQUERY:
while ((*code & 0xc0) == 0x80) code++;
break;
}
}
}
}
/*************************************************
* Scan compiled branch for non-emptiness *
*************************************************/
/* This function scans through a branch of a compiled pattern to see whether it
can match the empty string or not. It is called only from could_be_empty()
below. Note that first_significant_code() skips over assertions. If we hit an
unclosed bracket, we return "empty" - this means we've struck an inner bracket
whose current branch will already have been scanned.
Arguments:
code points to start of search
endcode points to where to stop
utf8 TRUE if in UTF8 mode
Returns: TRUE if what is matched could be empty
*/
static BOOL
could_be_empty_branch(const uschar *code, const uschar *endcode, BOOL utf8)
{
register int c;
for (code = first_significant_code(code + 1 + LINK_SIZE, NULL, 0, TRUE);
code < endcode;
code = first_significant_code(code + _pcre_OP_lengths[c], NULL, 0, TRUE))
{
const uschar *ccode;
c = *code;
if (c >= OP_BRA)
{
BOOL empty_branch;
if (GET(code, 1) == 0) return TRUE; /* Hit unclosed bracket */
/* Scan a closed bracket */
empty_branch = FALSE;
do
{
if (!empty_branch && could_be_empty_branch(code, endcode, utf8))
empty_branch = TRUE;
code += GET(code, 1);
}
while (*code == OP_ALT);
if (!empty_branch) return FALSE; /* All branches are non-empty */
code += 1 + LINK_SIZE;
c = *code;
}
else switch (c)
{
/* Check for quantifiers after a class */
#ifdef SUPPORT_UTF8
case OP_XCLASS:
ccode = code + GET(code, 1);
goto CHECK_CLASS_REPEAT;
#endif
case OP_CLASS:
case OP_NCLASS:
ccode = code + 33;
#ifdef SUPPORT_UTF8
CHECK_CLASS_REPEAT:
#endif
switch (*ccode)
{
case OP_CRSTAR: /* These could be empty; continue */
case OP_CRMINSTAR:
case OP_CRQUERY:
case OP_CRMINQUERY:
break;
default: /* Non-repeat => class must match */
case OP_CRPLUS: /* These repeats aren't empty */
case OP_CRMINPLUS:
return FALSE;
case OP_CRRANGE:
case OP_CRMINRANGE:
if (GET2(ccode, 1) > 0) return FALSE; /* Minimum > 0 */
break;
}
break;
/* Opcodes that must match a character */
case OP_PROP:
case OP_NOTPROP:
case OP_EXTUNI:
case OP_NOT_DIGIT:
case OP_DIGIT:
case OP_NOT_WHITESPACE:
case OP_WHITESPACE:
case OP_NOT_WORDCHAR:
case OP_WORDCHAR:
case OP_ANY:
case OP_ANYBYTE:
case OP_CHAR:
case OP_CHARNC:
case OP_NOT:
case OP_PLUS:
case OP_MINPLUS:
case OP_EXACT:
case OP_NOTPLUS:
case OP_NOTMINPLUS:
case OP_NOTEXACT:
case OP_TYPEPLUS:
case OP_TYPEMINPLUS:
case OP_TYPEEXACT:
return FALSE;
/* End of branch */
case OP_KET:
case OP_KETRMAX:
case OP_KETRMIN:
case OP_ALT:
return TRUE;
/* In UTF-8 mode, STAR, MINSTAR, QUERY, MINQUERY, UPTO, and MINUPTO may be
followed by a multibyte character */
#ifdef SUPPORT_UTF8
case OP_STAR:
case OP_MINSTAR:
case OP_QUERY:
case OP_MINQUERY:
case OP_UPTO:
case OP_MINUPTO:
if (utf8) while ((code[2] & 0xc0) == 0x80) code++;
break;
#endif
}
}
return TRUE;
}
/*************************************************
* Scan compiled regex for non-emptiness *
*************************************************/
/* This function is called to check for left recursive calls. We want to check
the current branch of the current pattern to see if it could match the empty
string. If it could, we must look outwards for branches at other levels,
stopping when we pass beyond the bracket which is the subject of the recursion.
Arguments:
code points to start of the recursion
endcode points to where to stop (current RECURSE item)
bcptr points to the chain of current (unclosed) branch starts
utf8 TRUE if in UTF-8 mode
Returns: TRUE if what is matched could be empty
*/
static BOOL
could_be_empty(const uschar *code, const uschar *endcode, branch_chain *bcptr,
BOOL utf8)
{
while (bcptr != NULL && bcptr->current >= code)
{
if (!could_be_empty_branch(bcptr->current, endcode, utf8)) return FALSE;
bcptr = bcptr->outer;
}
return TRUE;
}
/*************************************************
* Check for POSIX class syntax *
*************************************************/
/* This function is called when the sequence "[:" or "[." or "[=" is
encountered in a character class. It checks whether this is followed by an
optional ^ and then a sequence of letters, terminated by a matching ":]" or
".]" or "=]".
Argument:
ptr pointer to the initial [
endptr where to return the end pointer
cd pointer to compile data
Returns: TRUE or FALSE
*/
static BOOL
check_posix_syntax(const uschar *ptr, const uschar **endptr, compile_data *cd)
{
int terminator; /* Don't combine these lines; the Solaris cc */
terminator = *(++ptr); /* compiler warns about "non-constant" initializer. */
if (*(++ptr) == '^') ptr++;
while ((cd->ctypes[*ptr] & ctype_letter) != 0) ptr++;
if (*ptr == terminator && ptr[1] == ']')
{
*endptr = ptr;
return TRUE;
}
return FALSE;
}
/*************************************************
* Check POSIX class name *
*************************************************/
/* This function is called to check the name given in a POSIX-style class entry
such as [:alnum:].
Arguments:
ptr points to the first letter
len the length of the name
Returns: a value representing the name, or -1 if unknown
*/
static int
check_posix_name(const uschar *ptr, int len)
{
register int yield = 0;
while (posix_name_lengths[yield] != 0)
{
if (len == posix_name_lengths[yield] &&
strncmp((const char *)ptr, posix_names[yield], len) == 0) return yield;
yield++;
}
return -1;
}
/*************************************************
* Adjust OP_RECURSE items in repeated group *
*************************************************/
/* OP_RECURSE items contain an offset from the start of the regex to the group
that is referenced. This means that groups can be replicated for fixed
repetition simply by copying (because the recursion is allowed to refer to
earlier groups that are outside the current group). However, when a group is
optional (i.e. the minimum quantifier is zero), OP_BRAZERO is inserted before
it, after it has been compiled. This means that any OP_RECURSE items within it
that refer to the group itself or any contained groups have to have their
offsets adjusted. That is the job of this function. Before it is called, the
partially compiled regex must be temporarily terminated with OP_END.
Arguments:
group points to the start of the group
adjust the amount by which the group is to be moved
utf8 TRUE in UTF-8 mode
cd contains pointers to tables etc.
Returns: nothing
*/
static void
adjust_recurse(uschar *group, int adjust, BOOL utf8, compile_data *cd)
{
uschar *ptr = group;
while ((ptr = (uschar *)find_recurse(ptr, utf8)) != NULL)
{
int offset = GET(ptr, 1);
if (cd->start_code + offset >= group) PUT(ptr, 1, offset + adjust);
ptr += 1 + LINK_SIZE;
}
}
/*************************************************
* Insert an automatic callout point *
*************************************************/
/* This function is called when the PCRE_AUTO_CALLOUT option is set, to insert
callout points before each pattern item.
Arguments:
code current code pointer
ptr current pattern pointer
cd pointers to tables etc
Returns: new code pointer
*/
static uschar *
auto_callout(uschar *code, const uschar *ptr, compile_data *cd)
{
*code++ = OP_CALLOUT;
*code++ = 255;
PUT(code, 0, ptr - cd->start_pattern); /* Pattern offset */
PUT(code, LINK_SIZE, 0); /* Default length */
return code + 2*LINK_SIZE;
}
/*************************************************
* Complete a callout item *
*************************************************/
/* A callout item contains the length of the next item in the pattern, which
we can't fill in till after we have reached the relevant point. This is used
for both automatic and manual callouts.
Arguments:
previous_callout points to previous callout item
ptr current pattern pointer
cd pointers to tables etc
Returns: nothing
*/
static void
complete_callout(uschar *previous_callout, const uschar *ptr, compile_data *cd)
{
int length = ptr - cd->start_pattern - GET(previous_callout, 2);
PUT(previous_callout, 2 + LINK_SIZE, length);
}
#ifdef SUPPORT_UCP
/*************************************************
* Get othercase range *
*************************************************/
/* This function is passed the start and end of a class range, in UTF-8 mode
with UCP support. It searches up the characters, looking for internal ranges of
characters in the "other" case. Each call returns the next one, updating the
start address.
Arguments:
cptr points to starting character value; updated
d end value
ocptr where to put start of othercase range
odptr where to put end of othercase range
Yield: TRUE when range returned; FALSE when no more
*/
static BOOL
get_othercase_range(int *cptr, int d, int *ocptr, int *odptr)
{
int c, othercase, next;
for (c = *cptr; c <= d; c++)
{ if ((othercase = _pcre_ucp_othercase(c)) >= 0) break; }
if (c > d) return FALSE;
*ocptr = othercase;
next = othercase + 1;
for (++c; c <= d; c++)
{
if (_pcre_ucp_othercase(c) != next) break;
next++;
}
*odptr = next - 1;
*cptr = c;
return TRUE;
}
#endif /* SUPPORT_UCP */
/*************************************************
* Compile one branch *
*************************************************/
/* Scan the pattern, compiling it into the code vector. If the options are
changed during the branch, the pointer is used to change the external options
bits.
Arguments:
optionsptr pointer to the option bits
brackets points to number of extracting brackets used
codeptr points to the pointer to the current code point
ptrptr points to the current pattern pointer
errorcodeptr points to error code variable
firstbyteptr set to initial literal character, or < 0 (REQ_UNSET, REQ_NONE)
reqbyteptr set to the last literal character required, else < 0
bcptr points to current branch chain
cd contains pointers to tables etc.
Returns: TRUE on success
FALSE, with *errorcodeptr set non-zero on error
*/
static BOOL
compile_branch(int *optionsptr, int *brackets, uschar **codeptr,
const uschar **ptrptr, int *errorcodeptr, int *firstbyteptr,
int *reqbyteptr, branch_chain *bcptr, compile_data *cd)
{
int repeat_type, op_type;
int repeat_min = 0, repeat_max = 0; /* To please picky compilers */
int bravalue = 0;
int greedy_default, greedy_non_default;
int firstbyte, reqbyte;
int zeroreqbyte, zerofirstbyte;
int req_caseopt, reqvary, tempreqvary;
int options = *optionsptr;
int after_manual_callout = 0;
register int c;
register uschar *code = *codeptr;
uschar *tempcode;
BOOL inescq = FALSE;
BOOL groupsetfirstbyte = FALSE;
const uschar *ptr = *ptrptr;
const uschar *tempptr;
uschar *previous = NULL;
uschar *previous_callout = NULL;
uschar classbits[32];
#ifdef SUPPORT_UTF8
BOOL class_utf8;
BOOL utf8 = (options & PCRE_UTF8) != 0;
uschar *class_utf8data;
uschar utf8_char[6];
#else
BOOL utf8 = FALSE;
#endif
/* Set up the default and non-default settings for greediness */
greedy_default = ((options & PCRE_UNGREEDY) != 0);
greedy_non_default = greedy_default ^ 1;
/* Initialize no first byte, no required byte. REQ_UNSET means "no char
matching encountered yet". It gets changed to REQ_NONE if we hit something that
matches a non-fixed char first char; reqbyte just remains unset if we never
find one.
When we hit a repeat whose minimum is zero, we may have to adjust these values
to take the zero repeat into account. This is implemented by setting them to
zerofirstbyte and zeroreqbyte when such a repeat is encountered. The individual
item types that can be repeated set these backoff variables appropriately. */
firstbyte = reqbyte = zerofirstbyte = zeroreqbyte = REQ_UNSET;
/* The variable req_caseopt contains either the REQ_CASELESS value or zero,
according to the current setting of the caseless flag. REQ_CASELESS is a bit
value > 255. It is added into the firstbyte or reqbyte variables to record the
case status of the value. This is used only for ASCII characters. */
req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;
/* Switch on next character until the end of the branch */
for (;; ptr++)
{
BOOL negate_class;
BOOL possessive_quantifier;
BOOL is_quantifier;
int class_charcount;
int class_lastchar;
int newoptions;
int recno;
int skipbytes;
int subreqbyte;
int subfirstbyte;
int mclength;
uschar mcbuffer[8];
/* Next byte in the pattern */
c = *ptr;
/* If in \Q...\E, check for the end; if not, we have a literal */
if (inescq && c != 0)
{
if (c == '\\' && ptr[1] == 'E')
{
inescq = FALSE;
ptr++;
continue;
}
else
{
if (previous_callout != NULL)
{
complete_callout(previous_callout, ptr, cd);
previous_callout = NULL;
}
if ((options & PCRE_AUTO_CALLOUT) != 0)
{
previous_callout = code;
code = auto_callout(code, ptr, cd);
}
goto NORMAL_CHAR;
}
}
/* Fill in length of a previous callout, except when the next thing is
a quantifier. */
is_quantifier = c == '*' || c == '+' || c == '?' ||
(c == '{' && is_counted_repeat(ptr+1));
if (!is_quantifier && previous_callout != NULL &&
after_manual_callout-- <= 0)
{
complete_callout(previous_callout, ptr, cd);
previous_callout = NULL;
}
/* In extended mode, skip white space and comments */
if ((options & PCRE_EXTENDED) != 0)
{
if ((cd->ctypes[c] & ctype_space) != 0) continue;
if (c == '#')
{
while (*(++ptr) != 0) if (IS_NEWLINE(ptr)) break;
if (*ptr != 0)
{
ptr += cd->nllen - 1;
continue;
}
/* Else fall through to handle end of string */
c = 0;
}
}
/* No auto callout for quantifiers. */
if ((options & PCRE_AUTO_CALLOUT) != 0 && !is_quantifier)
{
previous_callout = code;
code = auto_callout(code, ptr, cd);
}
switch(c)
{
/* The branch terminates at end of string, |, or ). */
case 0:
case '|':
case ')':
*firstbyteptr = firstbyte;
*reqbyteptr = reqbyte;
*codeptr = code;
*ptrptr = ptr;
return TRUE;
/* Handle single-character metacharacters. In multiline mode, ^ disables
the setting of any following char as a first character. */
case '^':
if ((options & PCRE_MULTILINE) != 0)
{
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
}
previous = NULL;
*code++ = OP_CIRC;
break;
case '$':
previous = NULL;
*code++ = OP_DOLL;
break;
/* There can never be a first char if '.' is first, whatever happens about
repeats. The value of reqbyte doesn't change either. */
case '.':
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
zerofirstbyte = firstbyte;
zeroreqbyte = reqbyte;
previous = code;
*code++ = OP_ANY;
break;
/* Character classes. If the included characters are all < 256, we build a
32-byte bitmap of the permitted characters, except in the special case
where there is only one such character. For negated classes, we build the
map as usual, then invert it at the end. However, we use a different opcode
so that data characters > 255 can be handled correctly.
If the class contains characters outside the 0-255 range, a different
opcode is compiled. It may optionally have a bit map for characters < 256,
but those above are are explicitly listed afterwards. A flag byte tells
whether the bitmap is present, and whether this is a negated class or not.
*/
case '[':
previous = code;
/* PCRE supports POSIX class stuff inside a class. Perl gives an error if
they are encountered at the top level, so we'll do that too. */
if ((ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') &&
check_posix_syntax(ptr, &tempptr, cd))
{
*errorcodeptr = (ptr[1] == ':')? ERR13 : ERR31;
goto FAILED;
}
/* If the first character is '^', set the negation flag and skip it. */
if ((c = *(++ptr)) == '^')
{
negate_class = TRUE;
c = *(++ptr);
}
else
{
negate_class = FALSE;
}
/* Keep a count of chars with values < 256 so that we can optimize the case
of just a single character (as long as it's < 256). For higher valued UTF-8
characters, we don't yet do any optimization. */
class_charcount = 0;
class_lastchar = -1;
#ifdef SUPPORT_UTF8
class_utf8 = FALSE; /* No chars >= 256 */
class_utf8data = code + LINK_SIZE + 34; /* For UTF-8 items */
#endif
/* Initialize the 32-char bit map to all zeros. We have to build the
map in a temporary bit of store, in case the class contains only 1
character (< 256), because in that case the compiled code doesn't use the
bit map. */
memset(classbits, 0, 32 * sizeof(uschar));
/* Process characters until ] is reached. By writing this as a "do" it
means that an initial ] is taken as a data character. The first pass
through the regex checked the overall syntax, so we don't need to be very
strict here. At the start of the loop, c contains the first byte of the
character. */
do
{
#ifdef SUPPORT_UTF8
if (utf8 && c > 127)
{ /* Braces are required because the */
GETCHARLEN(c, ptr, ptr); /* macro generates multiple statements */
}
#endif
/* Inside \Q...\E everything is literal except \E */
if (inescq)
{
if (c == '\\' && ptr[1] == 'E')
{
inescq = FALSE;
ptr++;
continue;
}
else goto LONE_SINGLE_CHARACTER;
}
/* Handle POSIX class names. Perl allows a negation extension of the
form [:^name:]. A square bracket that doesn't match the syntax is
treated as a literal. We also recognize the POSIX constructions
[.ch.] and [=ch=] ("collating elements") and fault them, as Perl
5.6 and 5.8 do. */
if (c == '[' &&
(ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') &&
check_posix_syntax(ptr, &tempptr, cd))
{
BOOL local_negate = FALSE;
int posix_class, taboffset, tabopt;
register const uschar *cbits = cd->cbits;
uschar pbits[32];
if (ptr[1] != ':')
{
*errorcodeptr = ERR31;
goto FAILED;
}
ptr += 2;
if (*ptr == '^')
{
local_negate = TRUE;
ptr++;
}
posix_class = check_posix_name(ptr, tempptr - ptr);
if (posix_class < 0)
{
*errorcodeptr = ERR30;
goto FAILED;
}
/* If matching is caseless, upper and lower are converted to
alpha. This relies on the fact that the class table starts with
alpha, lower, upper as the first 3 entries. */
if ((options & PCRE_CASELESS) != 0 && posix_class <= 2)
posix_class = 0;
/* We build the bit map for the POSIX class in a chunk of local store
because we may be adding and subtracting from it, and we don't want to
subtract bits that may be in the main map already. At the end we or the
result into the bit map that is being built. */
posix_class *= 3;
/* Copy in the first table (always present) */
memcpy(pbits, cbits + posix_class_maps[posix_class],
32 * sizeof(uschar));
/* If there is a second table, add or remove it as required. */
taboffset = posix_class_maps[posix_class + 1];
tabopt = posix_class_maps[posix_class + 2];
if (taboffset >= 0)
{
if (tabopt >= 0)
for (c = 0; c < 32; c++) pbits[c] |= cbits[c + taboffset];
else
for (c = 0; c < 32; c++) pbits[c] &= ~cbits[c + taboffset];
}
/* Not see if we need to remove any special characters. An option
value of 1 removes vertical space and 2 removes underscore. */
if (tabopt < 0) tabopt = -tabopt;
if (tabopt == 1) pbits[1] &= ~0x3c;
else if (tabopt == 2) pbits[11] &= 0x7f;
/* Add the POSIX table or its complement into the main table that is
being built and we are done. */
if (local_negate)
for (c = 0; c < 32; c++) classbits[c] |= ~pbits[c];
else
for (c = 0; c < 32; c++) classbits[c] |= pbits[c];
ptr = tempptr + 1;
class_charcount = 10; /* Set > 1; assumes more than 1 per class */
continue; /* End of POSIX syntax handling */
}
/* Backslash may introduce a single character, or it may introduce one
of the specials, which just set a flag. Escaped items are checked for
validity in the pre-compiling pass. The sequence \b is a special case.
Inside a class (and only there) it is treated as backspace. Elsewhere
it marks a word boundary. Other escapes have preset maps ready to
or into the one we are building. We assume they have more than one
character in them, so set class_charcount bigger than one. */
if (c == '\\')
{
c = check_escape(&ptr, errorcodeptr, *brackets, options, TRUE);
if (-c == ESC_b) c = '\b'; /* \b is backslash in a class */
else if (-c == ESC_X) c = 'X'; /* \X is literal X in a class */
else if (-c == ESC_Q) /* Handle start of quoted string */
{
if (ptr[1] == '\\' && ptr[2] == 'E')
{
ptr += 2; /* avoid empty string */
}
else inescq = TRUE;
continue;
}
if (c < 0)
{
register const uschar *cbits = cd->cbits;
class_charcount += 2; /* Greater than 1 is what matters */
switch (-c)
{
case ESC_d:
for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_digit];
continue;
case ESC_D:
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_digit];
continue;
case ESC_w:
for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_word];
continue;
case ESC_W:
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_word];
continue;
case ESC_s:
for (c = 0; c < 32; c++) classbits[c] |= cbits[c+cbit_space];
classbits[1] &= ~0x08; /* Perl 5.004 onwards omits VT from \s */
continue;
case ESC_S:
for (c = 0; c < 32; c++) classbits[c] |= ~cbits[c+cbit_space];
classbits[1] |= 0x08; /* Perl 5.004 onwards omits VT from \s */
continue;
#ifdef SUPPORT_UCP
case ESC_p:
case ESC_P:
{
BOOL negated;
int pdata;
int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
if (ptype < 0) goto FAILED;
class_utf8 = TRUE;
*class_utf8data++ = ((-c == ESC_p) != negated)?
XCL_PROP : XCL_NOTPROP;
*class_utf8data++ = ptype;
*class_utf8data++ = pdata;
class_charcount -= 2; /* Not a < 256 character */
}
continue;
#endif
/* Unrecognized escapes are faulted if PCRE is running in its
strict mode. By default, for compatibility with Perl, they are
treated as literals. */
default:
if ((options & PCRE_EXTRA) != 0)
{
*errorcodeptr = ERR7;
goto FAILED;
}
c = *ptr; /* The final character */
class_charcount -= 2; /* Undo the default count from above */
}
}
/* Fall through if we have a single character (c >= 0). This may be
> 256 in UTF-8 mode. */
} /* End of backslash handling */
/* A single character may be followed by '-' to form a range. However,
Perl does not permit ']' to be the end of the range. A '-' character
here is treated as a literal. */
if (ptr[1] == '-' && ptr[2] != ']')
{
int d;
ptr += 2;
#ifdef SUPPORT_UTF8
if (utf8)
{ /* Braces are required because the */
GETCHARLEN(d, ptr, ptr); /* macro generates multiple statements */
}
else
#endif
d = *ptr; /* Not UTF-8 mode */
/* The second part of a range can be a single-character escape, but
not any of the other escapes. Perl 5.6 treats a hyphen as a literal
in such circumstances. */
if (d == '\\')
{
const uschar *oldptr = ptr;
d = check_escape(&ptr, errorcodeptr, *brackets, options, TRUE);
/* \b is backslash; \X is literal X; any other special means the '-'
was literal */
if (d < 0)
{
if (d == -ESC_b) d = '\b';
else if (d == -ESC_X) d = 'X'; else
{
ptr = oldptr - 2;
goto LONE_SINGLE_CHARACTER; /* A few lines below */
}
}
}
/* The check that the two values are in the correct order happens in
the pre-pass. Optimize one-character ranges */
if (d == c) goto LONE_SINGLE_CHARACTER; /* A few lines below */
/* In UTF-8 mode, if the upper limit is > 255, or > 127 for caseless
matching, we have to use an XCLASS with extra data items. Caseless
matching for characters > 127 is available only if UCP support is
available. */
#ifdef SUPPORT_UTF8
if (utf8 && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))
{
class_utf8 = TRUE;
/* With UCP support, we can find the other case equivalents of
the relevant characters. There may be several ranges. Optimize how
they fit with the basic range. */
#ifdef SUPPORT_UCP
if ((options & PCRE_CASELESS) != 0)
{
int occ, ocd;
int cc = c;
int origd = d;
while (get_othercase_range(&cc, origd, &occ, &ocd))
{
if (occ >= c && ocd <= d) continue; /* Skip embedded ranges */
if (occ < c && ocd >= c - 1) /* Extend the basic range */
{ /* if there is overlap, */
c = occ; /* noting that if occ < c */
continue; /* we can't have ocd > d */
} /* because a subrange is */
if (ocd > d && occ <= d + 1) /* always shorter than */
{ /* the basic range. */
d = ocd;
continue;
}
if (occ == ocd)
{
*class_utf8data++ = XCL_SINGLE;
}
else
{
*class_utf8data++ = XCL_RANGE;
class_utf8data += _pcre_ord2utf8(occ, class_utf8data);
}
class_utf8data += _pcre_ord2utf8(ocd, class_utf8data);
}
}
#endif /* SUPPORT_UCP */
/* Now record the original range, possibly modified for UCP caseless
overlapping ranges. */
*class_utf8data++ = XCL_RANGE;
class_utf8data += _pcre_ord2utf8(c, class_utf8data);
class_utf8data += _pcre_ord2utf8(d, class_utf8data);
/* With UCP support, we are done. Without UCP support, there is no
caseless matching for UTF-8 characters > 127; we can use the bit map
for the smaller ones. */
#ifdef SUPPORT_UCP
continue; /* With next character in the class */
#else
if ((options & PCRE_CASELESS) == 0 || c > 127) continue;
/* Adjust upper limit and fall through to set up the map */
d = 127;
#endif /* SUPPORT_UCP */
}
#endif /* SUPPORT_UTF8 */
/* We use the bit map for all cases when not in UTF-8 mode; else
ranges that lie entirely within 0-127 when there is UCP support; else
for partial ranges without UCP support. */
for (; c <= d; c++)
{
classbits[c/8] |= (1 << (c&7));
if ((options & PCRE_CASELESS) != 0)
{
int uc = cd->fcc[c]; /* flip case */
classbits[uc/8] |= (1 << (uc&7));
}
class_charcount++; /* in case a one-char range */
class_lastchar = c;
}
continue; /* Go get the next char in the class */
}
/* Handle a lone single character - we can get here for a normal
non-escape char, or after \ that introduces a single character or for an
apparent range that isn't. */
LONE_SINGLE_CHARACTER:
/* Handle a character that cannot go in the bit map */
#ifdef SUPPORT_UTF8
if (utf8 && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))
{
class_utf8 = TRUE;
*class_utf8data++ = XCL_SINGLE;
class_utf8data += _pcre_ord2utf8(c, class_utf8data);
#ifdef SUPPORT_UCP
if ((options & PCRE_CASELESS) != 0)
{
int othercase;
if ((othercase = _pcre_ucp_othercase(c)) >= 0)
{
*class_utf8data++ = XCL_SINGLE;
class_utf8data += _pcre_ord2utf8(othercase, class_utf8data);
}
}
#endif /* SUPPORT_UCP */
}
else
#endif /* SUPPORT_UTF8 */
/* Handle a single-byte character */
{
classbits[c/8] |= (1 << (c&7));
if ((options & PCRE_CASELESS) != 0)
{
c = cd->fcc[c]; /* flip case */
classbits[c/8] |= (1 << (c&7));
}
class_charcount++;
class_lastchar = c;
}
}
/* Loop until ']' reached; the check for end of string happens inside the
loop. This "while" is the end of the "do" above. */
while ((c = *(++ptr)) != ']' || inescq);
/* If class_charcount is 1, we saw precisely one character whose value is
less than 256. In non-UTF-8 mode we can always optimize. In UTF-8 mode, we
can optimize the negative case only if there were no characters >= 128
because OP_NOT and the related opcodes like OP_NOTSTAR operate on
single-bytes only. This is an historical hangover. Maybe one day we can
tidy these opcodes to handle multi-byte characters.
The optimization throws away the bit map. We turn the item into a
1-character OP_CHAR[NC] if it's positive, or OP_NOT if it's negative. Note
that OP_NOT does not support multibyte characters. In the positive case, it
can cause firstbyte to be set. Otherwise, there can be no first char if
this item is first, whatever repeat count may follow. In the case of
reqbyte, save the previous value for reinstating. */
#ifdef SUPPORT_UTF8
if (class_charcount == 1 &&
(!utf8 ||
(!class_utf8 && (!negate_class || class_lastchar < 128))))
#else
if (class_charcount == 1)
#endif
{
zeroreqbyte = reqbyte;
/* The OP_NOT opcode works on one-byte characters only. */
if (negate_class)
{
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
zerofirstbyte = firstbyte;
*code++ = OP_NOT;
*code++ = class_lastchar;
break;
}
/* For a single, positive character, get the value into mcbuffer, and
then we can handle this with the normal one-character code. */
#ifdef SUPPORT_UTF8
if (utf8 && class_lastchar > 127)
mclength = _pcre_ord2utf8(class_lastchar, mcbuffer);
else
#endif
{
mcbuffer[0] = class_lastchar;
mclength = 1;
}
goto ONE_CHAR;
} /* End of 1-char optimization */
/* The general case - not the one-char optimization. If this is the first
thing in the branch, there can be no first char setting, whatever the
repeat count. Any reqbyte setting must remain unchanged after any kind of
repeat. */
if (firstbyte == REQ_UNSET) firstbyte = REQ_NONE;
zerofirstbyte = firstbyte;
zeroreqbyte = reqbyte;
/* If there are characters with values > 255, we have to compile an
extended class, with its own opcode. If there are no characters < 256,
we can omit the bitmap. */
#ifdef SUPPORT_UTF8
if (class_utf8)
{
*class_utf8data++ = XCL_END; /* Marks the end of extra data */
*code++ = OP_XCLASS;
code += LINK_SIZE;
*code = negate_class? XCL_NOT : 0;
/* If the map is required, install it, and move on to the end of
the extra data */
if (class_charcount > 0)
{
*code++ |= XCL_MAP;
memcpy(code, classbits, 32);
code = class_utf8data;
}
/* If the map is not required, slide down the extra data. */
else
{
int len = class_utf8data - (code + 33);
memmove(code + 1, code + 33, len);
code += len + 1;
}
/* Now fill in the complete length of the item */
PUT(previous, 1, code - previous);
break; /* End of class handling */
}
#endif
/* If there are no characters > 255, negate the 32-byte map if necessary,
and copy it into the code vector. If this is the first thing in the branch,
there can be no first char setting, whatever the repeat count. Any reqbyte
setting must remain unchanged after any kind of repeat. */
if (negate_class)
{
*code++ = OP_NCLASS;
for (c = 0; c < 32; c++) code[c] = ~classbits[c];
}
else
{
*code++ = OP_CLASS;
memcpy(code, classbits, 32);
}
code += 32;
break;
/* Various kinds of repeat; '{' is not necessarily a quantifier, but this
has been tested above. */
case '{':
if (!is_quantifier) goto NORMAL_CHAR;
ptr = read_repeat_counts(ptr+1, &repeat_min, &repeat_max, errorcodeptr);
if (*errorcodeptr != 0) goto FAILED;
goto REPEAT;
case '*':
repeat_min = 0;
repeat_max = -1;
goto REPEAT;
case '+':
repeat_min = 1;
repeat_max = -1;
goto REPEAT;
case '?':
repeat_min = 0;
repeat_max = 1;
REPEAT:
if (previous == NULL)
{
*errorcodeptr = ERR9;
goto FAILED;
}
if (repeat_min == 0)
{
firstbyte = zerofirstbyte; /* Adjust for zero repeat */
reqbyte = zeroreqbyte; /* Ditto */
}
/* Remember whether this is a variable length repeat */
reqvary = (repeat_min == repeat_max)? 0 : REQ_VARY;
op_type = 0; /* Default single-char op codes */
possessive_quantifier = FALSE; /* Default not possessive quantifier */
/* Save start of previous item, in case we have to move it up to make space
for an inserted OP_ONCE for the additional '+' extension. */
tempcode = previous;
/* If the next character is '+', we have a possessive quantifier. This
implies greediness, whatever the setting of the PCRE_UNGREEDY option.
If the next character is '?' this is a minimizing repeat, by default,
but if PCRE_UNGREEDY is set, it works the other way round. We change the
repeat type to the non-default. */
if (ptr[1] == '+')
{
repeat_type = 0; /* Force greedy */
possessive_quantifier = TRUE;
ptr++;
}
else if (ptr[1] == '?')
{
repeat_type = greedy_non_default;
ptr++;
}
else repeat_type = greedy_default;
/* If previous was a recursion, we need to wrap it inside brackets so that
it can be replicated if necessary. */
if (*previous == OP_RECURSE)
{
memmove(previous + 1 + LINK_SIZE, previous, 1 + LINK_SIZE);
code += 1 + LINK_SIZE;
*previous = OP_BRA;
PUT(previous, 1, code - previous);
*code = OP_KET;
PUT(code, 1, code - previous);
code += 1 + LINK_SIZE;
}
/* If previous was a character match, abolish the item and generate a
repeat item instead. If a char item has a minumum of more than one, ensure
that it is set in reqbyte - it might not be if a sequence such as x{3} is
the first thing in a branch because the x will have gone into firstbyte
instead. */
if (*previous == OP_CHAR || *previous == OP_CHARNC)
{
/* Deal with UTF-8 characters that take up more than one byte. It's
easier to write this out separately than try to macrify it. Use c to
hold the length of the character in bytes, plus 0x80 to flag that it's a
length rather than a small character. */
#ifdef SUPPORT_UTF8
if (utf8 && (code[-1] & 0x80) != 0)
{
uschar *lastchar = code - 1;
while((*lastchar & 0xc0) == 0x80) lastchar--;
c = code - lastchar; /* Length of UTF-8 character */
memcpy(utf8_char, lastchar, c); /* Save the char */
c |= 0x80; /* Flag c as a length */
}
else
#endif
/* Handle the case of a single byte - either with no UTF8 support, or
with UTF-8 disabled, or for a UTF-8 character < 128. */
{
c = code[-1];
if (repeat_min > 1) reqbyte = c | req_caseopt | cd->req_varyopt;
}
goto OUTPUT_SINGLE_REPEAT; /* Code shared with single character types */
}
/* If previous was a single negated character ([^a] or similar), we use
one of the special opcodes, replacing it. The code is shared with single-
character repeats by setting opt_type to add a suitable offset into
repeat_type. OP_NOT is currently used only for single-byte chars. */
else if (*previous == OP_NOT)
{
op_type = OP_NOTSTAR - OP_STAR; /* Use "not" opcodes */
c = previous[1];
goto OUTPUT_SINGLE_REPEAT;
}
/* If previous was a character type match (\d or similar), abolish it and
create a suitable repeat item. The code is shared with single-character
repeats by setting op_type to add a suitable offset into repeat_type. Note
the the Unicode property types will be present only when SUPPORT_UCP is
defined, but we don't wrap the little bits of code here because it just
makes it horribly messy. */
else if (*previous < OP_EODN)
{
uschar *oldcode;
int prop_type, prop_value;
op_type = OP_TYPESTAR - OP_STAR; /* Use type opcodes */
c = *previous;
OUTPUT_SINGLE_REPEAT:
if (*previous == OP_PROP || *previous == OP_NOTPROP)
{
prop_type = previous[1];
prop_value = previous[2];
}
else prop_type = prop_value = -1;
oldcode = code;
code = previous; /* Usually overwrite previous item */
/* If the maximum is zero then the minimum must also be zero; Perl allows
this case, so we do too - by simply omitting the item altogether. */
if (repeat_max == 0) goto END_REPEAT;
/* All real repeats make it impossible to handle partial matching (maybe
one day we will be able to remove this restriction). */
if (repeat_max != 1) cd->nopartial = TRUE;
/* Combine the op_type with the repeat_type */
repeat_type += op_type;
/* A minimum of zero is handled either as the special case * or ?, or as
an UPTO, with the maximum given. */
if (repeat_min == 0)
{
if (repeat_max == -1) *code++ = OP_STAR + repeat_type;
else if (repeat_max == 1) *code++ = OP_QUERY + repeat_type;
else
{
*code++ = OP_UPTO + repeat_type;
PUT2INC(code, 0, repeat_max);
}
}
/* A repeat minimum of 1 is optimized into some special cases. If the
maximum is unlimited, we use OP_PLUS. Otherwise, the original item it
left in place and, if the maximum is greater than 1, we use OP_UPTO with
one less than the maximum. */
else if (repeat_min == 1)
{
if (repeat_max == -1)
*code++ = OP_PLUS + repeat_type;
else
{
code = oldcode; /* leave previous item in place */
if (repeat_max == 1) goto END_REPEAT;
*code++ = OP_UPTO + repeat_type;
PUT2INC(code, 0, repeat_max - 1);
}
}
/* The case {n,n} is just an EXACT, while the general case {n,m} is
handled as an EXACT followed by an UPTO. */
else
{
*code++ = OP_EXACT + op_type; /* NB EXACT doesn't have repeat_type */
PUT2INC(code, 0, repeat_min);
/* If the maximum is unlimited, insert an OP_STAR. Before doing so,
we have to insert the character for the previous code. For a repeated
Unicode property match, there are two extra bytes that define the
required property. In UTF-8 mode, long characters have their length in
c, with the 0x80 bit as a flag. */
if (repeat_max < 0)
{
#ifdef SUPPORT_UTF8
if (utf8 && c >= 128)
{
memcpy(code, utf8_char, c & 7);
code += c & 7;
}
else
#endif
{
*code++ = c;
if (prop_type >= 0)
{
*code++ = prop_type;
*code++ = prop_value;
}
}
*code++ = OP_STAR + repeat_type;
}
/* Else insert an UPTO if the max is greater than the min, again
preceded by the character, for the previously inserted code. */
else if (repeat_max != repeat_min)
{
#ifdef SUPPORT_UTF8
if (utf8 && c >= 128)
{
memcpy(code, utf8_char, c & 7);
code += c & 7;
}
else
#endif
*code++ = c;
if (prop_type >= 0)
{
*code++ = prop_type;
*code++ = prop_value;
}
repeat_max -= repeat_min;
*code++ = OP_UPTO + repeat_type;
PUT2INC(code, 0, repeat_max);
}
}
/* The character or character type itself comes last in all cases. */
#ifdef SUPPORT_UTF8
if (utf8 && c >= 128)
{
memcpy(code, utf8_char, c & 7);
code += c & 7;
}
else
#endif
*code++ = c;
/* For a repeated Unicode property match, there are two extra bytes that
define the required property. */
#ifdef SUPPORT_UCP
if (prop_type >= 0)
{
*code++ = prop_type;
*code++ = prop_value;
}
#endif
}
/* If previous was a character class or a back reference, we put the repeat
stuff after it, but just skip the item if the repeat was {0,0}. */
else if (*previous == OP_CLASS ||
*previous == OP_NCLASS ||
#ifdef SUPPORT_UTF8
*previous == OP_XCLASS ||
#endif
*previous == OP_REF)
{
if (repeat_max == 0)
{
code = previous;
goto END_REPEAT;
}
/* All real repeats make it impossible to handle partial matching (maybe
one day we will be able to remove this restriction). */
if (repeat_max != 1) cd->nopartial = TRUE;
if (repeat_min == 0 && repeat_max == -1)
*code++ = OP_CRSTAR + repeat_type;
else if (repeat_min == 1 && repeat_max == -1)
*code++ = OP_CRPLUS + repeat_type;
else if (repeat_min == 0 && repeat_max == 1)
*code++ = OP_CRQUERY + repeat_type;
else
{
*code++ = OP_CRRANGE + repeat_type;
PUT2INC(code, 0, repeat_min);
if (repeat_max == -1) repeat_max = 0; /* 2-byte encoding for max */
PUT2INC(code, 0, repeat_max);
}
}
/* If previous was a bracket group, we may have to replicate it in certain
cases. */
else if (*previous >= OP_BRA || *previous == OP_ONCE ||
*previous == OP_COND)
{
register int i;
int ketoffset = 0;
int len = code - previous;
uschar *bralink = NULL;
/* If the maximum repeat count is unlimited, find the end of the bracket
by scanning through from the start, and compute the offset back to it
from the current code pointer. There may be an OP_OPT setting following
the final KET, so we can't find the end just by going back from the code
pointer. */
if (repeat_max == -1)
{
register uschar *ket = previous;
do ket += GET(ket, 1); while (*ket != OP_KET);
ketoffset = code - ket;
}
/* The case of a zero minimum is special because of the need to stick
OP_BRAZERO in front of it, and because the group appears once in the
data, whereas in other cases it appears the minimum number of times. For
this reason, it is simplest to treat this case separately, as otherwise
the code gets far too messy. There are several special subcases when the
minimum is zero. */
if (repeat_min == 0)
{
/* If the maximum is also zero, we just omit the group from the output
altogether. */
if (repeat_max == 0)
{
code = previous;
goto END_REPEAT;
}
/* If the maximum is 1 or unlimited, we just have to stick in the
BRAZERO and do no more at this point. However, we do need to adjust
any OP_RECURSE calls inside the group that refer to the group itself or
any internal group, because the offset is from the start of the whole
regex. Temporarily terminate the pattern while doing this. */
if (repeat_max <= 1)
{
*code = OP_END;
adjust_recurse(previous, 1, utf8, cd);
memmove(previous+1, previous, len);
code++;
*previous++ = OP_BRAZERO + repeat_type;
}
/* If the maximum is greater than 1 and limited, we have to replicate
in a nested fashion, sticking OP_BRAZERO before each set of brackets.
The first one has to be handled carefully because it's the original
copy, which has to be moved up. The remainder can be handled by code
that is common with the non-zero minimum case below. We have to
adjust the value or repeat_max, since one less copy is required. Once
again, we may have to adjust any OP_RECURSE calls inside the group. */
else
{
int offset;
*code = OP_END;
adjust_recurse(previous, 2 + LINK_SIZE, utf8, cd);
memmove(previous + 2 + LINK_SIZE, previous, len);
code += 2 + LINK_SIZE;
*previous++ = OP_BRAZERO + repeat_type;
*previous++ = OP_BRA;
/* We chain together the bracket offset fields that have to be
filled in later when the ends of the brackets are reached. */
offset = (bralink == NULL)? 0 : previous - bralink;
bralink = previous;
PUTINC(previous, 0, offset);
}
repeat_max--;
}
/* If the minimum is greater than zero, replicate the group as many
times as necessary, and adjust the maximum to the number of subsequent
copies that we need. If we set a first char from the group, and didn't
set a required char, copy the latter from the former. */
else
{
if (repeat_min > 1)
{
if (groupsetfirstbyte && reqbyte < 0) reqbyte = firstbyte;
for (i = 1; i < repeat_min; i++)
{
memcpy(code, previous, len);
code += len;
}
}
if (repeat_max > 0) repeat_max -= repeat_min;
}
/* This code is common to both the zero and non-zero minimum cases. If
the maximum is limited, it replicates the group in a nested fashion,
remembering the bracket starts on a stack. In the case of a zero minimum,
the first one was set up above. In all cases the repeat_max now specifies
the number of additional copies needed. */
if (repeat_max >= 0)
{
for (i = repeat_max - 1; i >= 0; i--)
{
*code++ = OP_BRAZERO + repeat_type;
/* All but the final copy start a new nesting, maintaining the
chain of brackets outstanding. */
if (i != 0)
{
int offset;
*code++ = OP_BRA;
offset = (bralink == NULL)? 0 : code - bralink;
bralink = code;
PUTINC(code, 0, offset);
}
memcpy(code, previous, len);
code += len;
}
/* Now chain through the pending brackets, and fill in their length
fields (which are holding the chain links pro tem). */
while (bralink != NULL)
{
int oldlinkoffset;
int offset = code - bralink + 1;
uschar *bra = code - offset;
oldlinkoffset = GET(bra, 1);
bralink = (oldlinkoffset == 0)? NULL : bralink - oldlinkoffset;
*code++ = OP_KET;
PUTINC(code, 0, offset);
PUT(bra, 1, offset);
}
}
/* If the maximum is unlimited, set a repeater in the final copy. We
can't just offset backwards from the current code point, because we
don't know if there's been an options resetting after the ket. The
correct offset was computed above. */
else code[-ketoffset] = OP_KETRMAX + repeat_type;
}
/* Else there's some kind of shambles */
else
{
*errorcodeptr = ERR11;
goto FAILED;
}
/* If the character following a repeat is '+', we wrap the entire repeated
item inside OP_ONCE brackets. This is just syntactic sugar, taken from
Sun's Java package. The repeated item starts at tempcode, not at previous,
which might be the first part of a string whose (former) last char we
repeated. However, we don't support '+' after a greediness '?'. */
if (possessive_quantifier)
{
int len = code - tempcode;
memmove(tempcode + 1+LINK_SIZE, tempcode, len);
code += 1 + LINK_SIZE;
len += 1 + LINK_SIZE;
tempcode[0] = OP_ONCE;
*code++ = OP_KET;
PUTINC(code, 0, len);
PUT(tempcode, 1, len);
}
/* In all case we no longer have a previous item. We also set the
"follows varying string" flag for subsequently encountered reqbytes if
it isn't already set and we have just passed a varying length item. */
END_REPEAT:
previous = NULL;
cd->req_varyopt |= reqvary;
break;
/* Start of nested bracket sub-expression, or comment or lookahead or
lookbehind or option setting or condition. First deal with special things
that can come after a bracket; all are introduced by ?, and the appearance
of any of them means that this is not a referencing group. They were
checked for validity in the first pass over the string, so we don't have to
check for syntax errors here. */
case '(':
newoptions = options;
skipbytes = 0;
if (*(++ptr) == '?')
{
int set, unset;
int *optset;
switch (*(++ptr))
{
case '#': /* Comment; skip to ket */
ptr++;
while (*ptr != ')') ptr++;
continue;
case ':': /* Non-extracting bracket */
bravalue = OP_BRA;
ptr++;
break;
case '(':
bravalue = OP_COND; /* Conditional group */
/* A condition can be a number, referring to a numbered group, a name,
referring to a named group, 'R', referring to recursion, or an
assertion. There are two unfortunate ambiguities, caused by history.
(a) 'R' can be the recursive thing or the name 'R', and (b) a number
could be a name that consists of digits. In both cases, we look for a
name first; if not found, we try the other cases. If the first
character after (?( is a word character, we know the rest up to ) will
also be word characters because the syntax was checked in the first
pass. */
if ((cd->ctypes[ptr[1]] & ctype_word) != 0)
{
int i, namelen;
int condref = 0;
const uschar *name;
uschar *slot = cd->name_table;
/* This is needed for all successful cases. */
skipbytes = 3;
/* Read the name, but also get it as a number if it's all digits */
name = ++ptr;
while (*ptr != ')')
{
if (condref >= 0)
condref = ((digitab[*ptr] & ctype_digit) != 0)?
condref * 10 + *ptr - '0' : -1;
ptr++;
}
namelen = ptr - name;
ptr++;
for (i = 0; i < cd->names_found; i++)
{
if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break;
slot += cd->name_entry_size;
}
/* Found a previous named subpattern */
if (i < cd->names_found)
{
condref = GET2(slot, 0);
code[1+LINK_SIZE] = OP_CREF;
PUT2(code, 2+LINK_SIZE, condref);
}
/* Search the pattern for a forward reference */
else if ((i = find_named_parens(ptr, *brackets, name, namelen)) > 0)
{
code[1+LINK_SIZE] = OP_CREF;
PUT2(code, 2+LINK_SIZE, i);
}
/* Check for 'R' for recursion */
else if (namelen == 1 && *name == 'R')
{
code[1+LINK_SIZE] = OP_CREF;
PUT2(code, 2+LINK_SIZE, CREF_RECURSE);
}
/* Check for a subpattern number */
else if (condref > 0)
{
code[1+LINK_SIZE] = OP_CREF;
PUT2(code, 2+LINK_SIZE, condref);
}
/* Either an unidentified subpattern, or a reference to (?(0) */
else
{
*errorcodeptr = (condref == 0)? ERR35: ERR15;
goto FAILED;
}
}
/* For conditions that are assertions, we just fall through, having
set bravalue above. */
break;
case '=': /* Positive lookahead */
bravalue = OP_ASSERT;
ptr++;
break;
case '!': /* Negative lookahead */
bravalue = OP_ASSERT_NOT;
ptr++;
break;
case '<': /* Lookbehinds */
switch (*(++ptr))
{
case '=': /* Positive lookbehind */
bravalue = OP_ASSERTBACK;
ptr++;
break;
case '!': /* Negative lookbehind */
bravalue = OP_ASSERTBACK_NOT;
ptr++;
break;
}
break;
case '>': /* One-time brackets */
bravalue = OP_ONCE;
ptr++;
break;
case 'C': /* Callout - may be followed by digits; */
previous_callout = code; /* Save for later completion */
after_manual_callout = 1; /* Skip one item before completing */
*code++ = OP_CALLOUT; /* Already checked that the terminating */
{ /* closing parenthesis is present. */
int n = 0;
while ((digitab[*(++ptr)] & ctype_digit) != 0)
n = n * 10 + *ptr - '0';
if (n > 255)
{
*errorcodeptr = ERR38;
goto FAILED;
}
*code++ = n;
PUT(code, 0, ptr - cd->start_pattern + 1); /* Pattern offset */
PUT(code, LINK_SIZE, 0); /* Default length */
code += 2 * LINK_SIZE;
}
previous = NULL;
continue;
case 'P': /* Named subpattern handling */
if (*(++ptr) == '<') /* Definition */
{
int i, namelen;
uschar *slot = cd->name_table;
const uschar *name; /* Don't amalgamate; some compilers */
name = ++ptr; /* grumble at autoincrement in declaration */
while (*ptr++ != '>');
namelen = ptr - name - 1;
for (i = 0; i < cd->names_found; i++)
{
int crc = memcmp(name, slot+2, namelen);
if (crc == 0)
{
if (slot[2+namelen] == 0)
{
if ((options & PCRE_DUPNAMES) == 0)
{
*errorcodeptr = ERR43;
goto FAILED;
}
}
else crc = -1; /* Current name is substring */
}
if (crc < 0)
{
memmove(slot + cd->name_entry_size, slot,
(cd->names_found - i) * cd->name_entry_size);
break;
}
slot += cd->name_entry_size;
}
PUT2(slot, 0, *brackets + 1);
memcpy(slot + 2, name, namelen);
slot[2+namelen] = 0;
cd->names_found++;
goto NUMBERED_GROUP;
}
if (*ptr == '=' || *ptr == '>') /* Reference or recursion */
{
int i, namelen;
int type = *ptr++;
const uschar *name = ptr;
uschar *slot = cd->name_table;
while (*ptr != ')') ptr++;
namelen = ptr - name;
for (i = 0; i < cd->names_found; i++)
{
if (strncmp((char *)name, (char *)slot+2, namelen) == 0) break;
slot += cd->name_entry_size;
}
if (i < cd->names_found) /* Back reference */
{
recno = GET2(slot, 0);
}
else if ((recno = /* Forward back reference */
find_named_parens(ptr, *brackets, name, namelen)) <= 0)
{
*errorcodeptr = ERR15;
goto FAILED;
}
if (type == '>') goto HANDLE_RECURSION; /* A few lines below */
/* Back reference */
previous = code;
*code++ = OP_REF;
PUT2INC(code, 0, recno);
cd->backref_map |= (recno < 32)? (1 << recno) : 1;
if (recno > cd->top_backref) cd->top_backref = recno;
continue;
}
/* Should never happen */
break;
case 'R': /* Pattern recursion */
ptr++; /* Same as (?0) */
/* Fall through */
/* Recursion or "subroutine" call */
case '0': case '1': case '2': case '3': case '4':
case '5': case '6': case '7': case '8': case '9':
{
const uschar *called;
recno = 0;
while((digitab[*ptr] & ctype_digit) != 0)
recno = recno * 10 + *ptr++ - '0';
/* Come here from code above that handles a named recursion */
HANDLE_RECURSION:
previous = code;
/* Find the bracket that is being referenced. Temporarily end the
regex in case it doesn't exist. */
*code = OP_END;
called = (recno == 0)? cd->start_code :
find_bracket(cd->start_code, utf8, recno);
if (called == NULL)
{
*errorcodeptr = ERR15;
goto FAILED;
}
/* If the subpattern is still open, this is a recursive call. We
check to see if this is a left recursion that could loop for ever,
and diagnose that case. */
if (GET(called, 1) == 0 && could_be_empty(called, code, bcptr, utf8))
{
*errorcodeptr = ERR40;
goto FAILED;
}
/* Insert the recursion/subroutine item, automatically wrapped inside
"once" brackets. */
*code = OP_ONCE;
PUT(code, 1, 2 + 2*LINK_SIZE);
code += 1 + LINK_SIZE;
*code = OP_RECURSE;
PUT(code, 1, called - cd->start_code);
code += 1 + LINK_SIZE;
*code = OP_KET;
PUT(code, 1, 2 + 2*LINK_SIZE);
code += 1 + LINK_SIZE;
}
continue;
/* Character after (? not specially recognized */
default: /* Option setting */
set = unset = 0;
optset = &set;
while (*ptr != ')' && *ptr != ':')
{
switch (*ptr++)
{
case '-': optset = &unset; break;
case 'i': *optset |= PCRE_CASELESS; break;
case 'J': *optset |= PCRE_DUPNAMES; break;
case 'm': *optset |= PCRE_MULTILINE; break;
case 's': *optset |= PCRE_DOTALL; break;
case 'x': *optset |= PCRE_EXTENDED; break;
case 'U': *optset |= PCRE_UNGREEDY; break;
case 'X': *optset |= PCRE_EXTRA; break;
}
}
/* Set up the changed option bits, but don't change anything yet. */
newoptions = (options | set) & (~unset);
/* If the options ended with ')' this is not the start of a nested
group with option changes, so the options change at this level. Compile
code to change the ims options if this setting actually changes any of
them. We also pass the new setting back so that it can be put at the
start of any following branches, and when this group ends (if we are in
a group), a resetting item can be compiled.
Note that if this item is right at the start of the pattern, the
options will have been abstracted and made global, so there will be no
change to compile. */
if (*ptr == ')')
{
if ((options & PCRE_IMS) != (newoptions & PCRE_IMS))
{
*code++ = OP_OPT;
*code++ = newoptions & PCRE_IMS;
}
/* Change options at this level, and pass them back for use
in subsequent branches. Reset the greedy defaults and the case
value for firstbyte and reqbyte. */
*optionsptr = options = newoptions;
greedy_default = ((newoptions & PCRE_UNGREEDY) != 0);
greedy_non_default = greedy_default ^ 1;
req_caseopt = ((options & PCRE_CASELESS) != 0)? REQ_CASELESS : 0;
previous = NULL; /* This item can't be repeated */
continue; /* It is complete */
}
/* If the options ended with ':' we are heading into a nested group
with possible change of options. Such groups are non-capturing and are
not assertions of any kind. All we need to do is skip over the ':';
the newoptions value is handled below. */
bravalue = OP_BRA;
ptr++;
}
}
/* If PCRE_NO_AUTO_CAPTURE is set, all unadorned brackets become
non-capturing and behave like (?:...) brackets */
else if ((options & PCRE_NO_AUTO_CAPTURE) != 0)
{
bravalue = OP_BRA;
}
/* Else we have a referencing group; adjust the opcode. If the bracket
number is greater than EXTRACT_BASIC_MAX, we set the opcode one higher, and
arrange for the true number to follow later, in an OP_BRANUMBER item. */
else
{
NUMBERED_GROUP:
if (++(*brackets) > EXTRACT_BASIC_MAX)
{
bravalue = OP_BRA + EXTRACT_BASIC_MAX + 1;
code[1+LINK_SIZE] = OP_BRANUMBER;
PUT2(code, 2+LINK_SIZE, *brackets);
skipbytes = 3;
}
else bravalue = OP_BRA + *brackets;
}
/* Process nested bracketed re. Assertions may not be repeated, but other
kinds can be. We copy code into a non-register variable in order to be able
to pass its address because some compilers complain otherwise. Pass in a
new setting for the ims options if they have changed. */
previous = (bravalue >= OP_ONCE)? code : NULL;
*code = bravalue;
tempcode = code;
tempreqvary = cd->req_varyopt; /* Save value before bracket */
if (!compile_regex(
newoptions, /* The complete new option state */
options & PCRE_IMS, /* The previous ims option state */
brackets, /* Extracting bracket count */
&tempcode, /* Where to put code (updated) */
&ptr, /* Input pointer (updated) */
errorcodeptr, /* Where to put an error message */
(bravalue == OP_ASSERTBACK ||
bravalue == OP_ASSERTBACK_NOT), /* TRUE if back assert */
skipbytes, /* Skip over OP_COND/OP_BRANUMBER */
&subfirstbyte, /* For possible first char */
&subreqbyte, /* For possible last char */
bcptr, /* Current branch chain */
cd)) /* Tables block */
goto FAILED;
/* At the end of compiling, code is still pointing to the start of the
group, while tempcode has been updated to point past the end of the group
and any option resetting that may follow it. The pattern pointer (ptr)
is on the bracket. */
/* If this is a conditional bracket, check that there are no more than
two branches in the group. */
else if (bravalue == OP_COND)
{
uschar *tc = code;
int condcount = 0;
do {
condcount++;
tc += GET(tc,1);
}
while (*tc != OP_KET);
if (condcount > 2)
{
*errorcodeptr = ERR27;
goto FAILED;
}
/* If there is just one branch, we must not make use of its firstbyte or
reqbyte, because this is equivalent to an empty second branch. */
if (condcount == 1) subfirstbyte = subreqbyte = REQ_NONE;
}
/* Handle updating of the required and first characters. Update for normal
brackets of all kinds, and conditions with two branches (see code above).
If the bracket is followed by a quantifier with zero repeat, we have to
back off. Hence the definition of zeroreqbyte and zerofirstbyte outside the
main loop so that they can be accessed for the back off. */
zeroreqbyte = reqbyte;
zerofirstbyte = firstbyte;
groupsetfirstbyte = FALSE;
if (bravalue >= OP_BRA || bravalue == OP_ONCE || bravalue == OP_COND)
{
/* If we have not yet set a firstbyte in this branch, take it from the
subpattern, remembering that it was set here so that a repeat of more
than one can replicate it as reqbyte if necessary. If the subpattern has
no firstbyte, set "none" for the whole branch. In both cases, a zero
repeat forces firstbyte to "none". */
if (firstbyte == REQ_UNSET)
{
if (subfirstbyte >= 0)
{
firstbyte = subfirstbyte;
groupsetfirstbyte = TRUE;
}
else firstbyte = REQ_NONE;
zerofirstbyte = REQ_NONE;
}
/* If firstbyte was previously set, convert the subpattern's firstbyte
into reqbyte if there wasn't one, using the vary flag that was in
existence beforehand. */
else if (subfirstbyte >= 0 && subreqbyte < 0)
subreqbyte = subfirstbyte | tempreqvary;
/* If the subpattern set a required byte (or set a first byte that isn't
really the first byte - see above), set it. */
if (subreqbyte >= 0) reqbyte = subreqbyte;
}
/* For a forward assertion, we take the reqbyte, if set. This can be
helpful if the pattern that follows the assertion doesn't set a different
char. For example, it's useful for /(?=abcde).+/. We can't set firstbyte
for an assertion, however because it leads to incorrect effect for patterns
such as /(?=a)a.+/ when the "real" "a" would then become a reqbyte instead
of a firstbyte. This is overcome by a scan at the end if there's no
firstbyte, looking for an asserted first char. */
else if (bravalue == OP_ASSERT && subreqbyte >= 0) reqbyte = subreqbyte;
/* Now update the main code pointer to the end of the group. */
code = tempcode;
/* Error if hit end of pattern */
if (*ptr != ')')
{
*errorcodeptr = ERR14;
goto FAILED;
}
break;
/* Check \ for being a real metacharacter; if not, fall through and handle
it as a data character at the start of a string. Escape items are checked
for validity in the pre-compiling pass. */
case '\\':
tempptr = ptr;
c = check_escape(&ptr, errorcodeptr, *brackets, options, FALSE);
/* Handle metacharacters introduced by \. For ones like \d, the ESC_ values
are arranged to be the negation of the corresponding OP_values. For the
back references, the values are ESC_REF plus the reference number. Only
back references and those types that consume a character may be repeated.
We can test for values between ESC_b and ESC_Z for the latter; this may
have to change if any new ones are ever created. */
if (c < 0)
{
if (-c == ESC_Q) /* Handle start of quoted string */
{
if (ptr[1] == '\\' && ptr[2] == 'E') ptr += 2; /* avoid empty string */
else inescq = TRUE;
continue;
}
/* For metasequences that actually match a character, we disable the
setting of a first character if it hasn't already been set. */
if (firstbyte == REQ_UNSET && -c > ESC_b && -c < ESC_Z)
firstbyte = REQ_NONE;
/* Set values to reset to if this is followed by a zero repeat. */
zerofirstbyte = firstbyte;
zeroreqbyte = reqbyte;
/* Back references are handled specially */
if (-c >= ESC_REF)
{
int number = -c - ESC_REF;
previous = code;
*code++ = OP_REF;
PUT2INC(code, 0, number);
}
/* So are Unicode property matches, if supported. We know that get_ucp
won't fail because it was tested in the pre-pass. */
#ifdef SUPPORT_UCP
else if (-c == ESC_P || -c == ESC_p)
{
BOOL negated;
int pdata;
int ptype = get_ucp(&ptr, &negated, &pdata, errorcodeptr);
previous = code;
*code++ = ((-c == ESC_p) != negated)? OP_PROP : OP_NOTPROP;
*code++ = ptype;
*code++ = pdata;
}
#endif
/* For the rest, we can obtain the OP value by negating the escape
value */
else
{
previous = (-c > ESC_b && -c < ESC_Z)? code : NULL;
*code++ = -c;
}
continue;
}
/* We have a data character whose value is in c. In UTF-8 mode it may have
a value > 127. We set its representation in the length/buffer, and then
handle it as a data character. */
#ifdef SUPPORT_UTF8
if (utf8 && c > 127)
mclength = _pcre_ord2utf8(c, mcbuffer);
else
#endif
{
mcbuffer[0] = c;
mclength = 1;
}
goto ONE_CHAR;
/* Handle a literal character. It is guaranteed not to be whitespace or #
when the extended flag is set. If we are in UTF-8 mode, it may be a
multi-byte literal character. */
default:
NORMAL_CHAR:
mclength = 1;
mcbuffer[0] = c;
#ifdef SUPPORT_UTF8
if (utf8 && (c & 0xc0) == 0xc0)
{
while ((ptr[1] & 0xc0) == 0x80)
mcbuffer[mclength++] = *(++ptr);
}
#endif
/* At this point we have the character's bytes in mcbuffer, and the length
in mclength. When not in UTF-8 mode, the length is always 1. */
ONE_CHAR:
previous = code;
*code++ = ((options & PCRE_CASELESS) != 0)? OP_CHARNC : OP_CHAR;
for (c = 0; c < mclength; c++) *code++ = mcbuffer[c];
/* Set the first and required bytes appropriately. If no previous first
byte, set it from this character, but revert to none on a zero repeat.
Otherwise, leave the firstbyte value alone, and don't change it on a zero
repeat. */
if (firstbyte == REQ_UNSET)
{
zerofirstbyte = REQ_NONE;
zeroreqbyte = reqbyte;
/* If the character is more than one byte long, we can set firstbyte
only if it is not to be matched caselessly. */
if (mclength == 1 || req_caseopt == 0)
{
firstbyte = mcbuffer[0] | req_caseopt;
if (mclength != 1) reqbyte = code[-1] | cd->req_varyopt;
}
else firstbyte = reqbyte = REQ_NONE;
}
/* firstbyte was previously set; we can set reqbyte only the length is
1 or the matching is caseful. */
else
{
zerofirstbyte = firstbyte;
zeroreqbyte = reqbyte;
if (mclength == 1 || req_caseopt == 0)
reqbyte = code[-1] | req_caseopt | cd->req_varyopt;
}
break; /* End of literal character handling */
}
} /* end of big loop */
/* Control never reaches here by falling through, only by a goto for all the
error states. Pass back the position in the pattern so that it can be displayed
to the user for diagnosing the error. */
FAILED:
*ptrptr = ptr;
return FALSE;
}
/*************************************************
* Compile sequence of alternatives *
*************************************************/
/* On entry, ptr is pointing past the bracket character, but on return
it points to the closing bracket, or vertical bar, or end of string.
The code variable is pointing at the byte into which the BRA operator has been
stored. If the ims options are changed at the start (for a (?ims: group) or
during any branch, we need to insert an OP_OPT item at the start of every
following branch to ensure they get set correctly at run time, and also pass
the new options into every subsequent branch compile.
Argument:
options option bits, including any changes for this subpattern
oldims previous settings of ims option bits
brackets -> int containing the number of extracting brackets used
codeptr -> the address of the current code pointer
ptrptr -> the address of the current pattern pointer
errorcodeptr -> pointer to error code variable
lookbehind TRUE if this is a lookbehind assertion
skipbytes skip this many bytes at start (for OP_COND, OP_BRANUMBER)
firstbyteptr place to put the first required character, or a negative number
reqbyteptr place to put the last required character, or a negative number
bcptr pointer to the chain of currently open branches
cd points to the data block with tables pointers etc.
Returns: TRUE on success
*/
static BOOL
compile_regex(int options, int oldims, int *brackets, uschar **codeptr,
const uschar **ptrptr, int *errorcodeptr, BOOL lookbehind, int skipbytes,
int *firstbyteptr, int *reqbyteptr, branch_chain *bcptr, compile_data *cd)
{
const uschar *ptr = *ptrptr;
uschar *code = *codeptr;
uschar *last_branch = code;
uschar *start_bracket = code;
uschar *reverse_count = NULL;
int firstbyte, reqbyte;
int branchfirstbyte, branchreqbyte;
branch_chain bc;
bc.outer = bcptr;
bc.current = code;
firstbyte = reqbyte = REQ_UNSET;
/* Offset is set zero to mark that this bracket is still open */
PUT(code, 1, 0);
code += 1 + LINK_SIZE + skipbytes;
/* Loop for each alternative branch */
for (;;)
{
/* Handle a change of ims options at the start of the branch */
if ((options & PCRE_IMS) != oldims)
{
*code++ = OP_OPT;
*code++ = options & PCRE_IMS;
}
/* Set up dummy OP_REVERSE if lookbehind assertion */
if (lookbehind)
{
*code++ = OP_REVERSE;
reverse_count = code;
PUTINC(code, 0, 0);
}
/* Now compile the branch */
if (!compile_branch(&options, brackets, &code, &ptr, errorcodeptr,
&branchfirstbyte, &branchreqbyte, &bc, cd))
{
*ptrptr = ptr;
return FALSE;
}
/* If this is the first branch, the firstbyte and reqbyte values for the
branch become the values for the regex. */
if (*last_branch != OP_ALT)
{
firstbyte = branchfirstbyte;
reqbyte = branchreqbyte;
}
/* If this is not the first branch, the first char and reqbyte have to
match the values from all the previous branches, except that if the previous
value for reqbyte didn't have REQ_VARY set, it can still match, and we set
REQ_VARY for the regex. */
else
{
/* If we previously had a firstbyte, but it doesn't match the new branch,
we have to abandon the firstbyte for the regex, but if there was previously
no reqbyte, it takes on the value of the old firstbyte. */
if (firstbyte >= 0 && firstbyte != branchfirstbyte)
{
if (reqbyte < 0) reqbyte = firstbyte;
firstbyte = REQ_NONE;
}
/* If we (now or from before) have no firstbyte, a firstbyte from the
branch becomes a reqbyte if there isn't a branch reqbyte. */
if (firstbyte < 0 && branchfirstbyte >= 0 && branchreqbyte < 0)
branchreqbyte = branchfirstbyte;
/* Now ensure that the reqbytes match */
if ((reqbyte & ~REQ_VARY) != (branchreqbyte & ~REQ_VARY))
reqbyte = REQ_NONE;
else reqbyte |= branchreqbyte; /* To "or" REQ_VARY */
}
/* If lookbehind, check that this branch matches a fixed-length string,
and put the length into the OP_REVERSE item. Temporarily mark the end of
the branch with OP_END. */
if (lookbehind)
{
int length;
*code = OP_END;
length = find_fixedlength(last_branch, options);
DPRINTF(("fixed length = %d\n", length));
if (length < 0)
{
*errorcodeptr = (length == -2)? ERR36 : ERR25;
*ptrptr = ptr;
return FALSE;
}
PUT(reverse_count, 0, length);
}
/* Reached end of expression, either ')' or end of pattern. Go back through
the alternative branches and reverse the chain of offsets, with the field in
the BRA item now becoming an offset to the first alternative. If there are
no alternatives, it points to the end of the group. The length in the
terminating ket is always the length of the whole bracketed item. If any of
the ims options were changed inside the group, compile a resetting op-code
following, except at the very end of the pattern. Return leaving the pointer
at the terminating char. */
if (*ptr != '|')
{
int length = code - last_branch;
do
{
int prev_length = GET(last_branch, 1);
PUT(last_branch, 1, length);
length = prev_length;
last_branch -= length;
}
while (length > 0);
/* Fill in the ket */
*code = OP_KET;
PUT(code, 1, code - start_bracket);
code += 1 + LINK_SIZE;
/* Resetting option if needed */
if ((options & PCRE_IMS) != oldims && *ptr == ')')
{
*code++ = OP_OPT;
*code++ = oldims;
}
/* Set values to pass back */
*codeptr = code;
*ptrptr = ptr;
*firstbyteptr = firstbyte;
*reqbyteptr = reqbyte;
return TRUE;
}
/* Another branch follows; insert an "or" node. Its length field points back
to the previous branch while the bracket remains open. At the end the chain
is reversed. It's done like this so that the start of the bracket has a
zero offset until it is closed, making it possible to detect recursion. */
*code = OP_ALT;
PUT(code, 1, code - last_branch);
bc.current = last_branch = code;
code += 1 + LINK_SIZE;
ptr++;
}
/* Control never reaches here */
}
/*************************************************
* Check for anchored expression *
*************************************************/
/* Try to find out if this is an anchored regular expression. Consider each
alternative branch. If they all start with OP_SOD or OP_CIRC, or with a bracket
all of whose alternatives start with OP_SOD or OP_CIRC (recurse ad lib), then
it's anchored. However, if this is a multiline pattern, then only OP_SOD
counts, since OP_CIRC can match in the middle.
We can also consider a regex to be anchored if OP_SOM starts all its branches.
This is the code for \G, which means "match at start of match position, taking
into account the match offset".
A branch is also implicitly anchored if it starts with .* and DOTALL is set,
because that will try the rest of the pattern at all possible matching points,
so there is no point trying again.... er ....
.... except when the .* appears inside capturing parentheses, and there is a
subsequent back reference to those parentheses. We haven't enough information
to catch that case precisely.
At first, the best we could do was to detect when .* was in capturing brackets
and the highest back reference was greater than or equal to that level.
However, by keeping a bitmap of the first 31 back references, we can catch some
of the more common cases more precisely.
Arguments:
code points to start of expression (the bracket)
options points to the options setting
bracket_map a bitmap of which brackets we are inside while testing; this
handles up to substring 31; after that we just have to take
the less precise approach
backref_map the back reference bitmap
Returns: TRUE or FALSE
*/
static BOOL
is_anchored(register const uschar *code, int *options, unsigned int bracket_map,
unsigned int backref_map)
{
do {
const uschar *scode =
first_significant_code(code + 1+LINK_SIZE, options, PCRE_MULTILINE, FALSE);
register int op = *scode;
/* Capturing brackets */
if (op > OP_BRA)
{
int new_map;
op -= OP_BRA;
if (op > EXTRACT_BASIC_MAX) op = GET2(scode, 2+LINK_SIZE);
new_map = bracket_map | ((op < 32)? (1 << op) : 1);
if (!is_anchored(scode, options, new_map, backref_map)) return FALSE;
}
/* Other brackets */
else if (op == OP_BRA || op == OP_ASSERT || op == OP_ONCE || op == OP_COND)
{
if (!is_anchored(scode, options, bracket_map, backref_map)) return FALSE;
}
/* .* is not anchored unless DOTALL is set and it isn't in brackets that
are or may be referenced. */
else if ((op == OP_TYPESTAR || op == OP_TYPEMINSTAR) &&
(*options & PCRE_DOTALL) != 0)
{
if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE;
}
/* Check for explicit anchoring */
else if (op != OP_SOD && op != OP_SOM &&
((*options & PCRE_MULTILINE) != 0 || op != OP_CIRC))
return FALSE;
code += GET(code, 1);
}
while (*code == OP_ALT); /* Loop for each alternative */
return TRUE;
}
/*************************************************
* Check for starting with ^ or .* *
*************************************************/
/* This is called to find out if every branch starts with ^ or .* so that
"first char" processing can be done to speed things up in multiline
matching and for non-DOTALL patterns that start with .* (which must start at
the beginning or after \n). As in the case of is_anchored() (see above), we
have to take account of back references to capturing brackets that contain .*
because in that case we can't make the assumption.
Arguments:
code points to start of expression (the bracket)
bracket_map a bitmap of which brackets we are inside while testing; this
handles up to substring 31; after that we just have to take
the less precise approach
backref_map the back reference bitmap
Returns: TRUE or FALSE
*/
static BOOL
is_startline(const uschar *code, unsigned int bracket_map,
unsigned int backref_map)
{
do {
const uschar *scode = first_significant_code(code + 1+LINK_SIZE, NULL, 0,
FALSE);
register int op = *scode;
/* Capturing brackets */
if (op > OP_BRA)
{
int new_map;
op -= OP_BRA;
if (op > EXTRACT_BASIC_MAX) op = GET2(scode, 2+LINK_SIZE);
new_map = bracket_map | ((op < 32)? (1 << op) : 1);
if (!is_startline(scode, new_map, backref_map)) return FALSE;
}
/* Other brackets */
else if (op == OP_BRA || op == OP_ASSERT || op == OP_ONCE || op == OP_COND)
{ if (!is_startline(scode, bracket_map, backref_map)) return FALSE; }
/* .* means "start at start or after \n" if it isn't in brackets that
may be referenced. */
else if (op == OP_TYPESTAR || op == OP_TYPEMINSTAR)
{
if (scode[1] != OP_ANY || (bracket_map & backref_map) != 0) return FALSE;
}
/* Check for explicit circumflex */
else if (op != OP_CIRC) return FALSE;
/* Move on to the next alternative */
code += GET(code, 1);
}
while (*code == OP_ALT); /* Loop for each alternative */
return TRUE;
}
/*************************************************
* Check for asserted fixed first char *
*************************************************/
/* During compilation, the "first char" settings from forward assertions are
discarded, because they can cause conflicts with actual literals that follow.
However, if we end up without a first char setting for an unanchored pattern,
it is worth scanning the regex to see if there is an initial asserted first
char. If all branches start with the same asserted char, or with a bracket all
of whose alternatives start with the same asserted char (recurse ad lib), then
we return that char, otherwise -1.
Arguments:
code points to start of expression (the bracket)
options pointer to the options (used to check casing changes)
inassert TRUE if in an assertion
Returns: -1 or the fixed first char
*/
static int
find_firstassertedchar(const uschar *code, int *options, BOOL inassert)
{
register int c = -1;
do {
int d;
const uschar *scode =
first_significant_code(code + 1+LINK_SIZE, options, PCRE_CASELESS, TRUE);
register int op = *scode;
if (op >= OP_BRA) op = OP_BRA;
switch(op)
{
default:
return -1;
case OP_BRA:
case OP_ASSERT:
case OP_ONCE:
case OP_COND:
if ((d = find_firstassertedchar(scode, options, op == OP_ASSERT)) < 0)
return -1;
if (c < 0) c = d; else if (c != d) return -1;
break;
case OP_EXACT: /* Fall through */
scode += 2;
case OP_CHAR:
case OP_CHARNC:
case OP_PLUS:
case OP_MINPLUS:
if (!inassert) return -1;
if (c < 0)
{
c = scode[1];
if ((*options & PCRE_CASELESS) != 0) c |= REQ_CASELESS;
}
else if (c != scode[1]) return -1;
break;
}
code += GET(code, 1);
}
while (*code == OP_ALT);
return c;
}
/*************************************************
* Compile a Regular Expression *
*************************************************/
/* This function takes a string and returns a pointer to a block of store
holding a compiled version of the expression. The original API for this
function had no error code return variable; it is retained for backwards
compatibility. The new function is given a new name.
Arguments:
pattern the regular expression
options various option bits
errorcodeptr pointer to error code variable (pcre_compile2() only)
can be NULL if you don't want a code value
errorptr pointer to pointer to error text
erroroffset ptr offset in pattern where error was detected
tables pointer to character tables or NULL
Returns: pointer to compiled data block, or NULL on error,
with errorptr and erroroffset set
*/
PCRE_DATA_SCOPE pcre *
pcre_compile(const char *pattern, int options, const char **errorptr,
int *erroroffset, const unsigned char *tables)
{
return pcre_compile2(pattern, options, NULL, errorptr, erroroffset, tables);
}
PCRE_DATA_SCOPE pcre *
pcre_compile2(const char *pattern, int options, int *errorcodeptr,
const char **errorptr, int *erroroffset, const unsigned char *tables)
{
real_pcre *re;
int length = 1 + LINK_SIZE; /* For initial BRA plus length */
int c, firstbyte, reqbyte, newline;
int bracount = 0;
int branch_extra = 0;
int branch_newextra;
int item_count = -1;
int name_count = 0;
int max_name_size = 0;
int lastitemlength = 0;
int errorcode = 0;
#ifdef SUPPORT_UTF8
BOOL utf8;
BOOL class_utf8;
#endif
BOOL inescq = FALSE;
BOOL capturing;
unsigned int brastackptr = 0;
size_t size;
uschar *code;
const uschar *codestart;
const uschar *ptr;
compile_data compile_block;
compile_data *cd = &compile_block;
int brastack[BRASTACK_SIZE];
uschar bralenstack[BRASTACK_SIZE];
/* We can't pass back an error message if errorptr is NULL; I guess the best we
can do is just return NULL, but we can set a code value if there is a code
pointer. */
if (errorptr == NULL)
{
if (errorcodeptr != NULL) *errorcodeptr = 99;
return NULL;
}
*errorptr = NULL;
if (errorcodeptr != NULL) *errorcodeptr = ERR0;
/* However, we can give a message for this error */
if (erroroffset == NULL)
{
errorcode = ERR16;
goto PCRE_EARLY_ERROR_RETURN;
}
*erroroffset = 0;
/* Can't support UTF8 unless PCRE has been compiled to include the code. */
#ifdef SUPPORT_UTF8
utf8 = (options & PCRE_UTF8) != 0;
if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0 &&
(*erroroffset = _pcre_valid_utf8((uschar *)pattern, -1)) >= 0)
{
errorcode = ERR44;
goto PCRE_EARLY_ERROR_RETURN;
}
#else
if ((options & PCRE_UTF8) != 0)
{
errorcode = ERR32;
goto PCRE_EARLY_ERROR_RETURN;
}
#endif
if ((options & ~PUBLIC_OPTIONS) != 0)
{
errorcode = ERR17;
goto PCRE_EARLY_ERROR_RETURN;
}
/* Set up pointers to the individual character tables */
if (tables == NULL) tables = _pcre_default_tables;
cd->lcc = tables + lcc_offset;
cd->fcc = tables + fcc_offset;
cd->cbits = tables + cbits_offset;
cd->ctypes = tables + ctypes_offset;
/* Handle different types of newline. The two bits give four cases. The current
code allows for one- or two-byte sequences. */
switch (options & PCRE_NEWLINE_CRLF)
{
default: newline = NEWLINE; break; /* Compile-time default */
case PCRE_NEWLINE_CR: newline = '\r'; break;
case PCRE_NEWLINE_LF: newline = '\n'; break;
case PCRE_NEWLINE_CR+
PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
}
if (newline > 255)
{
cd->nllen = 2;
cd->nl[0] = (newline >> 8) & 255;
cd->nl[1] = newline & 255;
}
else
{
cd->nllen = 1;
cd->nl[0] = newline;
}
/* Maximum back reference and backref bitmap. This is updated for numeric
references during the first pass, but for named references during the actual
compile pass. The bitmap records up to 31 back references to help in deciding
whether (.*) can be treated as anchored or not. */
cd->top_backref = 0;
cd->backref_map = 0;
/* Reflect pattern for debugging output */
DPRINTF(("------------------------------------------------------------------\n"));
DPRINTF(("%s\n", pattern));
/* The first thing to do is to make a pass over the pattern to compute the
amount of store required to hold the compiled code. This does not have to be
perfect as long as errors are overestimates. At the same time we can detect any
flag settings right at the start, and extract them. Make an attempt to correct
for any counted white space if an "extended" flag setting appears late in the
pattern. We can't be so clever for #-comments. */
ptr = (const uschar *)(pattern - 1);
while ((c = *(++ptr)) != 0)
{
int min, max;
int class_optcount;
int bracket_length;
int duplength;
/* If we are inside a \Q...\E sequence, all chars are literal */
if (inescq)
{
if ((options & PCRE_AUTO_CALLOUT) != 0) length += 2 + 2*LINK_SIZE;
goto NORMAL_CHAR;
}
/* Otherwise, first check for ignored whitespace and comments */
if ((options & PCRE_EXTENDED) != 0)
{
if ((cd->ctypes[c] & ctype_space) != 0) continue;
if (c == '#')
{
while (*(++ptr) != 0) if (IS_NEWLINE(ptr)) break;
if (*ptr != 0)
{
ptr += cd->nllen - 1;
continue;
}
break; /* End loop at end of pattern */
}
}
item_count++; /* Is zero for the first non-comment item */
/* Allow space for auto callout before every item except quantifiers. */
if ((options & PCRE_AUTO_CALLOUT) != 0 &&
c != '*' && c != '+' && c != '?' &&
(c != '{' || !is_counted_repeat(ptr + 1)))
length += 2 + 2*LINK_SIZE;
switch(c)
{
/* A backslashed item may be an escaped data character or it may be a
character type. */
case '\\':
c = check_escape(&ptr, &errorcode, bracount, options, FALSE);
if (errorcode != 0) goto PCRE_ERROR_RETURN;
lastitemlength = 1; /* Default length of last item for repeats */
if (c >= 0) /* Data character */
{
length += 2; /* For a one-byte character */
#ifdef SUPPORT_UTF8
if (utf8 && c > 127)
{
int i;
for (i = 0; i < _pcre_utf8_table1_size; i++)
if (c <= _pcre_utf8_table1[i]) break;
length += i;
lastitemlength += i;
}
#endif
continue;
}
/* If \Q, enter "literal" mode */
if (-c == ESC_Q)
{
inescq = TRUE;
continue;
}
/* \X is supported only if Unicode property support is compiled */
#ifndef SUPPORT_UCP
if (-c == ESC_X)
{
errorcode = ERR45;
goto PCRE_ERROR_RETURN;
}
#endif
/* \P and \p are for Unicode properties, but only when the support has
been compiled. Each item needs 3 bytes. */
else if (-c == ESC_P || -c == ESC_p)
{
#ifdef SUPPORT_UCP
BOOL negated;
BOOL pdata;
length += 3;
lastitemlength = 3;
if (get_ucp(&ptr, &negated, &pdata, &errorcode) < 0)
goto PCRE_ERROR_RETURN;
continue;
#else
errorcode = ERR45;
goto PCRE_ERROR_RETURN;
#endif
}
/* Other escapes need one byte */
length++;
/* A back reference needs an additional 2 bytes, plus either one or 5
bytes for a repeat. We also need to keep the value of the highest
back reference. */
if (c <= -ESC_REF)
{
int refnum = -c - ESC_REF;
cd->backref_map |= (refnum < 32)? (1 << refnum) : 1;
if (refnum > cd->top_backref)
cd->top_backref = refnum;
length += 2; /* For single back reference */
if (ptr[1] == '{' && is_counted_repeat(ptr+2))
{
ptr = read_repeat_counts(ptr+2, &min, &max, &errorcode);
if (errorcode != 0) goto PCRE_ERROR_RETURN;
if ((min == 0 && (max == 1 || max == -1)) ||
(min == 1 && max == -1))
length++;
else length += 5;
if (ptr[1] == '?') ptr++;
}
}
continue;
case '^': /* Single-byte metacharacters */
case '.':
case '$':
length++;
lastitemlength = 1;
continue;
case '*': /* These repeats won't be after brackets; */
case '+': /* those are handled separately */
case '?':
length++;
goto POSESSIVE; /* A few lines below */
/* This covers the cases of braced repeats after a single char, metachar,
class, or back reference. */
case '{':
if (!is_counted_repeat(ptr+1)) goto NORMAL_CHAR;
ptr = read_repeat_counts(ptr+1, &min, &max, &errorcode);
if (errorcode != 0) goto PCRE_ERROR_RETURN;
/* These special cases just insert one extra opcode */
if ((min == 0 && (max == 1 || max == -1)) ||
(min == 1 && max == -1))
length++;
/* These cases might insert additional copies of a preceding character. */
else
{
if (min != 1)
{
length -= lastitemlength; /* Uncount the original char or metachar */
if (min > 0) length += 3 + lastitemlength;
}
length += lastitemlength + ((max > 0)? 3 : 1);
}
if (ptr[1] == '?') ptr++; /* Needs no extra length */
POSESSIVE: /* Test for possessive quantifier */
if (ptr[1] == '+')
{
ptr++;
length += 2 + 2*LINK_SIZE; /* Allow for atomic brackets */
}
continue;
/* An alternation contains an offset to the next branch or ket. If any ims
options changed in the previous branch(es), and/or if we are in a
lookbehind assertion, extra space will be needed at the start of the
branch. This is handled by branch_extra. */
case '|':
length += 1 + LINK_SIZE + branch_extra;
continue;
/* A character class uses 33 characters provided that all the character
values are less than 256. Otherwise, it uses a bit map for low valued
characters, and individual items for others. Don't worry about character
types that aren't allowed in classes - they'll get picked up during the
compile. A character class that contains only one single-byte character
uses 2 or 3 bytes, depending on whether it is negated or not. Notice this
where we can. (In UTF-8 mode we can do this only for chars < 128.) */
case '[':
if (*(++ptr) == '^')
{
class_optcount = 10; /* Greater than one */
ptr++;
}
else class_optcount = 0;
#ifdef SUPPORT_UTF8
class_utf8 = FALSE;
#endif
/* Written as a "do" so that an initial ']' is taken as data */
if (*ptr != 0) do
{
/* Inside \Q...\E everything is literal except \E */
if (inescq)
{
if (*ptr != '\\' || ptr[1] != 'E') goto GET_ONE_CHARACTER;
inescq = FALSE;
ptr += 1;
continue;
}
/* Outside \Q...\E, check for escapes */
if (*ptr == '\\')
{
c = check_escape(&ptr, &errorcode, bracount, options, TRUE);
if (errorcode != 0) goto PCRE_ERROR_RETURN;
/* \b is backspace inside a class; \X is literal */
if (-c == ESC_b) c = '\b';
else if (-c == ESC_X) c = 'X';
/* \Q enters quoting mode */
else if (-c == ESC_Q)
{
inescq = TRUE;
continue;
}
/* Handle escapes that turn into characters */
if (c >= 0) goto NON_SPECIAL_CHARACTER;
/* Escapes that are meta-things. The normal ones just affect the
bit map, but Unicode properties require an XCLASS extended item. */
else
{
class_optcount = 10; /* \d, \s etc; make sure > 1 */
#ifdef SUPPORT_UTF8
if (-c == ESC_p || -c == ESC_P)
{
if (!class_utf8)
{
class_utf8 = TRUE;
length += LINK_SIZE + 2;
}
length += 3;
}
#endif
}
}
/* Check the syntax for POSIX stuff. The bits we actually handle are
checked during the real compile phase. */
else if (*ptr == '[' &&
(ptr[1] == ':' || ptr[1] == '.' || ptr[1] == '=') &&
check_posix_syntax(ptr, &ptr, cd))
{
ptr++;
class_optcount = 10; /* Make sure > 1 */
}
/* Anything else increments the possible optimization count. We have to
detect ranges here so that we can compute the number of extra ranges for
caseless wide characters when UCP support is available. If there are wide
characters, we are going to have to use an XCLASS, even for single
characters. */
else
{
int d;
GET_ONE_CHARACTER:
#ifdef SUPPORT_UTF8
if (utf8)
{
int extra = 0;
GETCHARLEN(c, ptr, extra);
ptr += extra;
}
else c = *ptr;
#else
c = *ptr;
#endif
/* Come here from handling \ above when it escapes to a char value */
NON_SPECIAL_CHARACTER:
class_optcount++;
d = -1;
if (ptr[1] == '-')
{
uschar const *hyptr = ptr++;
if (ptr[1] == '\\')
{
ptr++;
d = check_escape(&ptr, &errorcode, bracount, options, TRUE);
if (errorcode != 0) goto PCRE_ERROR_RETURN;
if (-d == ESC_b) d = '\b'; /* backspace */
else if (-d == ESC_X) d = 'X'; /* literal X in a class */
}
else if (ptr[1] != 0 && ptr[1] != ']')
{
ptr++;
#ifdef SUPPORT_UTF8
if (utf8)
{
int extra = 0;
GETCHARLEN(d, ptr, extra);
ptr += extra;
}
else
#endif
d = *ptr;
}
if (d < 0) ptr = hyptr; /* go back to hyphen as data */
}
/* If d >= 0 we have a range. In UTF-8 mode, if the end is > 255, or >
127 for caseless matching, we will need to use an XCLASS. */
if (d >= 0)
{
class_optcount = 10; /* Ensure > 1 */
if (d < c)
{
errorcode = ERR8;
goto PCRE_ERROR_RETURN;
}
#ifdef SUPPORT_UTF8
if (utf8 && (d > 255 || ((options & PCRE_CASELESS) != 0 && d > 127)))
{
uschar buffer[6];
if (!class_utf8) /* Allow for XCLASS overhead */
{
class_utf8 = TRUE;
length += LINK_SIZE + 2;
}
#ifdef SUPPORT_UCP
/* If we have UCP support, find out how many extra ranges are
needed to map the other case of characters within this range. We
have to mimic the range optimization here, because extending the
range upwards might push d over a boundary that makes is use
another byte in the UTF-8 representation. */
if ((options & PCRE_CASELESS) != 0)
{
int occ, ocd;
int cc = c;
int origd = d;
while (get_othercase_range(&cc, origd, &occ, &ocd))
{
if (occ >= c && ocd <= d) continue; /* Skip embedded */
if (occ < c && ocd >= c - 1) /* Extend the basic range */
{ /* if there is overlap, */
c = occ; /* noting that if occ < c */
continue; /* we can't have ocd > d */
} /* because a subrange is */
if (ocd > d && occ <= d + 1) /* always shorter than */
{ /* the basic range. */
d = ocd;
continue;
}
/* An extra item is needed */
length += 1 + _pcre_ord2utf8(occ, buffer) +
((occ == ocd)? 0 : _pcre_ord2utf8(ocd, buffer));
}
}
#endif /* SUPPORT_UCP */
/* The length of the (possibly extended) range */
length += 1 + _pcre_ord2utf8(c, buffer) + _pcre_ord2utf8(d, buffer);
}
#endif /* SUPPORT_UTF8 */
}
/* We have a single character. There is nothing to be done unless we
are in UTF-8 mode. If the char is > 255, or 127 when caseless, we must
allow for an XCL_SINGLE item, doubled for caselessness if there is UCP
support. */
else
{
#ifdef SUPPORT_UTF8
if (utf8 && (c > 255 || ((options & PCRE_CASELESS) != 0 && c > 127)))
{
uschar buffer[6];
class_optcount = 10; /* Ensure > 1 */
if (!class_utf8) /* Allow for XCLASS overhead */
{
class_utf8 = TRUE;
length += LINK_SIZE + 2;
}
#ifdef SUPPORT_UCP
length += (((options & PCRE_CASELESS) != 0)? 2 : 1) *
(1 + _pcre_ord2utf8(c, buffer));
#else /* SUPPORT_UCP */
length += 1 + _pcre_ord2utf8(c, buffer);
#endif /* SUPPORT_UCP */
}
#endif /* SUPPORT_UTF8 */
}
}
}
while (*(++ptr) != 0 && (inescq || *ptr != ']')); /* Concludes "do" above */
if (*ptr == 0) /* Missing terminating ']' */
{
errorcode = ERR6;
goto PCRE_ERROR_RETURN;
}
/* We can optimize when there was only one optimizable character. Repeats
for positive and negated single one-byte chars are handled by the general
code. Here, we handle repeats for the class opcodes. */
if (class_optcount == 1) length += 3; else
{
length += 33;
/* A repeat needs either 1 or 5 bytes. If it is a possessive quantifier,
we also need extra for wrapping the whole thing in a sub-pattern. */
if (*ptr != 0 && ptr[1] == '{' && is_counted_repeat(ptr+2))
{
ptr = read_repeat_counts(ptr+2, &min, &max, &errorcode);
if (errorcode != 0) goto PCRE_ERROR_RETURN;
if ((min == 0 && (max == 1 || max == -1)) ||
(min == 1 && max == -1))
length++;
else length += 5;
if (ptr[1] == '+')
{
ptr++;
length += 2 + 2*LINK_SIZE;
}
else if (ptr[1] == '?') ptr++;
}
}
continue;
/* Brackets may be genuine groups or special things */
case '(':
branch_newextra = 0;
bracket_length = 1 + LINK_SIZE;
capturing = FALSE;
/* Handle special forms of bracket, which all start (? */
if (ptr[1] == '?')
{
int set, unset;
int *optset;
switch (c = ptr[2])
{
/* Skip over comments entirely */
case '#':
ptr += 3;
while (*ptr != 0 && *ptr != ')') ptr++;
if (*ptr == 0)
{
errorcode = ERR18;
goto PCRE_ERROR_RETURN;
}
continue;
/* Non-referencing groups and lookaheads just move the pointer on, and
then behave like a non-special bracket, except that they don't increment
the count of extracting brackets. Ditto for the "once only" bracket,
which is in Perl from version 5.005. */
case ':':
case '=':
case '!':
case '>':
ptr += 2;
break;
/* Named subpatterns are an extension copied from Python */
case 'P':
ptr += 3;
/* Handle the definition of a named subpattern */
if (*ptr == '<')
{
const uschar *p; /* Don't amalgamate; some compilers */
p = ++ptr; /* grumble at autoincrement in declaration */
while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
if (*ptr != '>')
{
errorcode = ERR42;
goto PCRE_ERROR_RETURN;
}
name_count++;
if (name_count > MAX_NAME_COUNT)
{
errorcode = ERR49;
goto PCRE_ERROR_RETURN;
}
if (ptr - p > max_name_size)
{
max_name_size = (ptr - p);
if (max_name_size > MAX_NAME_SIZE)
{
errorcode = ERR48;
goto PCRE_ERROR_RETURN;
}
}
capturing = TRUE; /* Named parentheses are always capturing */
break; /* Go handle capturing parentheses */
}
/* Handle back references and recursive calls to named subpatterns */
if (*ptr == '=' || *ptr == '>')
{
length += 3 + 3*LINK_SIZE; /* Allow for the automatic "once" */
while ((cd->ctypes[*(++ptr)] & ctype_word) != 0);
if (*ptr != ')')
{
errorcode = ERR42;
goto PCRE_ERROR_RETURN;
}
goto RECURSE_CHECK_QUANTIFIED;
}
/* Unknown character after (?P */
errorcode = ERR41;
goto PCRE_ERROR_RETURN;
/* (?R) specifies a recursive call to the regex, which is an extension
to provide the facility which can be obtained by (?p{perl-code}) in
Perl 5.6. In Perl 5.8 this has become (??{perl-code}).
From PCRE 4.00, items such as (?3) specify subroutine-like "calls" to
the appropriate numbered brackets. This includes both recursive and
non-recursive calls. (?R) is now synonymous with (?0). */
case 'R':
ptr++;
case '0': case '1': case '2': case '3': case '4':
case '5': case '6': case '7': case '8': case '9':
ptr += 2;
if (c != 'R')
while ((digitab[*(++ptr)] & ctype_digit) != 0);
if (*ptr != ')')
{
errorcode = ERR29;
goto PCRE_ERROR_RETURN;
}
length += 3 + 3*LINK_SIZE; /* Allows for the automatic "once" */
/* If this item is quantified, it will get wrapped inside brackets so
as to use the code for quantified brackets. We jump down and use the
code that handles this for real brackets. Come here from code for
named recursions/subroutines. */
RECURSE_CHECK_QUANTIFIED:
if (ptr[1] == '+' || ptr[1] == '*' || ptr[1] == '?' || ptr[1] == '{')
{
length += 2 + 2 * LINK_SIZE; /* to make bracketed */
duplength = 5 + 3 * LINK_SIZE;
goto HANDLE_QUANTIFIED_BRACKETS;
}
continue;
/* (?C) is an extension which provides "callout" - to provide a bit of
the functionality of the Perl (?{...}) feature. An optional number may
follow (default is zero). */
case 'C':
ptr += 2;
while ((digitab[*(++ptr)] & ctype_digit) != 0);
if (*ptr != ')')
{
errorcode = ERR39;
goto PCRE_ERROR_RETURN;
}
length += 2 + 2*LINK_SIZE;
continue;
/* Lookbehinds are in Perl from version 5.005 */
case '<':
ptr += 3;
if (*ptr == '=' || *ptr == '!')
{
branch_newextra = 1 + LINK_SIZE;
length += 1 + LINK_SIZE; /* For the first branch */
break;
}
errorcode = ERR24;
goto PCRE_ERROR_RETURN;
/* Conditionals are in Perl from version 5.005. The bracket must either
be followed by a number (for bracket reference) or by an assertion
group. PCRE extends this by allowing a name to reference a named group;
unfortunately, previously 'R' was implemented for a recursion test.
When this is compiled, we look for the named group 'R' first. At this
point we just do a basic syntax check. */
case '(':
if ((cd->ctypes[ptr[3]] & ctype_word) != 0)
{
ptr += 4;
length += 3;
while ((cd->ctypes[*ptr] & ctype_word) != 0) ptr++;
if (*ptr != ')')
{
errorcode = ERR26;
goto PCRE_ERROR_RETURN;
}
}
else /* An assertion must follow */
{
ptr++; /* Can treat like ':' as far as spacing is concerned */
if (ptr[2] != '?' ||
(ptr[3] != '=' && ptr[3] != '!' && ptr[3] != '<') )
{
ptr += 2; /* To get right offset in message */
errorcode = ERR28;
goto PCRE_ERROR_RETURN;
}
}
break;
/* Else loop checking valid options until ) is met. Anything else is an
error. If we are without any brackets, i.e. at top level, the settings
act as if specified in the options, so massage the options immediately.
This is for backward compatibility with Perl 5.004. */
default:
set = unset = 0;
optset = &set;
ptr += 2;
for (;; ptr++)
{
c = *ptr;
switch (c)
{
case 'i':
*optset |= PCRE_CASELESS;
continue;
case 'J':
*optset |= PCRE_DUPNAMES;
options |= PCRE_JCHANGED; /* Record that it changed */
continue;
case 'm':
*optset |= PCRE_MULTILINE;
continue;
case 's':
*optset |= PCRE_DOTALL;
continue;
case 'x':
*optset |= PCRE_EXTENDED;
continue;
case 'X':
*optset |= PCRE_EXTRA;
continue;
case 'U':
*optset |= PCRE_UNGREEDY;
continue;
case '-':
optset = &unset;
continue;
/* A termination by ')' indicates an options-setting-only item; if
this is at the very start of the pattern (indicated by item_count
being zero), we use it to set the global options. This is helpful
when analyzing the pattern for first characters, etc. Otherwise
nothing is done here and it is handled during the compiling
process.
We allow for more than one options setting at the start. If such
settings do not change the existing options, nothing is compiled.
However, we must leave space just in case something is compiled.
This can happen for pathological sequences such as (?i)(?-i)
because the global options will end up with -i set. The space is
small and not significant. (Before I did this there was a reported
bug with (?i)(?-i) in a machine-generated pattern.)
[Historical note: Up to Perl 5.8, options settings at top level
were always global settings, wherever they appeared in the pattern.
That is, they were equivalent to an external setting. From 5.8
onwards, they apply only to what follows (which is what you might
expect).] */
case ')':
if (item_count == 0)
{
options = (options | set) & (~unset);
set = unset = 0; /* To save length */
item_count--; /* To allow for several */
length += 2;
}
/* Fall through */
/* A termination by ':' indicates the start of a nested group with
the given options set. This is again handled at compile time, but
we must allow for compiled space if any of the ims options are
set. We also have to allow for resetting space at the end of
the group, which is why 4 is added to the length and not just 2.
If there are several changes of options within the same group, this
will lead to an over-estimate on the length, but this shouldn't
matter very much. We also have to allow for resetting options at
the start of any alternations, which we do by setting
branch_newextra to 2. */
case ':':
if (((set|unset) & PCRE_IMS) != 0)
{
length += 4;
branch_newextra = 2;
}
goto END_OPTIONS;
/* Unrecognized option character */
default:
errorcode = ERR12;
goto PCRE_ERROR_RETURN;
}
}
/* If we hit a closing bracket, that's it - this is a freestanding
option-setting. We need to ensure that branch_extra is updated if
necessary. The only values branch_newextra can have here are 0 or 2.
If the value is 2, then branch_extra must either be 2 or 5, depending
on whether this is a lookbehind group or not. */
END_OPTIONS:
if (c == ')')
{
if (branch_newextra == 2 &&
(branch_extra == 0 || branch_extra == 1+LINK_SIZE))
branch_extra += branch_newextra;
continue;
}
/* If options were terminated by ':' control comes here. This is a
non-capturing group with an options change. There is nothing more that
needs to be done because "capturing" is already set FALSE by default;
we can just fall through. */
}
}
/* Ordinary parentheses, not followed by '?', are capturing unless
PCRE_NO_AUTO_CAPTURE is set. */
else capturing = (options & PCRE_NO_AUTO_CAPTURE) == 0;
/* Capturing brackets must be counted so we can process escapes in a
Perlish way. If the number exceeds EXTRACT_BASIC_MAX we are going to need
an additional 3 bytes of memory per capturing bracket. */
if (capturing)
{
bracount++;
if (bracount > EXTRACT_BASIC_MAX) bracket_length += 3;
}
/* Save length for computing whole length at end if there's a repeat that
requires duplication of the group. Also save the current value of
branch_extra, and start the new group with the new value. If non-zero, this
will either be 2 for a (?imsx: group, or 3 for a lookbehind assertion. */
if (brastackptr >= sizeof(brastack)/sizeof(int))
{
errorcode = ERR19;
goto PCRE_ERROR_RETURN;
}
bralenstack[brastackptr] = branch_extra;
branch_extra = branch_newextra;
brastack[brastackptr++] = length;
length += bracket_length;
continue;
/* Handle ket. Look for subsequent max/min; for certain sets of values we
have to replicate this bracket up to that many times. If brastackptr is
0 this is an unmatched bracket which will generate an error, but take care
not to try to access brastack[-1] when computing the length and restoring
the branch_extra value. */
case ')':
length += 1 + LINK_SIZE;
if (brastackptr > 0)
{
duplength = length - brastack[--brastackptr];
branch_extra = bralenstack[brastackptr];
/* This is a paranoid check to stop integer overflow later on */
if (duplength > MAX_DUPLENGTH)
{
errorcode = ERR50;
goto PCRE_ERROR_RETURN;
}
}
else duplength = 0;
/* The following code is also used when a recursion such as (?3) is
followed by a quantifier, because in that case, it has to be wrapped inside
brackets so that the quantifier works. The value of duplength must be
set before arrival. */
HANDLE_QUANTIFIED_BRACKETS:
/* Leave ptr at the final char; for read_repeat_counts this happens
automatically; for the others we need an increment. */
if ((c = ptr[1]) == '{' && is_counted_repeat(ptr+2))
{
ptr = read_repeat_counts(ptr+2, &min, &max, &errorcode);
if (errorcode != 0) goto PCRE_ERROR_RETURN;
}
else if (c == '*') { min = 0; max = -1; ptr++; }
else if (c == '+') { min = 1; max = -1; ptr++; }
else if (c == '?') { min = 0; max = 1; ptr++; }
else { min = 1; max = 1; }
/* If the minimum is zero, we have to allow for an OP_BRAZERO before the
group, and if the maximum is greater than zero, we have to replicate
maxval-1 times; each replication acquires an OP_BRAZERO plus a nesting
bracket set. */
if (min == 0)
{
length++;
if (max > 0) length += (max - 1) * (duplength + 3 + 2*LINK_SIZE);
}
/* When the minimum is greater than zero, we have to replicate up to
minval-1 times, with no additions required in the copies. Then, if there
is a limited maximum we have to replicate up to maxval-1 times allowing
for a BRAZERO item before each optional copy and nesting brackets for all
but one of the optional copies. */
else
{
length += (min - 1) * duplength;
if (max > min) /* Need this test as max=-1 means no limit */
length += (max - min) * (duplength + 3 + 2*LINK_SIZE)
- (2 + 2*LINK_SIZE);
}
/* Allow space for once brackets for "possessive quantifier" */
if (ptr[1] == '+')
{
ptr++;
length += 2 + 2*LINK_SIZE;
}
continue;
/* Non-special character. It won't be space or # in extended mode, so it is
always a genuine character. If we are in a \Q...\E sequence, check for the
end; if not, we have a literal. */
default:
NORMAL_CHAR:
if (inescq && c == '\\' && ptr[1] == 'E')
{
inescq = FALSE;
ptr++;
continue;
}
length += 2; /* For a one-byte character */
lastitemlength = 1; /* Default length of last item for repeats */
/* In UTF-8 mode, check for additional bytes. */
#ifdef SUPPORT_UTF8
if (utf8 && (c & 0xc0) == 0xc0)
{
while ((ptr[1] & 0xc0) == 0x80) /* Can't flow over the end */
{ /* because the end is marked */
lastitemlength++; /* by a zero byte. */
length++;
ptr++;
}
}
#endif
continue;
}
}
length += 2 + LINK_SIZE; /* For final KET and END */
if ((options & PCRE_AUTO_CALLOUT) != 0)
length += 2 + 2*LINK_SIZE; /* For final callout */
if (length > MAX_PATTERN_SIZE)
{
errorcode = ERR20;
goto PCRE_EARLY_ERROR_RETURN;
}
/* Compute the size of data block needed and get it, either from malloc or
externally provided function. Integer overflow should no longer be possible
because nowadays we limit the maximum value of name_count and max_name size. */
size = length + sizeof(real_pcre) + name_count * (max_name_size + 3);
re = (real_pcre *)(pcre_malloc)(size);
if (re == NULL)
{
errorcode = ERR21;
goto PCRE_EARLY_ERROR_RETURN;
}
/* Put in the magic number, and save the sizes, options, and character table
pointer. NULL is used for the default character tables. The nullpad field is at
the end; it's there to help in the case when a regex compiled on a system with
4-byte pointers is run on another with 8-byte pointers. */
re->magic_number = MAGIC_NUMBER;
re->size = size;
re->options = options;
re->dummy1 = 0;
re->name_table_offset = sizeof(real_pcre);
re->name_entry_size = max_name_size + 3;
re->name_count = name_count;
re->ref_count = 0;
re->tables = (tables == _pcre_default_tables)? NULL : tables;
re->nullpad = NULL;
/* The starting points of the name/number translation table and of the code are
passed around in the compile data block. */
cd->names_found = 0;
cd->name_entry_size = max_name_size + 3;
cd->name_table = (uschar *)re + re->name_table_offset;
codestart = cd->name_table + re->name_entry_size * re->name_count;
cd->start_code = codestart;
cd->start_pattern = (const uschar *)pattern;
cd->req_varyopt = 0;
cd->nopartial = FALSE;
/* Set up a starting, non-extracting bracket, then compile the expression. On
error, errorcode will be set non-zero, so we don't need to look at the result
of the function here. */
ptr = (const uschar *)pattern;
code = (uschar *)codestart;
*code = OP_BRA;
bracount = 0;
(void)compile_regex(options, options & PCRE_IMS, &bracount, &code, &ptr,
&errorcode, FALSE, 0, &firstbyte, &reqbyte, NULL, cd);
re->top_bracket = bracount;
re->top_backref = cd->top_backref;
if (cd->nopartial) re->options |= PCRE_NOPARTIAL;
/* If not reached end of pattern on success, there's an excess bracket. */
if (errorcode == 0 && *ptr != 0) errorcode = ERR22;
/* Fill in the terminating state and check for disastrous overflow, but
if debugging, leave the test till after things are printed out. */
*code++ = OP_END;
#ifndef DEBUG
if (code - codestart > length) errorcode = ERR23;
#endif
/* Give an error if there's back reference to a non-existent capturing
subpattern. */
if (re->top_backref > re->top_bracket) errorcode = ERR15;
/* Failed to compile, or error while post-processing */
if (errorcode != 0)
{
(pcre_free)(re);
PCRE_ERROR_RETURN:
*erroroffset = ptr - (const uschar *)pattern;
PCRE_EARLY_ERROR_RETURN:
*errorptr = error_texts[errorcode];
if (errorcodeptr != NULL) *errorcodeptr = errorcode;
return NULL;
}
/* If the anchored option was not passed, set the flag if we can determine that
the pattern is anchored by virtue of ^ characters or \A or anything else (such
as starting with .* when DOTALL is set).
Otherwise, if we know what the first character has to be, save it, because that
speeds up unanchored matches no end. If not, see if we can set the
PCRE_STARTLINE flag. This is helpful for multiline matches when all branches
start with ^. and also when all branches start with .* for non-DOTALL matches.
*/
if ((options & PCRE_ANCHORED) == 0)
{
int temp_options = options;
if (is_anchored(codestart, &temp_options, 0, cd->backref_map))
re->options |= PCRE_ANCHORED;
else
{
if (firstbyte < 0)
firstbyte = find_firstassertedchar(codestart, &temp_options, FALSE);
if (firstbyte >= 0) /* Remove caseless flag for non-caseable chars */
{
int ch = firstbyte & 255;
re->first_byte = ((firstbyte & REQ_CASELESS) != 0 &&
cd->fcc[ch] == ch)? ch : firstbyte;
re->options |= PCRE_FIRSTSET;
}
else if (is_startline(codestart, 0, cd->backref_map))
re->options |= PCRE_STARTLINE;
}
}
/* For an anchored pattern, we use the "required byte" only if it follows a
variable length item in the regex. Remove the caseless flag for non-caseable
bytes. */
if (reqbyte >= 0 &&
((re->options & PCRE_ANCHORED) == 0 || (reqbyte & REQ_VARY) != 0))
{
int ch = reqbyte & 255;
re->req_byte = ((reqbyte & REQ_CASELESS) != 0 &&
cd->fcc[ch] == ch)? (reqbyte & ~REQ_CASELESS) : reqbyte;
re->options |= PCRE_REQCHSET;
}
/* Print out the compiled data if debugging is enabled. This is never the
case when building a production library. */
#ifdef DEBUG
printf("Length = %d top_bracket = %d top_backref = %d\n",
length, re->top_bracket, re->top_backref);
if (re->options != 0)
{
printf("%s%s%s%s%s%s%s%s%s\n",
((re->options & PCRE_NOPARTIAL) != 0)? "nopartial " : "",
((re->options & PCRE_ANCHORED) != 0)? "anchored " : "",
((re->options & PCRE_CASELESS) != 0)? "caseless " : "",
((re->options & PCRE_EXTENDED) != 0)? "extended " : "",
((re->options & PCRE_MULTILINE) != 0)? "multiline " : "",
((re->options & PCRE_DOTALL) != 0)? "dotall " : "",
((re->options & PCRE_DOLLAR_ENDONLY) != 0)? "endonly " : "",
((re->options & PCRE_EXTRA) != 0)? "extra " : "",
((re->options & PCRE_UNGREEDY) != 0)? "ungreedy " : "");
}
if ((re->options & PCRE_FIRSTSET) != 0)
{
int ch = re->first_byte & 255;
const char *caseless = ((re->first_byte & REQ_CASELESS) == 0)?
"" : " (caseless)";
if (isprint(ch)) printf("First char = %c%s\n", ch, caseless);
else printf("First char = \\x%02x%s\n", ch, caseless);
}
if ((re->options & PCRE_REQCHSET) != 0)
{
int ch = re->req_byte & 255;
const char *caseless = ((re->req_byte & REQ_CASELESS) == 0)?
"" : " (caseless)";
if (isprint(ch)) printf("Req char = %c%s\n", ch, caseless);
else printf("Req char = \\x%02x%s\n", ch, caseless);
}
pcre_printint(re, stdout);
/* This check is done here in the debugging case so that the code that
was compiled can be seen. */
if (code - codestart > length)
{
(pcre_free)(re);
*errorptr = error_texts[ERR23];
*erroroffset = ptr - (uschar *)pattern;
if (errorcodeptr != NULL) *errorcodeptr = ERR23;
return NULL;
}
#endif
return (pcre *)re;
}
/* End of pcre_compile.c */
|