aboutsummaryrefslogtreecommitdiff
path: root/src/common/util.c
blob: e676530b29d7481ec028587f2c836e23707961a2 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
/* Copyright (c) 2003, Roger Dingledine
 * Copyright (c) 2004-2006, Roger Dingledine, Nick Mathewson.
 * Copyright (c) 2007-2011, The Tor Project, Inc. */
/* See LICENSE for licensing information */

/**
 * \file util.c
 * \brief Common functions for strings, IO, network, data structures,
 * process control.
 **/

/* This is required on rh7 to make strptime not complain.
 */
#define _GNU_SOURCE

#include "orconfig.h"
#define UTIL_PRIVATE
#include "util.h"
#include "torlog.h"
#undef log
#include "crypto.h"
#include "torint.h"
#include "container.h"
#include "address.h"

#ifdef MS_WINDOWS
#include <io.h>
#include <direct.h>
#include <process.h>
#include <tchar.h>
#else
#include <dirent.h>
#include <pwd.h>
#endif

/* math.h needs this on Linux */
#ifndef __USE_ISOC99
#define __USE_ISOC99 1
#endif
#include <math.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <assert.h>

#ifdef HAVE_NETINET_IN_H
#include <netinet/in.h>
#endif
#ifdef HAVE_ARPA_INET_H
#include <arpa/inet.h>
#endif
#ifdef HAVE_ERRNO_H
#include <errno.h>
#endif
#ifdef HAVE_SYS_SOCKET_H
#include <sys/socket.h>
#endif
#ifdef HAVE_SYS_TIME_H
#include <sys/time.h>
#endif
#ifdef HAVE_UNISTD_H
#include <unistd.h>
#endif
#ifdef HAVE_SYS_STAT_H
#include <sys/stat.h>
#endif
#ifdef HAVE_SYS_FCNTL_H
#include <sys/fcntl.h>
#endif
#ifdef HAVE_FCNTL_H
#include <fcntl.h>
#endif
#ifdef HAVE_TIME_H
#include <time.h>
#endif
#ifdef HAVE_MALLOC_MALLOC_H
#include <malloc/malloc.h>
#endif
#ifdef HAVE_MALLOC_H
#ifndef OPENBSD
/* OpenBSD has a malloc.h, but for our purposes, it only exists in order to
 * scold us for being so stupid as to autodetect its presence.  To be fair,
 * they've done this since 1996, when autoconf was only 5 years old. */
#include <malloc.h>
#endif
#endif
#ifdef HAVE_MALLOC_NP_H
#include <malloc_np.h>
#endif

/* =====
 * Memory management
 * ===== */
#ifdef USE_DMALLOC
 #undef strndup
 #include <dmalloc.h>
 /* Macro to pass the extra dmalloc args to another function. */
 #define DMALLOC_FN_ARGS , file, line

 #if defined(HAVE_DMALLOC_STRDUP)
 /* the dmalloc_strdup should be fine as defined */
 #elif defined(HAVE_DMALLOC_STRNDUP)
 #define dmalloc_strdup(file, line, string, xalloc_b) \
         dmalloc_strndup(file, line, (string), -1, xalloc_b)
 #else
 #error "No dmalloc_strdup or equivalent"
 #endif

#else /* not using dmalloc */

 #define DMALLOC_FN_ARGS
#endif

/** Allocate a chunk of <b>size</b> bytes of memory, and return a pointer to
 * result.  On error, log and terminate the process.  (Same as malloc(size),
 * but never returns NULL.)
 *
 * <b>file</b> and <b>line</b> are used if dmalloc is enabled, and
 * ignored otherwise.
 */
void *
_tor_malloc(size_t size DMALLOC_PARAMS)
{
  void *result;

  tor_assert(size < SIZE_T_CEILING);

#ifndef MALLOC_ZERO_WORKS
  /* Some libc mallocs don't work when size==0. Override them. */
  if (size==0) {
    size=1;
  }
#endif

#ifdef USE_DMALLOC
  result = dmalloc_malloc(file, line, size, DMALLOC_FUNC_MALLOC, 0, 0);
#else
  result = malloc(size);
#endif

  if (PREDICT_UNLIKELY(result == NULL)) {
    log_err(LD_MM,"Out of memory on malloc(). Dying.");
    /* If these functions die within a worker process, they won't call
     * spawn_exit, but that's ok, since the parent will run out of memory soon
     * anyway. */
    exit(1);
  }
  return result;
}

/** Allocate a chunk of <b>size</b> bytes of memory, fill the memory with
 * zero bytes, and return a pointer to the result.  Log and terminate
 * the process on error.  (Same as calloc(size,1), but never returns NULL.)
 */
void *
_tor_malloc_zero(size_t size DMALLOC_PARAMS)
{
  /* You may ask yourself, "wouldn't it be smart to use calloc instead of
   * malloc+memset?  Perhaps libc's calloc knows some nifty optimization trick
   * we don't!"  Indeed it does, but its optimizations are only a big win when
   * we're allocating something very big (it knows if it just got the memory
   * from the OS in a pre-zeroed state).  We don't want to use tor_malloc_zero
   * for big stuff, so we don't bother with calloc. */
  void *result = _tor_malloc(size DMALLOC_FN_ARGS);
  memset(result, 0, size);
  return result;
}

/** Change the size of the memory block pointed to by <b>ptr</b> to <b>size</b>
 * bytes long; return the new memory block.  On error, log and
 * terminate. (Like realloc(ptr,size), but never returns NULL.)
 */
void *
_tor_realloc(void *ptr, size_t size DMALLOC_PARAMS)
{
  void *result;

  tor_assert(size < SIZE_T_CEILING);

#ifdef USE_DMALLOC
  result = dmalloc_realloc(file, line, ptr, size, DMALLOC_FUNC_REALLOC, 0);
#else
  result = realloc(ptr, size);
#endif

  if (PREDICT_UNLIKELY(result == NULL)) {
    log_err(LD_MM,"Out of memory on realloc(). Dying.");
    exit(1);
  }
  return result;
}

/** Return a newly allocated copy of the NUL-terminated string s. On
 * error, log and terminate.  (Like strdup(s), but never returns
 * NULL.)
 */
char *
_tor_strdup(const char *s DMALLOC_PARAMS)
{
  char *dup;
  tor_assert(s);

#ifdef USE_DMALLOC
  dup = dmalloc_strdup(file, line, s, 0);
#else
  dup = strdup(s);
#endif
  if (PREDICT_UNLIKELY(dup == NULL)) {
    log_err(LD_MM,"Out of memory on strdup(). Dying.");
    exit(1);
  }
  return dup;
}

/** Allocate and return a new string containing the first <b>n</b>
 * characters of <b>s</b>.  If <b>s</b> is longer than <b>n</b>
 * characters, only the first <b>n</b> are copied.  The result is
 * always NUL-terminated.  (Like strndup(s,n), but never returns
 * NULL.)
 */
char *
_tor_strndup(const char *s, size_t n DMALLOC_PARAMS)
{
  char *dup;
  tor_assert(s);
  tor_assert(n < SIZE_T_CEILING);
  dup = _tor_malloc((n+1) DMALLOC_FN_ARGS);
  /* Performance note: Ordinarily we prefer strlcpy to strncpy.  But
   * this function gets called a whole lot, and platform strncpy is
   * much faster than strlcpy when strlen(s) is much longer than n.
   */
  strncpy(dup, s, n);
  dup[n]='\0';
  return dup;
}

/** Allocate a chunk of <b>len</b> bytes, with the same contents as the
 * <b>len</b> bytes starting at <b>mem</b>. */
void *
_tor_memdup(const void *mem, size_t len DMALLOC_PARAMS)
{
  char *dup;
  tor_assert(len < SIZE_T_CEILING);
  tor_assert(mem);
  dup = _tor_malloc(len DMALLOC_FN_ARGS);
  memcpy(dup, mem, len);
  return dup;
}

/** Helper for places that need to take a function pointer to the right
 * spelling of "free()". */
void
_tor_free(void *mem)
{
  tor_free(mem);
}

#if defined(HAVE_MALLOC_GOOD_SIZE) && !defined(HAVE_MALLOC_GOOD_SIZE_PROTOTYPE)
/* Some version of Mac OSX have malloc_good_size in their libc, but not
 * actually defined in malloc/malloc.h.  We detect this and work around it by
 * prototyping.
 */
extern size_t malloc_good_size(size_t size);
#endif

/** Allocate and return a chunk of memory of size at least *<b>size</b>, using
 * the same resources we would use to malloc *<b>sizep</b>.  Set *<b>sizep</b>
 * to the number of usable bytes in the chunk of memory. */
void *
_tor_malloc_roundup(size_t *sizep DMALLOC_PARAMS)
{
#ifdef HAVE_MALLOC_GOOD_SIZE
  tor_assert(*sizep < SIZE_T_CEILING);
  *sizep = malloc_good_size(*sizep);
  return _tor_malloc(*sizep DMALLOC_FN_ARGS);
#elif 0 && defined(HAVE_MALLOC_USABLE_SIZE) && !defined(USE_DMALLOC)
  /* Never use malloc_usable_size(); it makes valgrind really unhappy,
   * and doesn't win much in terms of usable space where it exists. */
  void *result;
  tor_assert(*sizep < SIZE_T_CEILING);
  result = _tor_malloc(*sizep DMALLOC_FN_ARGS);
  *sizep = malloc_usable_size(result);
  return result;
#else
  return _tor_malloc(*sizep DMALLOC_FN_ARGS);
#endif
}

/** Call the platform malloc info function, and dump the results to the log at
 * level <b>severity</b>.  If no such function exists, do nothing. */
void
tor_log_mallinfo(int severity)
{
#ifdef HAVE_MALLINFO
  struct mallinfo mi;
  memset(&mi, 0, sizeof(mi));
  mi = mallinfo();
  tor_log(severity, LD_MM,
      "mallinfo() said: arena=%d, ordblks=%d, smblks=%d, hblks=%d, "
      "hblkhd=%d, usmblks=%d, fsmblks=%d, uordblks=%d, fordblks=%d, "
      "keepcost=%d",
      mi.arena, mi.ordblks, mi.smblks, mi.hblks,
      mi.hblkhd, mi.usmblks, mi.fsmblks, mi.uordblks, mi.fordblks,
      mi.keepcost);
#else
  (void)severity;
#endif
#ifdef USE_DMALLOC
  dmalloc_log_changed(0, /* Since the program started. */
                      1, /* Log info about non-freed pointers. */
                      0, /* Do not log info about freed pointers. */
                      0  /* Do not log individual pointers. */
                      );
#endif
}

/* =====
 * Math
 * ===== */

/**
 * Returns the natural logarithm of d base 2.  We define this wrapper here so
 * as to make it easier not to conflict with Tor's log() macro.
 */
double
tor_mathlog(double d)
{
  return log(d);
}

/** Return the long integer closest to d.  We define this wrapper here so
 * that not all users of math.h need to use the right incancations to get
 * the c99 functions. */
long
tor_lround(double d)
{
  return lround(d);
}

/** Returns floor(log2(u64)).  If u64 is 0, (incorrectly) returns 0. */
int
tor_log2(uint64_t u64)
{
  int r = 0;
  if (u64 >= (U64_LITERAL(1)<<32)) {
    u64 >>= 32;
    r = 32;
  }
  if (u64 >= (U64_LITERAL(1)<<16)) {
    u64 >>= 16;
    r += 16;
  }
  if (u64 >= (U64_LITERAL(1)<<8)) {
    u64 >>= 8;
    r += 8;
  }
  if (u64 >= (U64_LITERAL(1)<<4)) {
    u64 >>= 4;
    r += 4;
  }
  if (u64 >= (U64_LITERAL(1)<<2)) {
    u64 >>= 2;
    r += 2;
  }
  if (u64 >= (U64_LITERAL(1)<<1)) {
    u64 >>= 1;
    r += 1;
  }
  return r;
}

/** Return the power of 2 closest to <b>u64</b>. */
uint64_t
round_to_power_of_2(uint64_t u64)
{
  int lg2 = tor_log2(u64);
  uint64_t low = U64_LITERAL(1) << lg2, high = U64_LITERAL(1) << (lg2+1);
  if (high - u64 < u64 - low)
    return high;
  else
    return low;
}

/** Return the lowest x such that x is at least <b>number</b>, and x modulo
 * <b>divisor</b> == 0. */
unsigned
round_to_next_multiple_of(unsigned number, unsigned divisor)
{
  number += divisor - 1;
  number -= number % divisor;
  return number;
}

/** Return the lowest x such that x is at least <b>number</b>, and x modulo
 * <b>divisor</b> == 0. */
uint32_t
round_uint32_to_next_multiple_of(uint32_t number, uint32_t divisor)
{
  number += divisor - 1;
  number -= number % divisor;
  return number;
}

/** Return the lowest x such that x is at least <b>number</b>, and x modulo
 * <b>divisor</b> == 0. */
uint64_t
round_uint64_to_next_multiple_of(uint64_t number, uint64_t divisor)
{
  number += divisor - 1;
  number -= number % divisor;
  return number;
}

/* =====
 * String manipulation
 * ===== */

/** Remove from the string <b>s</b> every character which appears in
 * <b>strip</b>. */
void
tor_strstrip(char *s, const char *strip)
{
  char *read = s;
  while (*read) {
    if (strchr(strip, *read)) {
      ++read;
    } else {
      *s++ = *read++;
    }
  }
  *s = '\0';
}

/** Return a pointer to a NUL-terminated hexadecimal string encoding
 * the first <b>fromlen</b> bytes of <b>from</b>. (fromlen must be \<= 32.) The
 * result does not need to be deallocated, but repeated calls to
 * hex_str will trash old results.
 */
const char *
hex_str(const char *from, size_t fromlen)
{
  static char buf[65];
  if (fromlen>(sizeof(buf)-1)/2)
    fromlen = (sizeof(buf)-1)/2;
  base16_encode(buf,sizeof(buf),from,fromlen);
  return buf;
}

/** Convert all alphabetic characters in the nul-terminated string <b>s</b> to
 * lowercase. */
void
tor_strlower(char *s)
{
  while (*s) {
    *s = TOR_TOLOWER(*s);
    ++s;
  }
}

/** Convert all alphabetic characters in the nul-terminated string <b>s</b> to
 * lowercase. */
void
tor_strupper(char *s)
{
  while (*s) {
    *s = TOR_TOUPPER(*s);
    ++s;
  }
}

/** Return 1 if every character in <b>s</b> is printable, else return 0.
 */
int
tor_strisprint(const char *s)
{
  while (*s) {
    if (!TOR_ISPRINT(*s))
      return 0;
    s++;
  }
  return 1;
}

/** Return 1 if no character in <b>s</b> is uppercase, else return 0.
 */
int
tor_strisnonupper(const char *s)
{
  while (*s) {
    if (TOR_ISUPPER(*s))
      return 0;
    s++;
  }
  return 1;
}

/** Compares the first strlen(s2) characters of s1 with s2.  Returns as for
 * strcmp.
 */
int
strcmpstart(const char *s1, const char *s2)
{
  size_t n = strlen(s2);
  return strncmp(s1, s2, n);
}

/** Compare the s1_len-byte string <b>s1</b> with <b>s2</b>,
 * without depending on a terminating nul in s1.  Sorting order is first by
 * length, then lexically; return values are as for strcmp.
 */
int
strcmp_len(const char *s1, const char *s2, size_t s1_len)
{
  size_t s2_len = strlen(s2);
  if (s1_len < s2_len)
    return -1;
  if (s1_len > s2_len)
    return 1;
  return memcmp(s1, s2, s2_len);
}

/** Compares the first strlen(s2) characters of s1 with s2.  Returns as for
 * strcasecmp.
 */
int
strcasecmpstart(const char *s1, const char *s2)
{
  size_t n = strlen(s2);
  return strncasecmp(s1, s2, n);
}

/** Compares the last strlen(s2) characters of s1 with s2.  Returns as for
 * strcmp.
 */
int
strcmpend(const char *s1, const char *s2)
{
  size_t n1 = strlen(s1), n2 = strlen(s2);
  if (n2>n1)
    return strcmp(s1,s2);
  else
    return strncmp(s1+(n1-n2), s2, n2);
}

/** Compares the last strlen(s2) characters of s1 with s2.  Returns as for
 * strcasecmp.
 */
int
strcasecmpend(const char *s1, const char *s2)
{
  size_t n1 = strlen(s1), n2 = strlen(s2);
  if (n2>n1) /* then they can't be the same; figure out which is bigger */
    return strcasecmp(s1,s2);
  else
    return strncasecmp(s1+(n1-n2), s2, n2);
}

/** Compare the value of the string <b>prefix</b> with the start of the
 * <b>memlen</b>-byte memory chunk at <b>mem</b>.  Return as for strcmp.
 *
 * [As memcmp(mem, prefix, strlen(prefix)) but returns -1 if memlen is less
 * than strlen(prefix).]
 */
int
memcmpstart(const void *mem, size_t memlen,
                const char *prefix)
{
  size_t plen = strlen(prefix);
  if (memlen < plen)
    return -1;
  return memcmp(mem, prefix, plen);
}

/** Return a pointer to the first char of s that is not whitespace and
 * not a comment, or to the terminating NUL if no such character exists.
 */
const char *
eat_whitespace(const char *s)
{
  tor_assert(s);

  while (1) {
    switch (*s) {
    case '\0':
    default:
      return s;
    case ' ':
    case '\t':
    case '\n':
    case '\r':
      ++s;
      break;
    case '#':
      ++s;
      while (*s && *s != '\n')
        ++s;
    }
  }
}

/** Return a pointer to the first char of s that is not whitespace and
 * not a comment, or to the terminating NUL if no such character exists.
 */
const char *
eat_whitespace_eos(const char *s, const char *eos)
{
  tor_assert(s);
  tor_assert(eos && s <= eos);

  while (s < eos) {
    switch (*s) {
    case '\0':
    default:
      return s;
    case ' ':
    case '\t':
    case '\n':
    case '\r':
      ++s;
      break;
    case '#':
      ++s;
      while (s < eos && *s && *s != '\n')
        ++s;
    }
  }
  return s;
}

/** Return a pointer to the first char of s that is not a space or a tab
 * or a \\r, or to the terminating NUL if no such character exists. */
const char *
eat_whitespace_no_nl(const char *s)
{
  while (*s == ' ' || *s == '\t' || *s == '\r')
    ++s;
  return s;
}

/** As eat_whitespace_no_nl, but stop at <b>eos</b> whether we have
 * found a non-whitespace character or not. */
const char *
eat_whitespace_eos_no_nl(const char *s, const char *eos)
{
  while (s < eos && (*s == ' ' || *s == '\t' || *s == '\r'))
    ++s;
  return s;
}

/** Return a pointer to the first char of s that is whitespace or <b>#</b>,
 * or to the terminating NUL if no such character exists.
 */
const char *
find_whitespace(const char *s)
{
  /* tor_assert(s); */
  while (1) {
    switch (*s)
    {
    case '\0':
    case '#':
    case ' ':
    case '\r':
    case '\n':
    case '\t':
      return s;
    default:
      ++s;
    }
  }
}

/** As find_whitespace, but stop at <b>eos</b> whether we have found a
 * whitespace or not. */
const char *
find_whitespace_eos(const char *s, const char *eos)
{
  /* tor_assert(s); */
  while (s < eos) {
    switch (*s)
    {
    case '\0':
    case '#':
    case ' ':
    case '\r':
    case '\n':
    case '\t':
      return s;
    default:
      ++s;
    }
  }
  return s;
}

/** Return the first occurrence of <b>needle</b> in <b>haystack</b> that
 * occurs at the start of a line (that is, at the beginning of <b>haystack</b>
 * or immediately after a newline).  Return NULL if no such string is found.
 */
const char *
find_str_at_start_of_line(const char *haystack, const char *needle)
{
  size_t needle_len = strlen(needle);

  do {
    if (!strncmp(haystack, needle, needle_len))
      return haystack;

    haystack = strchr(haystack, '\n');
    if (!haystack)
      return NULL;
    else
      ++haystack;
  } while (*haystack);

  return NULL;
}

/** Return true iff the 'len' bytes at 'mem' are all zero. */
int
tor_mem_is_zero(const char *mem, size_t len)
{
  static const char ZERO[] = {
    0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0, 0,0,0,0,
  };
  while (len >= sizeof(ZERO)) {
    if (memcmp(mem, ZERO, sizeof(ZERO)))
      return 0;
    len -= sizeof(ZERO);
    mem += sizeof(ZERO);
  }
  /* Deal with leftover bytes. */
  if (len)
    return ! memcmp(mem, ZERO, len);

  return 1;
}

/** Return true iff the DIGEST_LEN bytes in digest are all zero. */
int
tor_digest_is_zero(const char *digest)
{
  return tor_mem_is_zero(digest, DIGEST_LEN);
}

/** Return true iff the DIGEST256_LEN bytes in digest are all zero. */
int
tor_digest256_is_zero(const char *digest)
{
  return tor_mem_is_zero(digest, DIGEST256_LEN);
}

/* Helper: common code to check whether the result of a strtol or strtoul or
 * strtoll is correct. */
#define CHECK_STRTOX_RESULT()                           \
  /* Was at least one character converted? */           \
  if (endptr == s)                                      \
    goto err;                                           \
  /* Were there unexpected unconverted characters? */   \
  if (!next && *endptr)                                 \
    goto err;                                           \
  /* Is r within limits? */                             \
  if (r < min || r > max)                               \
    goto err;                                           \
  if (ok) *ok = 1;                                      \
  if (next) *next = endptr;                             \
  return r;                                             \
 err:                                                   \
  if (ok) *ok = 0;                                      \
  if (next) *next = endptr;                             \
  return 0

/** Extract a long from the start of s, in the given numeric base.  If
 * there is unconverted data and next is provided, set *next to the
 * first unconverted character.  An error has occurred if no characters
 * are converted; or if there are unconverted characters and next is NULL; or
 * if the parsed value is not between min and max.  When no error occurs,
 * return the parsed value and set *ok (if provided) to 1.  When an error
 * occurs, return 0 and set *ok (if provided) to 0.
 */
long
tor_parse_long(const char *s, int base, long min, long max,
               int *ok, char **next)
{
  char *endptr;
  long r;

  r = strtol(s, &endptr, base);
  CHECK_STRTOX_RESULT();
}

/** As tor_parse_long(), but return an unsigned long. */
unsigned long
tor_parse_ulong(const char *s, int base, unsigned long min,
                unsigned long max, int *ok, char **next)
{
  char *endptr;
  unsigned long r;

  r = strtoul(s, &endptr, base);
  CHECK_STRTOX_RESULT();
}

/** As tor_parse_long(), but return a double. */
double
tor_parse_double(const char *s, double min, double max, int *ok, char **next)
{
  char *endptr;
  double r;

  r = strtod(s, &endptr);
  CHECK_STRTOX_RESULT();
}

/** As tor_parse_long, but return a uint64_t.  Only base 10 is guaranteed to
 * work for now. */
uint64_t
tor_parse_uint64(const char *s, int base, uint64_t min,
                 uint64_t max, int *ok, char **next)
{
  char *endptr;
  uint64_t r;

#ifdef HAVE_STRTOULL
  r = (uint64_t)strtoull(s, &endptr, base);
#elif defined(MS_WINDOWS)
#if defined(_MSC_VER) && _MSC_VER < 1300
  tor_assert(base <= 10);
  r = (uint64_t)_atoi64(s);
  endptr = (char*)s;
  while (TOR_ISSPACE(*endptr)) endptr++;
  while (TOR_ISDIGIT(*endptr)) endptr++;
#else
  r = (uint64_t)_strtoui64(s, &endptr, base);
#endif
#elif SIZEOF_LONG == 8
  r = (uint64_t)strtoul(s, &endptr, base);
#else
#error "I don't know how to parse 64-bit numbers."
#endif

  CHECK_STRTOX_RESULT();
}

/** Encode the <b>srclen</b> bytes at <b>src</b> in a NUL-terminated,
 * uppercase hexadecimal string; store it in the <b>destlen</b>-byte buffer
 * <b>dest</b>.
 */
void
base16_encode(char *dest, size_t destlen, const char *src, size_t srclen)
{
  const char *end;
  char *cp;

  tor_assert(destlen >= srclen*2+1);
  tor_assert(destlen < SIZE_T_CEILING);

  cp = dest;
  end = src+srclen;
  while (src<end) {
    *cp++ = "0123456789ABCDEF"[ (*(const uint8_t*)src) >> 4 ];
    *cp++ = "0123456789ABCDEF"[ (*(const uint8_t*)src) & 0xf ];
    ++src;
  }
  *cp = '\0';
}

/** Helper: given a hex digit, return its value, or -1 if it isn't hex. */
static INLINE int
_hex_decode_digit(char c)
{
  switch (c) {
    case '0': return 0;
    case '1': return 1;
    case '2': return 2;
    case '3': return 3;
    case '4': return 4;
    case '5': return 5;
    case '6': return 6;
    case '7': return 7;
    case '8': return 8;
    case '9': return 9;
    case 'A': case 'a': return 10;
    case 'B': case 'b': return 11;
    case 'C': case 'c': return 12;
    case 'D': case 'd': return 13;
    case 'E': case 'e': return 14;
    case 'F': case 'f': return 15;
    default:
      return -1;
  }
}

/** Helper: given a hex digit, return its value, or -1 if it isn't hex. */
int
hex_decode_digit(char c)
{
  return _hex_decode_digit(c);
}

/** Given a hexadecimal string of <b>srclen</b> bytes in <b>src</b>, decode it
 * and store the result in the <b>destlen</b>-byte buffer at <b>dest</b>.
 * Return 0 on success, -1 on failure. */
int
base16_decode(char *dest, size_t destlen, const char *src, size_t srclen)
{
  const char *end;

  int v1,v2;
  if ((srclen % 2) != 0)
    return -1;
  if (destlen < srclen/2 || destlen > SIZE_T_CEILING)
    return -1;
  end = src+srclen;
  while (src<end) {
    v1 = _hex_decode_digit(*src);
    v2 = _hex_decode_digit(*(src+1));
    if (v1<0||v2<0)
      return -1;
    *(uint8_t*)dest = (v1<<4)|v2;
    ++dest;
    src+=2;
  }
  return 0;
}

/** Allocate and return a new string representing the contents of <b>s</b>,
 * surrounded by quotes and using standard C escapes.
 *
 * Generally, we use this for logging values that come in over the network to
 * keep them from tricking users, and for sending certain values to the
 * controller.
 *
 * We trust values from the resolver, OS, configuration file, and command line
 * to not be maliciously ill-formed.  We validate incoming routerdescs and
 * SOCKS requests and addresses from BEGIN cells as they're parsed;
 * afterwards, we trust them as non-malicious.
 */
char *
esc_for_log(const char *s)
{
  const char *cp;
  char *result, *outp;
  size_t len = 3;
  if (!s) {
    return tor_strdup("(null)");
  }

  for (cp = s; *cp; ++cp) {
    switch (*cp) {
      case '\\':
      case '\"':
      case '\'':
      case '\r':
      case '\n':
      case '\t':
        len += 2;
        break;
      default:
        if (TOR_ISPRINT(*cp) && ((uint8_t)*cp)<127)
          ++len;
        else
          len += 4;
        break;
    }
  }

  result = outp = tor_malloc(len);
  *outp++ = '\"';
  for (cp = s; *cp; ++cp) {
    switch (*cp) {
      case '\\':
      case '\"':
      case '\'':
        *outp++ = '\\';
        *outp++ = *cp;
        break;
      case '\n':
        *outp++ = '\\';
        *outp++ = 'n';
        break;
      case '\t':
        *outp++ = '\\';
        *outp++ = 't';
        break;
      case '\r':
        *outp++ = '\\';
        *outp++ = 'r';
        break;
      default:
        if (TOR_ISPRINT(*cp) && ((uint8_t)*cp)<127) {
          *outp++ = *cp;
        } else {
          tor_snprintf(outp, 5, "\\%03o", (int)(uint8_t) *cp);
          outp += 4;
        }
        break;
    }
  }

  *outp++ = '\"';
  *outp++ = 0;

  return result;
}

/** Allocate and return a new string representing the contents of <b>s</b>,
 * surrounded by quotes and using standard C escapes.
 *
 * THIS FUNCTION IS NOT REENTRANT.  Don't call it from outside the main
 * thread.  Also, each call invalidates the last-returned value, so don't
 * try log_warn(LD_GENERAL, "%s %s", escaped(a), escaped(b));
 */
const char *
escaped(const char *s)
{
  static char *_escaped_val = NULL;
  tor_free(_escaped_val);

  if (s)
    _escaped_val = esc_for_log(s);
  else
    _escaped_val = NULL;

  return _escaped_val;
}

/** Rudimentary string wrapping code: given a un-wrapped <b>string</b> (no
 * newlines!), break the string into newline-terminated lines of no more than
 * <b>width</b> characters long (not counting newline) and insert them into
 * <b>out</b> in order.  Precede the first line with prefix0, and subsequent
 * lines with prefixRest.
 */
/* This uses a stupid greedy wrapping algorithm right now:
 *  - For each line:
 *    - Try to fit as much stuff as possible, but break on a space.
 *    - If the first "word" of the line will extend beyond the allowable
 *      width, break the word at the end of the width.
 */
void
wrap_string(smartlist_t *out, const char *string, size_t width,
            const char *prefix0, const char *prefixRest)
{
  size_t p0Len, pRestLen, pCurLen;
  const char *eos, *prefixCur;
  tor_assert(out);
  tor_assert(string);
  tor_assert(width);
  if (!prefix0)
    prefix0 = "";
  if (!prefixRest)
    prefixRest = "";

  p0Len = strlen(prefix0);
  pRestLen = strlen(prefixRest);
  tor_assert(width > p0Len && width > pRestLen);
  eos = strchr(string, '\0');
  tor_assert(eos);
  pCurLen = p0Len;
  prefixCur = prefix0;

  while ((eos-string)+pCurLen > width) {
    const char *eol = string + width - pCurLen;
    while (eol > string && *eol != ' ')
      --eol;
    /* eol is now the last space that can fit, or the start of the string. */
    if (eol > string) {
      size_t line_len = (eol-string) + pCurLen + 2;
      char *line = tor_malloc(line_len);
      memcpy(line, prefixCur, pCurLen);
      memcpy(line+pCurLen, string, eol-string);
      line[line_len-2] = '\n';
      line[line_len-1] = '\0';
      smartlist_add(out, line);
      string = eol + 1;
    } else {
      size_t line_len = width + 2;
      char *line = tor_malloc(line_len);
      memcpy(line, prefixCur, pCurLen);
      memcpy(line+pCurLen, string, width - pCurLen);
      line[line_len-2] = '\n';
      line[line_len-1] = '\0';
      smartlist_add(out, line);
      string += width-pCurLen;
    }
    prefixCur = prefixRest;
    pCurLen = pRestLen;
  }

  if (string < eos) {
    size_t line_len = (eos-string) + pCurLen + 2;
    char *line = tor_malloc(line_len);
    memcpy(line, prefixCur, pCurLen);
    memcpy(line+pCurLen, string, eos-string);
    line[line_len-2] = '\n';
    line[line_len-1] = '\0';
    smartlist_add(out, line);
  }
}

/* =====
 * Time
 * ===== */

/**
 * Converts struct timeval to a double value.
 * Preserves microsecond precision, but just barely.
 * Error is approx +/- 0.1 usec when dealing with epoch values.
 */
double
tv_to_double(const struct timeval *tv)
{
  double conv = tv->tv_sec;
  conv += tv->tv_usec/1000000.0;
  return conv;
}

/**
 * Converts timeval to milliseconds.
 */
int64_t
tv_to_msec(const struct timeval *tv)
{
  int64_t conv = ((int64_t)tv->tv_sec)*1000L;
  /* Round ghetto-style */
  conv += ((int64_t)tv->tv_usec+500)/1000L;
  return conv;
}

/**
 * Converts timeval to microseconds.
 */
int64_t
tv_to_usec(const struct timeval *tv)
{
  int64_t conv = ((int64_t)tv->tv_sec)*1000000L;
  conv += tv->tv_usec;
  return conv;
}

/** Return the number of microseconds elapsed between *start and *end.
 */
long
tv_udiff(const struct timeval *start, const struct timeval *end)
{
  long udiff;
  long secdiff = end->tv_sec - start->tv_sec;

  if (labs(secdiff+1) > LONG_MAX/1000000) {
    log_warn(LD_GENERAL, "comparing times on microsecond detail too far "
             "apart: %ld seconds", secdiff);
    return LONG_MAX;
  }

  udiff = secdiff*1000000L + (end->tv_usec - start->tv_usec);
  return udiff;
}

/** Return the number of milliseconds elapsed between *start and *end.
 */
long
tv_mdiff(const struct timeval *start, const struct timeval *end)
{
  long mdiff;
  long secdiff = end->tv_sec - start->tv_sec;

  if (labs(secdiff+1) > LONG_MAX/1000) {
    log_warn(LD_GENERAL, "comparing times on millisecond detail too far "
             "apart: %ld seconds", secdiff);
    return LONG_MAX;
  }

  /* Subtract and round */
  mdiff = secdiff*1000L +
      ((long)end->tv_usec - (long)start->tv_usec + 500L) / 1000L;
  return mdiff;
}

/** Yield true iff <b>y</b> is a leap-year. */
#define IS_LEAPYEAR(y) (!(y % 4) && ((y % 100) || !(y % 400)))
/** Helper: Return the number of leap-days between Jan 1, y1 and Jan 1, y2. */
static int
n_leapdays(int y1, int y2)
{
  --y1;
  --y2;
  return (y2/4 - y1/4) - (y2/100 - y1/100) + (y2/400 - y1/400);
}
/** Number of days per month in non-leap year; used by tor_timegm. */
static const int days_per_month[] =
  { 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31};

/** Return a time_t given a struct tm.  The result is given in GMT, and
 * does not account for leap seconds.
 */
time_t
tor_timegm(struct tm *tm)
{
  /* This is a pretty ironclad timegm implementation, snarfed from Python2.2.
   * It's way more brute-force than fiddling with tzset().
   */
  time_t year, days, hours, minutes, seconds;
  int i;
  year = tm->tm_year + 1900;
  if (year < 1970 || tm->tm_mon < 0 || tm->tm_mon > 11) {
    log_warn(LD_BUG, "Out-of-range argument to tor_timegm");
    return -1;
  }
  tor_assert(year < INT_MAX);
  days = 365 * (year-1970) + n_leapdays(1970,(int)year);
  for (i = 0; i < tm->tm_mon; ++i)
    days += days_per_month[i];
  if (tm->tm_mon > 1 && IS_LEAPYEAR(year))
    ++days;
  days += tm->tm_mday - 1;
  hours = days*24 + tm->tm_hour;

  minutes = hours*60 + tm->tm_min;
  seconds = minutes*60 + tm->tm_sec;
  return seconds;
}

/* strftime is locale-specific, so we need to replace those parts */

/** A c-locale array of 3-letter names of weekdays, starting with Sun. */
static const char *WEEKDAY_NAMES[] =
  { "Sun", "Mon", "Tue", "Wed", "Thu", "Fri", "Sat" };
/** A c-locale array of 3-letter names of months, starting with Jan. */
static const char *MONTH_NAMES[] =
  { "Jan", "Feb", "Mar", "Apr", "May", "Jun",
    "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" };

/** Set <b>buf</b> to the RFC1123 encoding of the GMT value of <b>t</b>.
 * The buffer must be at least RFC1123_TIME_LEN+1 bytes long.
 *
 * (RFC1123 format is Fri, 29 Sep 2006 15:54:20 GMT)
 */
void
format_rfc1123_time(char *buf, time_t t)
{
  struct tm tm;

  tor_gmtime_r(&t, &tm);

  strftime(buf, RFC1123_TIME_LEN+1, "___, %d ___ %Y %H:%M:%S GMT", &tm);
  tor_assert(tm.tm_wday >= 0);
  tor_assert(tm.tm_wday <= 6);
  memcpy(buf, WEEKDAY_NAMES[tm.tm_wday], 3);
  tor_assert(tm.tm_wday >= 0);
  tor_assert(tm.tm_mon <= 11);
  memcpy(buf+8, MONTH_NAMES[tm.tm_mon], 3);
}

/** Parse the RFC1123 encoding of some time (in GMT) from <b>buf</b>,
 * and store the result in *<b>t</b>.
 *
 * Return 0 on success, -1 on failure.
*/
int
parse_rfc1123_time(const char *buf, time_t *t)
{
  struct tm tm;
  char month[4];
  char weekday[4];
  int i, m;
  unsigned tm_mday, tm_year, tm_hour, tm_min, tm_sec;

  if (strlen(buf) != RFC1123_TIME_LEN)
    return -1;
  memset(&tm, 0, sizeof(tm));
  if (tor_sscanf(buf, "%3s, %2u %3s %u %2u:%2u:%2u GMT", weekday,
             &tm_mday, month, &tm_year, &tm_hour,
             &tm_min, &tm_sec) < 7) {
    char *esc = esc_for_log(buf);
    log_warn(LD_GENERAL, "Got invalid RFC1123 time %s", esc);
    tor_free(esc);
    return -1;
  }
  if (tm_mday > 31 || tm_hour > 23 || tm_min > 59 || tm_sec > 61) {
    char *esc = esc_for_log(buf);
    log_warn(LD_GENERAL, "Got invalid RFC1123 time %s", esc);
    tor_free(esc);
    return -1;
  }
  tm.tm_mday = (int)tm_mday;
  tm.tm_year = (int)tm_year;
  tm.tm_hour = (int)tm_hour;
  tm.tm_min = (int)tm_min;
  tm.tm_sec = (int)tm_sec;

  m = -1;
  for (i = 0; i < 12; ++i) {
    if (!strcmp(month, MONTH_NAMES[i])) {
      m = i;
      break;
    }
  }
  if (m<0) {
    char *esc = esc_for_log(buf);
    log_warn(LD_GENERAL, "Got invalid RFC1123 time %s: No such month", esc);
    tor_free(esc);
    return -1;
  }
  tm.tm_mon = m;

  if (tm.tm_year < 1970) {
    char *esc = esc_for_log(buf);
    log_warn(LD_GENERAL,
             "Got invalid RFC1123 time %s. (Before 1970)", esc);
    tor_free(esc);
    return -1;
  }
  tm.tm_year -= 1900;

  *t = tor_timegm(&tm);
  return 0;
}

/** Set <b>buf</b> to the ISO8601 encoding of the local value of <b>t</b>.
 * The buffer must be at least ISO_TIME_LEN+1 bytes long.
 *
 * (ISO8601 format is 2006-10-29 10:57:20)
 */
void
format_local_iso_time(char *buf, time_t t)
{
  struct tm tm;
  strftime(buf, ISO_TIME_LEN+1, "%Y-%m-%d %H:%M:%S", tor_localtime_r(&t, &tm));
}

/** Set <b>buf</b> to the ISO8601 encoding of the GMT value of <b>t</b>.
 * The buffer must be at least ISO_TIME_LEN+1 bytes long.
 */
void
format_iso_time(char *buf, time_t t)
{
  struct tm tm;
  strftime(buf, ISO_TIME_LEN+1, "%Y-%m-%d %H:%M:%S", tor_gmtime_r(&t, &tm));
}

/** Given an ISO-formatted UTC time value (after the epoch) in <b>cp</b>,
 * parse it and store its value in *<b>t</b>.  Return 0 on success, -1 on
 * failure.  Ignore extraneous stuff in <b>cp</b> separated by whitespace from
 * the end of the time string. */
int
parse_iso_time(const char *cp, time_t *t)
{
  struct tm st_tm;
  unsigned int year=0, month=0, day=0, hour=100, minute=100, second=100;
  if (tor_sscanf(cp, "%u-%2u-%2u %2u:%2u:%2u", &year, &month,
                &day, &hour, &minute, &second) < 6) {
    char *esc = esc_for_log(cp);
    log_warn(LD_GENERAL, "ISO time %s was unparseable", esc);
    tor_free(esc);
    return -1;
  }
  if (year < 1970 || month < 1 || month > 12 || day < 1 || day > 31 ||
          hour > 23 || minute > 59 || second > 61) {
    char *esc = esc_for_log(cp);
    log_warn(LD_GENERAL, "ISO time %s was nonsensical", esc);
    tor_free(esc);
    return -1;
  }
  st_tm.tm_year = year-1900;
  st_tm.tm_mon = month-1;
  st_tm.tm_mday = day;
  st_tm.tm_hour = hour;
  st_tm.tm_min = minute;
  st_tm.tm_sec = second;

  if (st_tm.tm_year < 70) {
    char *esc = esc_for_log(cp);
    log_warn(LD_GENERAL, "Got invalid ISO time %s. (Before 1970)", esc);
    tor_free(esc);
    return -1;
  }
  *t = tor_timegm(&st_tm);
  return 0;
}

/** Given a <b>date</b> in one of the three formats allowed by HTTP (ugh),
 * parse it into <b>tm</b>.  Return 0 on success, negative on failure. */
int
parse_http_time(const char *date, struct tm *tm)
{
  const char *cp;
  char month[4];
  char wkday[4];
  int i;
  unsigned tm_mday, tm_year, tm_hour, tm_min, tm_sec;

  tor_assert(tm);
  memset(tm, 0, sizeof(*tm));

  /* First, try RFC1123 or RFC850 format: skip the weekday.  */
  if ((cp = strchr(date, ','))) {
    ++cp;
    if (tor_sscanf(date, "%2u %3s %4u %2u:%2u:%2u GMT",
               &tm_mday, month, &tm_year,
               &tm_hour, &tm_min, &tm_sec) == 6) {
      /* rfc1123-date */
      tm_year -= 1900;
    } else if (tor_sscanf(date, "%2u-%3s-%2u %2u:%2u:%2u GMT",
                      &tm_mday, month, &tm_year,
                      &tm_hour, &tm_min, &tm_sec) == 6) {
      /* rfc850-date */
    } else {
      return -1;
    }
  } else {
    /* No comma; possibly asctime() format. */
    if (tor_sscanf(date, "%3s %3s %2u %2u:%2u:%2u %4u",
               wkday, month, &tm_mday,
               &tm_hour, &tm_min, &tm_sec, &tm_year) == 7) {
      tm_year -= 1900;
    } else {
      return -1;
    }
  }
  tm->tm_mday = (int)tm_mday;
  tm->tm_year = (int)tm_year;
  tm->tm_hour = (int)tm_hour;
  tm->tm_min = (int)tm_min;
  tm->tm_sec = (int)tm_sec;

  month[3] = '\0';
  /* Okay, now decode the month. */
  for (i = 0; i < 12; ++i) {
    if (!strcasecmp(MONTH_NAMES[i], month)) {
      tm->tm_mon = i+1;
    }
  }

  if (tm->tm_year < 0 ||
      tm->tm_mon < 1  || tm->tm_mon > 12 ||
      tm->tm_mday < 0 || tm->tm_mday > 31 ||
      tm->tm_hour < 0 || tm->tm_hour > 23 ||
      tm->tm_min < 0  || tm->tm_min > 59 ||
      tm->tm_sec < 0  || tm->tm_sec > 61)
    return -1; /* Out of range, or bad month. */

  return 0;
}

/** Given an <b>interval</b> in seconds, try to write it to the
 * <b>out_len</b>-byte buffer in <b>out</b> in a human-readable form.
 * Return 0 on success, -1 on failure.
 */
int
format_time_interval(char *out, size_t out_len, long interval)
{
  /* We only report seconds if there's no hours. */
  long sec = 0, min = 0, hour = 0, day = 0;
  if (interval < 0)
    interval = -interval;

  if (interval >= 86400) {
    day = interval / 86400;
    interval %= 86400;
  }
  if (interval >= 3600) {
    hour = interval / 3600;
    interval %= 3600;
  }
  if (interval >= 60) {
    min = interval / 60;
    interval %= 60;
  }
  sec = interval;

  if (day) {
    return tor_snprintf(out, out_len, "%ld days, %ld hours, %ld minutes",
                        day, hour, min);
  } else if (hour) {
    return tor_snprintf(out, out_len, "%ld hours, %ld minutes", hour, min);
  } else if (min) {
    return tor_snprintf(out, out_len, "%ld minutes, %ld seconds", min, sec);
  } else {
    return tor_snprintf(out, out_len, "%ld seconds", sec);
  }
}

/* =====
 * Cached time
 * ===== */

#ifndef TIME_IS_FAST
/** Cached estimate of the current time.  Updated around once per second;
 * may be a few seconds off if we are really busy.  This is a hack to avoid
 * calling time(NULL) (which not everybody has optimized) on critical paths.
 */
static time_t cached_approx_time = 0;

/** Return a cached estimate of the current time from when
 * update_approx_time() was last called.  This is a hack to avoid calling
 * time(NULL) on critical paths: please do not even think of calling it
 * anywhere else. */
time_t
approx_time(void)
{
  return cached_approx_time;
}

/** Update the cached estimate of the current time.  This function SHOULD be
 * called once per second, and MUST be called before the first call to
 * get_approx_time. */
void
update_approx_time(time_t now)
{
  cached_approx_time = now;
}
#endif

/* =====
 * Rate limiting
 * ===== */

/** If the rate-limiter <b>lim</b> is ready at <b>now</b>, return the number
 * of calls to rate_limit_is_ready (including this one!) since the last time
 * rate_limit_is_ready returned nonzero.  Otherwise return 0. */
static int
rate_limit_is_ready(ratelim_t *lim, time_t now)
{
  if (lim->rate + lim->last_allowed <= now) {
    int res = lim->n_calls_since_last_time + 1;
    lim->last_allowed = now;
    lim->n_calls_since_last_time = 0;
    return res;
  } else {
    ++lim->n_calls_since_last_time;
    return 0;
  }
}

/** If the rate-limiter <b>lim</b> is ready at <b>now</b>, return a newly
 * allocated string indicating how many messages were suppressed, suitable to
 * append to a log message.  Otherwise return NULL. */
char *
rate_limit_log(ratelim_t *lim, time_t now)
{
  int n;
  if ((n = rate_limit_is_ready(lim, now))) {
    if (n == 1) {
      return tor_strdup("");
    } else {
      char *cp=NULL;
      tor_asprintf(&cp,
                   " [%d similar message(s) suppressed in last %d seconds]",
                   n-1, lim->rate);
      return cp;
    }
  } else {
    return NULL;
  }
}

/* =====
 * File helpers
 * ===== */

/** Write <b>count</b> bytes from <b>buf</b> to <b>fd</b>.  <b>isSocket</b>
 * must be 1 if fd was returned by socket() or accept(), and 0 if fd
 * was returned by open().  Return the number of bytes written, or -1
 * on error.  Only use if fd is a blocking fd.  */
ssize_t
write_all(int fd, const char *buf, size_t count, int isSocket)
{
  size_t written = 0;
  ssize_t result;
  tor_assert(count < SSIZE_T_MAX);

  while (written != count) {
    if (isSocket)
      result = tor_socket_send(fd, buf+written, count-written, 0);
    else
      result = write(fd, buf+written, count-written);
    if (result<0)
      return -1;
    written += result;
  }
  return (ssize_t)count;
}

/** Read from <b>fd</b> to <b>buf</b>, until we get <b>count</b> bytes
 * or reach the end of the file. <b>isSocket</b> must be 1 if fd
 * was returned by socket() or accept(), and 0 if fd was returned by
 * open().  Return the number of bytes read, or -1 on error. Only use
 * if fd is a blocking fd. */
ssize_t
read_all(int fd, char *buf, size_t count, int isSocket)
{
  size_t numread = 0;
  ssize_t result;

  if (count > SIZE_T_CEILING || count > SSIZE_T_MAX)
    return -1;

  while (numread != count) {
    if (isSocket)
      result = tor_socket_recv(fd, buf+numread, count-numread, 0);
    else
      result = read(fd, buf+numread, count-numread);
    if (result<0)
      return -1;
    else if (result == 0)
      break;
    numread += result;
  }
  return (ssize_t)numread;
}

/*
 *    Filesystem operations.
 */

/** Clean up <b>name</b> so that we can use it in a call to "stat".  On Unix,
 * we do nothing.  On Windows, we remove a trailing slash, unless the path is
 * the root of a disk. */
static void
clean_name_for_stat(char *name)
{
#ifdef MS_WINDOWS
  size_t len = strlen(name);
  if (!len)
    return;
  if (name[len-1]=='\\' || name[len-1]=='/') {
    if (len == 1 || (len==3 && name[1]==':'))
      return;
    name[len-1]='\0';
  }
#else
  (void)name;
#endif
}

/** Return FN_ERROR if filename can't be read, FN_NOENT if it doesn't
 * exist, FN_FILE if it is a regular file, or FN_DIR if it's a
 * directory.  On FN_ERROR, sets errno. */
file_status_t
file_status(const char *fname)
{
  struct stat st;
  char *f;
  int r;
  f = tor_strdup(fname);
  clean_name_for_stat(f);
  r = stat(f, &st);
  tor_free(f);
  if (r) {
    if (errno == ENOENT) {
      return FN_NOENT;
    }
    return FN_ERROR;
  }
  if (st.st_mode & S_IFDIR)
    return FN_DIR;
  else if (st.st_mode & S_IFREG)
    return FN_FILE;
  else
    return FN_ERROR;
}

/** Check whether dirname exists and is private.  If yes return 0.  If
 * it does not exist, and check==CPD_CREATE is set, try to create it
 * and return 0 on success. If it does not exist, and
 * check==CPD_CHECK, and we think we can create it, return 0.  Else
 * return -1. */
int
check_private_dir(const char *dirname, cpd_check_t check)
{
  int r;
  struct stat st;
  char *f;
  tor_assert(dirname);
  f = tor_strdup(dirname);
  clean_name_for_stat(f);
  r = stat(f, &st);
  tor_free(f);
  if (r) {
    if (errno != ENOENT) {
      log_warn(LD_FS, "Directory %s cannot be read: %s", dirname,
               strerror(errno));
      return -1;
    }
    if (check == CPD_NONE) {
      log_warn(LD_FS, "Directory %s does not exist.", dirname);
      return -1;
    } else if (check == CPD_CREATE) {
      log_info(LD_GENERAL, "Creating directory %s", dirname);
#if defined (MS_WINDOWS) && !defined (WINCE)
      r = mkdir(dirname);
#else
      r = mkdir(dirname, 0700);
#endif
      if (r) {
        log_warn(LD_FS, "Error creating directory %s: %s", dirname,
            strerror(errno));
        return -1;
      }
    }
    /* XXXX In the case where check==CPD_CHECK, we should look at the
     * parent directory a little harder. */
    return 0;
  }
  if (!(st.st_mode & S_IFDIR)) {
    log_warn(LD_FS, "%s is not a directory", dirname);
    return -1;
  }
#ifndef MS_WINDOWS
  if (st.st_uid != getuid()) {
    struct passwd *pw = NULL;
    char *process_ownername = NULL;

    pw = getpwuid(getuid());
    process_ownername = pw ? tor_strdup(pw->pw_name) : tor_strdup("<unknown>");

    pw = getpwuid(st.st_uid);

    log_warn(LD_FS, "%s is not owned by this user (%s, %d) but by "
        "%s (%d). Perhaps you are running Tor as the wrong user?",
                         dirname, process_ownername, (int)getuid(),
                         pw ? pw->pw_name : "<unknown>", (int)st.st_uid);

    tor_free(process_ownername);
    return -1;
  }
  if (st.st_mode & 0077) {
    log_warn(LD_FS, "Fixing permissions on directory %s", dirname);
    if (chmod(dirname, 0700)) {
      log_warn(LD_FS, "Could not chmod directory %s: %s", dirname,
          strerror(errno));
      return -1;
    } else {
      return 0;
    }
  }
#endif
  return 0;
}

/** Create a file named <b>fname</b> with the contents <b>str</b>.  Overwrite
 * the previous <b>fname</b> if possible.  Return 0 on success, -1 on failure.
 *
 * This function replaces the old file atomically, if possible.  This
 * function, and all other functions in util.c that create files, create them
 * with mode 0600.
 */
int
write_str_to_file(const char *fname, const char *str, int bin)
{
#ifdef MS_WINDOWS
  if (!bin && strchr(str, '\r')) {
    log_warn(LD_BUG,
             "We're writing a text string that already contains a CR.");
  }
#endif
  return write_bytes_to_file(fname, str, strlen(str), bin);
}

/** Represents a file that we're writing to, with support for atomic commit:
 * we can write into a temporary file, and either remove the file on
 * failure, or replace the original file on success. */
struct open_file_t {
  char *tempname; /**< Name of the temporary file. */
  char *filename; /**< Name of the original file. */
  unsigned rename_on_close:1; /**< Are we using the temporary file or not? */
  unsigned binary:1; /**< Did we open in binary mode? */
  int fd; /**< fd for the open file. */
  FILE *stdio_file; /**< stdio wrapper for <b>fd</b>. */
};

/** Try to start writing to the file in <b>fname</b>, passing the flags
 * <b>open_flags</b> to the open() syscall, creating the file (if needed) with
 * access value <b>mode</b>.  If the O_APPEND flag is set, we append to the
 * original file.  Otherwise, we open a new temporary file in the same
 * directory, and either replace the original or remove the temporary file
 * when we're done.
 *
 * Return the fd for the newly opened file, and store working data in
 * *<b>data_out</b>.  The caller should not close the fd manually:
 * instead, call finish_writing_to_file() or abort_writing_to_file().
 * Returns -1 on failure.
 *
 * NOTE: When not appending, the flags O_CREAT and O_TRUNC are treated
 * as true and the flag O_EXCL is treated as false.
 *
 * NOTE: Ordinarily, O_APPEND means "seek to the end of the file before each
 * write()".  We don't do that.
 */
int
start_writing_to_file(const char *fname, int open_flags, int mode,
                      open_file_t **data_out)
{
  size_t tempname_len = strlen(fname)+16;
  open_file_t *new_file = tor_malloc_zero(sizeof(open_file_t));
  const char *open_name;
  int append = 0;

  tor_assert(fname);
  tor_assert(data_out);
#if (O_BINARY != 0 && O_TEXT != 0)
  tor_assert((open_flags & (O_BINARY|O_TEXT)) != 0);
#endif
  new_file->fd = -1;
  tor_assert(tempname_len > strlen(fname)); /*check for overflow*/
  new_file->filename = tor_strdup(fname);
  if (open_flags & O_APPEND) {
    open_name = fname;
    new_file->rename_on_close = 0;
    append = 1;
    open_flags &= ~O_APPEND;
  } else {
    open_name = new_file->tempname = tor_malloc(tempname_len);
    if (tor_snprintf(new_file->tempname, tempname_len, "%s.tmp", fname)<0) {
      log_warn(LD_GENERAL, "Failed to generate filename");
      goto err;
    }
    /* We always replace an existing temporary file if there is one. */
    open_flags |= O_CREAT|O_TRUNC;
    open_flags &= ~O_EXCL;
    new_file->rename_on_close = 1;
  }
  if (open_flags & O_BINARY)
    new_file->binary = 1;

  new_file->fd = tor_open_cloexec(open_name, open_flags, mode);
  if (new_file->fd < 0) {
    log_warn(LD_FS, "Couldn't open \"%s\" (%s) for writing: %s",
        open_name, fname, strerror(errno));
    goto err;
  }
  if (append) {
    if (tor_fd_seekend(new_file->fd) < 0) {
      log_warn(LD_FS, "Couldn't seek to end of file \"%s\": %s", open_name,
               strerror(errno));
      goto err;
    }
  }

  *data_out = new_file;

  return new_file->fd;

 err:
  if (new_file->fd >= 0)
    close(new_file->fd);
  *data_out = NULL;
  tor_free(new_file->filename);
  tor_free(new_file->tempname);
  tor_free(new_file);
  return -1;
}

/** Given <b>file_data</b> from start_writing_to_file(), return a stdio FILE*
 * that can be used to write to the same file.  The caller should not mix
 * stdio calls with non-stdio calls. */
FILE *
fdopen_file(open_file_t *file_data)
{
  tor_assert(file_data);
  if (file_data->stdio_file)
    return file_data->stdio_file;
  tor_assert(file_data->fd >= 0);
  if (!(file_data->stdio_file = fdopen(file_data->fd,
                                       file_data->binary?"ab":"a"))) {
    log_warn(LD_FS, "Couldn't fdopen \"%s\" [%d]: %s", file_data->filename,
             file_data->fd, strerror(errno));
  }
  return file_data->stdio_file;
}

/** Combines start_writing_to_file with fdopen_file(): arguments are as
 * for start_writing_to_file, but  */
FILE *
start_writing_to_stdio_file(const char *fname, int open_flags, int mode,
                            open_file_t **data_out)
{
  FILE *res;
  if (start_writing_to_file(fname, open_flags, mode, data_out)<0)
    return NULL;
  if (!(res = fdopen_file(*data_out))) {
    abort_writing_to_file(*data_out);
    *data_out = NULL;
  }
  return res;
}

/** Helper function: close and free the underlying file and memory in
 * <b>file_data</b>.  If we were writing into a temporary file, then delete
 * that file (if abort_write is true) or replaces the target file with
 * the temporary file (if abort_write is false). */
static int
finish_writing_to_file_impl(open_file_t *file_data, int abort_write)
{
  int r = 0;
  tor_assert(file_data && file_data->filename);
  if (file_data->stdio_file) {
    if (fclose(file_data->stdio_file)) {
      log_warn(LD_FS, "Error closing \"%s\": %s", file_data->filename,
               strerror(errno));
      abort_write = r = -1;
    }
  } else if (file_data->fd >= 0 && close(file_data->fd) < 0) {
    log_warn(LD_FS, "Error flushing \"%s\": %s", file_data->filename,
             strerror(errno));
    abort_write = r = -1;
  }

  if (file_data->rename_on_close) {
    tor_assert(file_data->tempname && file_data->filename);
    if (abort_write) {
      unlink(file_data->tempname);
    } else {
      tor_assert(strcmp(file_data->filename, file_data->tempname));
      if (replace_file(file_data->tempname, file_data->filename)) {
        log_warn(LD_FS, "Error replacing \"%s\": %s", file_data->filename,
                 strerror(errno));
        r = -1;
      }
    }
  }

  tor_free(file_data->filename);
  tor_free(file_data->tempname);
  tor_free(file_data);

  return r;
}

/** Finish writing to <b>file_data</b>: close the file handle, free memory as
 * needed, and if using a temporary file, replace the original file with
 * the temporary file. */
int
finish_writing_to_file(open_file_t *file_data)
{
  return finish_writing_to_file_impl(file_data, 0);
}

/** Finish writing to <b>file_data</b>: close the file handle, free memory as
 * needed, and if using a temporary file, delete it. */
int
abort_writing_to_file(open_file_t *file_data)
{
  return finish_writing_to_file_impl(file_data, 1);
}

/** Helper: given a set of flags as passed to open(2), open the file
 * <b>fname</b> and write all the sized_chunk_t structs in <b>chunks</b> to
 * the file.  Do so as atomically as possible e.g. by opening temp files and
 * renaming. */
static int
write_chunks_to_file_impl(const char *fname, const smartlist_t *chunks,
                          int open_flags)
{
  open_file_t *file = NULL;
  int fd;
  ssize_t result;
  fd = start_writing_to_file(fname, open_flags, 0600, &file);
  if (fd<0)
    return -1;
  SMARTLIST_FOREACH(chunks, sized_chunk_t *, chunk,
  {
    result = write_all(fd, chunk->bytes, chunk->len, 0);
    if (result < 0) {
      log_warn(LD_FS, "Error writing to \"%s\": %s", fname,
          strerror(errno));
      goto err;
    }
    tor_assert((size_t)result == chunk->len);
  });

  return finish_writing_to_file(file);
 err:
  abort_writing_to_file(file);
  return -1;
}

/** Given a smartlist of sized_chunk_t, write them atomically to a file
 * <b>fname</b>, overwriting or creating the file as necessary. */
int
write_chunks_to_file(const char *fname, const smartlist_t *chunks, int bin)
{
  int flags = OPEN_FLAGS_REPLACE|(bin?O_BINARY:O_TEXT);
  return write_chunks_to_file_impl(fname, chunks, flags);
}

/** As write_str_to_file, but does not assume a NUL-terminated
 * string. Instead, we write <b>len</b> bytes, starting at <b>str</b>. */
int
write_bytes_to_file(const char *fname, const char *str, size_t len,
                    int bin)
{
  int flags = OPEN_FLAGS_REPLACE|(bin?O_BINARY:O_TEXT);
  int r;
  sized_chunk_t c = { str, len };
  smartlist_t *chunks = smartlist_create();
  smartlist_add(chunks, &c);
  r = write_chunks_to_file_impl(fname, chunks, flags);
  smartlist_free(chunks);
  return r;
}

/** As write_bytes_to_file, but if the file already exists, append the bytes
 * to the end of the file instead of overwriting it. */
int
append_bytes_to_file(const char *fname, const char *str, size_t len,
                     int bin)
{
  int flags = OPEN_FLAGS_APPEND|(bin?O_BINARY:O_TEXT);
  int r;
  sized_chunk_t c = { str, len };
  smartlist_t *chunks = smartlist_create();
  smartlist_add(chunks, &c);
  r = write_chunks_to_file_impl(fname, chunks, flags);
  smartlist_free(chunks);
  return r;
}

/** Read the contents of <b>filename</b> into a newly allocated
 * string; return the string on success or NULL on failure.
 *
 * If <b>stat_out</b> is provided, store the result of stat()ing the
 * file into <b>stat_out</b>.
 *
 * If <b>flags</b> &amp; RFTS_BIN, open the file in binary mode.
 * If <b>flags</b> &amp; RFTS_IGNORE_MISSING, don't warn if the file
 * doesn't exist.
 */
/*
 * This function <em>may</em> return an erroneous result if the file
 * is modified while it is running, but must not crash or overflow.
 * Right now, the error case occurs when the file length grows between
 * the call to stat and the call to read_all: the resulting string will
 * be truncated.
 */
char *
read_file_to_str(const char *filename, int flags, struct stat *stat_out)
{
  int fd; /* router file */
  struct stat statbuf;
  char *string;
  ssize_t r;
  int bin = flags & RFTS_BIN;

  tor_assert(filename);

  fd = tor_open_cloexec(filename,O_RDONLY|(bin?O_BINARY:O_TEXT),0);
  if (fd<0) {
    int severity = LOG_WARN;
    int save_errno = errno;
    if (errno == ENOENT && (flags & RFTS_IGNORE_MISSING))
      severity = LOG_INFO;
    log_fn(severity, LD_FS,"Could not open \"%s\": %s ",filename,
           strerror(errno));
    errno = save_errno;
    return NULL;
  }

  if (fstat(fd, &statbuf)<0) {
    int save_errno = errno;
    close(fd);
    log_warn(LD_FS,"Could not fstat \"%s\".",filename);
    errno = save_errno;
    return NULL;
  }

  if ((uint64_t)(statbuf.st_size)+1 >= SIZE_T_CEILING)
    return NULL;

  string = tor_malloc((size_t)(statbuf.st_size+1));

  r = read_all(fd,string,(size_t)statbuf.st_size,0);
  if (r<0) {
    int save_errno = errno;
    log_warn(LD_FS,"Error reading from file \"%s\": %s", filename,
             strerror(errno));
    tor_free(string);
    close(fd);
    errno = save_errno;
    return NULL;
  }
  string[r] = '\0'; /* NUL-terminate the result. */

#ifdef MS_WINDOWS
  if (!bin && strchr(string, '\r')) {
    log_debug(LD_FS, "We didn't convert CRLF to LF as well as we hoped "
              "when reading %s. Coping.",
              filename);
    tor_strstrip(string, "\r");
    r = strlen(string);
  }
  if (!bin) {
    statbuf.st_size = (size_t) r;
  } else
#endif
    if (r != statbuf.st_size) {
      /* Unless we're using text mode on win32, we'd better have an exact
       * match for size. */
      int save_errno = errno;
      log_warn(LD_FS,"Could read only %d of %ld bytes of file \"%s\".",
               (int)r, (long)statbuf.st_size,filename);
      tor_free(string);
      close(fd);
      errno = save_errno;
      return NULL;
    }
  close(fd);
  if (stat_out) {
    memcpy(stat_out, &statbuf, sizeof(struct stat));
  }

  return string;
}

#define TOR_ISODIGIT(c) ('0' <= (c) && (c) <= '7')

/** Given a c-style double-quoted escaped string in <b>s</b>, extract and
 * decode its contents into a newly allocated string.  On success, assign this
 * string to *<b>result</b>, assign its length to <b>size_out</b> (if
 * provided), and return a pointer to the position in <b>s</b> immediately
 * after the string.  On failure, return NULL.
 */
static const char *
unescape_string(const char *s, char **result, size_t *size_out)
{
  const char *cp;
  char *out;
  if (s[0] != '\"')
    return NULL;
  cp = s+1;
  while (1) {
    switch (*cp) {
      case '\0':
      case '\n':
        return NULL;
      case '\"':
        goto end_of_loop;
      case '\\':
        if ((cp[1] == 'x' || cp[1] == 'X')
            && TOR_ISXDIGIT(cp[2]) && TOR_ISXDIGIT(cp[3])) {
          cp += 4;
        } else if (TOR_ISODIGIT(cp[1])) {
          cp += 2;
          if (TOR_ISODIGIT(*cp)) ++cp;
          if (TOR_ISODIGIT(*cp)) ++cp;
        } else if (cp[1]) {
          cp += 2;
        } else {
          return NULL;
        }
        break;
      default:
        ++cp;
        break;
    }
  }
 end_of_loop:
  out = *result = tor_malloc(cp-s + 1);
  cp = s+1;
  while (1) {
    switch (*cp)
      {
      case '\"':
        *out = '\0';
        if (size_out) *size_out = out - *result;
        return cp+1;
      case '\0':
        tor_fragile_assert();
        tor_free(*result);
        return NULL;
      case '\\':
        switch (cp[1])
          {
          case 'n': *out++ = '\n'; cp += 2; break;
          case 'r': *out++ = '\r'; cp += 2; break;
          case 't': *out++ = '\t'; cp += 2; break;
          case 'x': case 'X':
            *out++ = ((hex_decode_digit(cp[2])<<4) +
                      hex_decode_digit(cp[3]));
            cp += 4;
            break;
          case '0': case '1': case '2': case '3': case '4': case '5':
          case '6': case '7':
            {
              int n = cp[1]-'0';
              cp += 2;
              if (TOR_ISODIGIT(*cp)) { n = n*8 + *cp-'0'; cp++; }
              if (TOR_ISODIGIT(*cp)) { n = n*8 + *cp-'0'; cp++; }
              if (n > 255) { tor_free(*result); return NULL; }
              *out++ = (char)n;
            }
            break;
          case '\'':
          case '\"':
          case '\\':
          case '\?':
            *out++ = cp[1];
            cp += 2;
            break;
          default:
            tor_free(*result); return NULL;
          }
        break;
      default:
        *out++ = *cp++;
      }
  }
}

/** Given a string containing part of a configuration file or similar format,
 * advance past comments and whitespace and try to parse a single line.  If we
 * parse a line successfully, set *<b>key_out</b> to a new string holding the
 * key portion and *<b>value_out</b> to a new string holding the value portion
 * of the line, and return a pointer to the start of the next line.  If we run
 * out of data, return a pointer to the end of the string.  If we encounter an
 * error, return NULL.
 */
const char *
parse_config_line_from_str(const char *line, char **key_out, char **value_out)
{
  /* I believe the file format here is supposed to be:
     FILE = (EMPTYLINE | LINE)* (EMPTYLASTLINE | LASTLINE)?

     EMPTYLASTLINE = SPACE* | COMMENT
     EMPTYLINE = EMPTYLASTLINE NL
     SPACE = ' ' | '\r' | '\t'
     COMMENT = '#' NOT-NL*
     NOT-NL = Any character except '\n'
     NL = '\n'

     LASTLINE = SPACE* KEY SPACE* VALUES
     LINE = LASTLINE NL
     KEY = KEYCHAR+
     KEYCHAR = Any character except ' ', '\r', '\n', '\t', '#', "\"

     VALUES = QUOTEDVALUE | NORMALVALUE
     QUOTEDVALUE = QUOTE QVITEM* QUOTE EOLSPACE?
     QUOTE = '"'
     QVCHAR = KEYCHAR | ESC ('n' | 't' | 'r' | '"' | ESC |'\'' | OCTAL | HEX)
     ESC = "\\"
     OCTAL = ODIGIT (ODIGIT ODIGIT?)?
     HEX = ('x' | 'X') HEXDIGIT HEXDIGIT
     ODIGIT = '0' .. '7'
     HEXDIGIT = '0'..'9' | 'a' .. 'f' | 'A' .. 'F'
     EOLSPACE = SPACE* COMMENT?

     NORMALVALUE = (VALCHAR | ESC ESC_IGNORE | CONTINUATION)* EOLSPACE?
     VALCHAR = Any character except ESC, '#', and '\n'
     ESC_IGNORE = Any character except '#' or '\n'
     CONTINUATION = ESC NL ( COMMENT NL )*
   */

  const char *key, *val, *cp;
  int continuation = 0;

  tor_assert(key_out);
  tor_assert(value_out);

  *key_out = *value_out = NULL;
  key = val = NULL;
  /* Skip until the first keyword. */
  while (1) {
    while (TOR_ISSPACE(*line))
      ++line;
    if (*line == '#') {
      while (*line && *line != '\n')
        ++line;
    } else {
      break;
    }
  }

  if (!*line) { /* End of string? */
    *key_out = *value_out = NULL;
    return line;
  }

  /* Skip until the next space or \ followed by newline. */
  key = line;
  while (*line && !TOR_ISSPACE(*line) && *line != '#' &&
         ! (line[0] == '\\' && line[1] == '\n'))
    ++line;
  *key_out = tor_strndup(key, line-key);

  /* Skip until the value. */
  while (*line == ' ' || *line == '\t')
    ++line;

  val = line;

  /* Find the end of the line. */
  if (*line == '\"') { // XXX No continuation handling is done here
    if (!(line = unescape_string(line, value_out, NULL)))
       return NULL;
    while (*line == ' ' || *line == '\t')
      ++line;
    if (*line && *line != '#' && *line != '\n')
      return NULL;
  } else {
    /* Look for the end of the line. */
    while (*line && *line != '\n' && (*line != '#' || continuation)) {
      if (*line == '\\' && line[1] == '\n') {
        continuation = 1;
        line += 2;
      } else if (*line == '#') {
        do {
          ++line;
        } while (*line && *line != '\n');
        if (*line == '\n')
          ++line;
      } else {
        ++line;
      }
    }

    if (*line == '\n') {
      cp = line++;
    } else {
      cp = line;
    }
    /* Now back cp up to be the last nonspace character */
    while (cp>val && TOR_ISSPACE(*(cp-1)))
      --cp;

    tor_assert(cp >= val);

    /* Now copy out and decode the value. */
    *value_out = tor_strndup(val, cp-val);
    if (continuation) {
      char *v_out, *v_in;
      v_out = v_in = *value_out;
      while (*v_in) {
        if (*v_in == '#') {
          do {
            ++v_in;
          } while (*v_in && *v_in != '\n');
          if (*v_in == '\n')
            ++v_in;
        } else if (v_in[0] == '\\' && v_in[1] == '\n') {
          v_in += 2;
        } else {
          *v_out++ = *v_in++;
        }
      }
      *v_out = '\0';
    }
  }

  if (*line == '#') {
    do {
      ++line;
    } while (*line && *line != '\n');
  }
  while (TOR_ISSPACE(*line)) ++line;

  return line;
}

/** Expand any homedir prefix on <b>filename</b>; return a newly allocated
 * string. */
char *
expand_filename(const char *filename)
{
  tor_assert(filename);
#ifdef MS_WINDOWS
  return tor_strdup(filename);
#else
  if (*filename == '~') {
    char *home, *result=NULL;
    const char *rest;

    if (filename[1] == '/' || filename[1] == '\0') {
      home = getenv("HOME");
      if (!home) {
        log_warn(LD_CONFIG, "Couldn't find $HOME environment variable while "
                 "expanding \"%s\"; defaulting to \"\".", filename);
        home = tor_strdup("");
      } else {
        home = tor_strdup(home);
      }
      rest = strlen(filename)>=2?(filename+2):"";
    } else {
#ifdef HAVE_PWD_H
      char *username, *slash;
      slash = strchr(filename, '/');
      if (slash)
        username = tor_strndup(filename+1,slash-filename-1);
      else
        username = tor_strdup(filename+1);
      if (!(home = get_user_homedir(username))) {
        log_warn(LD_CONFIG,"Couldn't get homedir for \"%s\"",username);
        tor_free(username);
        return NULL;
      }
      tor_free(username);
      rest = slash ? (slash+1) : "";
#else
      log_warn(LD_CONFIG, "Couldn't expend homedir on system without pwd.h");
      return tor_strdup(filename);
#endif
    }
    tor_assert(home);
    /* Remove trailing slash. */
    if (strlen(home)>1 && !strcmpend(home,PATH_SEPARATOR)) {
      home[strlen(home)-1] = '\0';
    }
    tor_asprintf(&result,"%s"PATH_SEPARATOR"%s",home,rest);
    tor_free(home);
    return result;
  } else {
    return tor_strdup(filename);
  }
#endif
}

#define MAX_SCANF_WIDTH 9999

/** Helper: given an ASCII-encoded decimal digit, return its numeric value.
 * NOTE: requires that its input be in-bounds. */
static int
digit_to_num(char d)
{
  int num = ((int)d) - (int)'0';
  tor_assert(num <= 9 && num >= 0);
  return num;
}

/** Helper: Read an unsigned int from *<b>bufp</b> of up to <b>width</b>
 * characters.  (Handle arbitrary width if <b>width</b> is less than 0.)  On
 * success, store the result in <b>out</b>, advance bufp to the next
 * character, and return 0.  On failure, return -1. */
static int
scan_unsigned(const char **bufp, unsigned *out, int width, int base)
{
  unsigned result = 0;
  int scanned_so_far = 0;
  const int hex = base==16;
  tor_assert(base == 10 || base == 16);
  if (!bufp || !*bufp || !out)
    return -1;
  if (width<0)
    width=MAX_SCANF_WIDTH;

  while (**bufp && (hex?TOR_ISXDIGIT(**bufp):TOR_ISDIGIT(**bufp))
         && scanned_so_far < width) {
    int digit = hex?hex_decode_digit(*(*bufp)++):digit_to_num(*(*bufp)++);
    unsigned new_result = result * base + digit;
    if (new_result > UINT32_MAX || new_result < result)
      return -1; /* over/underflow. */
    result = new_result;
    ++scanned_so_far;
  }

  if (!scanned_so_far) /* No actual digits scanned */
    return -1;

  *out = result;
  return 0;
}

/** Helper: copy up to <b>width</b> non-space characters from <b>bufp</b> to
 * <b>out</b>.  Make sure <b>out</b> is nul-terminated. Advance <b>bufp</b>
 * to the next non-space character or the EOS. */
static int
scan_string(const char **bufp, char *out, int width)
{
  int scanned_so_far = 0;
  if (!bufp || !out || width < 0)
    return -1;
  while (**bufp && ! TOR_ISSPACE(**bufp) && scanned_so_far < width) {
    *out++ = *(*bufp)++;
    ++scanned_so_far;
  }
  *out = '\0';
  return 0;
}

/** Locale-independent, minimal, no-surprises scanf variant, accepting only a
 * restricted pattern format.  For more info on what it supports, see
 * tor_sscanf() documentation.  */
int
tor_vsscanf(const char *buf, const char *pattern, va_list ap)
{
  int n_matched = 0;

  while (*pattern) {
    if (*pattern != '%') {
      if (*buf == *pattern) {
        ++buf;
        ++pattern;
        continue;
      } else {
        return n_matched;
      }
    } else {
      int width = -1;
      ++pattern;
      if (TOR_ISDIGIT(*pattern)) {
        width = digit_to_num(*pattern++);
        while (TOR_ISDIGIT(*pattern)) {
          width *= 10;
          width += digit_to_num(*pattern++);
          if (width > MAX_SCANF_WIDTH)
            return -1;
        }
        if (!width) /* No zero-width things. */
          return -1;
      }
      if (*pattern == 'u' || *pattern == 'x') {
        unsigned *u = va_arg(ap, unsigned *);
        const int base = (*pattern == 'u') ? 10 : 16;
        if (!*buf)
          return n_matched;
        if (scan_unsigned(&buf, u, width, base)<0)
          return n_matched;
        ++pattern;
        ++n_matched;
      } else if (*pattern == 's') {
        char *s = va_arg(ap, char *);
        if (width < 0)
          return -1;
        if (scan_string(&buf, s, width)<0)
          return n_matched;
        ++pattern;
        ++n_matched;
      } else if (*pattern == 'c') {
        char *ch = va_arg(ap, char *);
        if (width != -1)
          return -1;
        if (!*buf)
          return n_matched;
        *ch = *buf++;
        ++pattern;
        ++n_matched;
      } else if (*pattern == '%') {
        if (*buf != '%')
          return -1;
        ++buf;
        ++pattern;
      } else {
        return -1; /* Unrecognized pattern component. */
      }
    }
  }

  return n_matched;
}

/** Minimal sscanf replacement: parse <b>buf</b> according to <b>pattern</b>
 * and store the results in the corresponding argument fields.  Differs from
 * sscanf in that it: Only handles %u and %x and %Ns.  Does not handle
 * arbitrarily long widths. %u and %x do not consume any space.  Is
 * locale-independent.  Returns -1 on malformed patterns.
 *
 * (As with other locale-independent functions, we need this to parse data that
 * is in ASCII without worrying that the C library's locale-handling will make
 * miscellaneous characters look like numbers, spaces, and so on.)
 */
int
tor_sscanf(const char *buf, const char *pattern, ...)
{
  int r;
  va_list ap;
  va_start(ap, pattern);
  r = tor_vsscanf(buf, pattern, ap);
  va_end(ap);
  return r;
}

/** Return a new list containing the filenames in the directory <b>dirname</b>.
 * Return NULL on error or if <b>dirname</b> is not a directory.
 */
smartlist_t *
tor_listdir(const char *dirname)
{
  smartlist_t *result;
#ifdef MS_WINDOWS
  char *pattern;
  TCHAR tpattern[MAX_PATH] = {0};
  char name[MAX_PATH] = {0};
  HANDLE handle;
  WIN32_FIND_DATA findData;
  size_t pattern_len = strlen(dirname)+16;
  pattern = tor_malloc(pattern_len);
  tor_snprintf(pattern, pattern_len, "%s\\*", dirname);
#ifdef UNICODE
  mbstowcs(tpattern,pattern,MAX_PATH);
#else
  strlcpy(tpattern, pattern, MAX_PATH);
#endif
  if (INVALID_HANDLE_VALUE == (handle = FindFirstFile(tpattern, &findData))) {
    tor_free(pattern);
    return NULL;
  }
  result = smartlist_create();
  while (1) {
#ifdef UNICODE
    wcstombs(name,findData.cFileName,MAX_PATH);
#else
    strlcpy(name,findData.cFileName,sizeof(name));
#endif
    if (strcmp(name, ".") &&
        strcmp(name, "..")) {
      smartlist_add(result, tor_strdup(name));
    }
    if (!FindNextFile(handle, &findData)) {
      DWORD err;
      if ((err = GetLastError()) != ERROR_NO_MORE_FILES) {
        char *errstr = format_win32_error(err);
        log_warn(LD_FS, "Error reading directory '%s': %s", dirname, errstr);
        tor_free(errstr);
      }
      break;
    }
  }
  FindClose(handle);
  tor_free(pattern);
#else
  DIR *d;
  struct dirent *de;
  if (!(d = opendir(dirname)))
    return NULL;

  result = smartlist_create();
  while ((de = readdir(d))) {
    if (!strcmp(de->d_name, ".") ||
        !strcmp(de->d_name, ".."))
      continue;
    smartlist_add(result, tor_strdup(de->d_name));
  }
  closedir(d);
#endif
  return result;
}

/** Return true iff <b>filename</b> is a relative path. */
int
path_is_relative(const char *filename)
{
  if (filename && filename[0] == '/')
    return 0;
#ifdef MS_WINDOWS
  else if (filename && filename[0] == '\\')
    return 0;
  else if (filename && strlen(filename)>3 && TOR_ISALPHA(filename[0]) &&
           filename[1] == ':' && filename[2] == '\\')
    return 0;
#endif
  else
    return 1;
}

/* =====
 * Process helpers
 * ===== */

#ifndef MS_WINDOWS
/* Based on code contributed by christian grothoff */
/** True iff we've called start_daemon(). */
static int start_daemon_called = 0;
/** True iff we've called finish_daemon(). */
static int finish_daemon_called = 0;
/** Socketpair used to communicate between parent and child process while
 * daemonizing. */
static int daemon_filedes[2];
/** Start putting the process into daemon mode: fork and drop all resources
 * except standard fds.  The parent process never returns, but stays around
 * until finish_daemon is called.  (Note: it's safe to call this more
 * than once: calls after the first are ignored.)
 */
void
start_daemon(void)
{
  pid_t pid;

  if (start_daemon_called)
    return;
  start_daemon_called = 1;

  if (pipe(daemon_filedes)) {
    log_err(LD_GENERAL,"pipe failed; exiting. Error was %s", strerror(errno));
    exit(1);
  }
  pid = fork();
  if (pid < 0) {
    log_err(LD_GENERAL,"fork failed. Exiting.");
    exit(1);
  }
  if (pid) {  /* Parent */
    int ok;
    char c;

    close(daemon_filedes[1]); /* we only read */
    ok = -1;
    while (0 < read(daemon_filedes[0], &c, sizeof(char))) {
      if (c == '.')
        ok = 1;
    }
    fflush(stdout);
    if (ok == 1)
      exit(0);
    else
      exit(1); /* child reported error */
  } else { /* Child */
    close(daemon_filedes[0]); /* we only write */

    pid = setsid(); /* Detach from controlling terminal */
    /*
     * Fork one more time, so the parent (the session group leader) can exit.
     * This means that we, as a non-session group leader, can never regain a
     * controlling terminal.   This part is recommended by Stevens's
     * _Advanced Programming in the Unix Environment_.
     */
    if (fork() != 0) {
      exit(0);
    }
    set_main_thread(); /* We are now the main thread. */

    return;
  }
}

/** Finish putting the process into daemon mode: drop standard fds, and tell
 * the parent process to exit.  (Note: it's safe to call this more than once:
 * calls after the first are ignored.  Calls start_daemon first if it hasn't
 * been called already.)
 */
void
finish_daemon(const char *desired_cwd)
{
  int nullfd;
  char c = '.';
  if (finish_daemon_called)
    return;
  if (!start_daemon_called)
    start_daemon();
  finish_daemon_called = 1;

  if (!desired_cwd)
    desired_cwd = "/";
   /* Don't hold the wrong FS mounted */
  if (chdir(desired_cwd) < 0) {
    log_err(LD_GENERAL,"chdir to \"%s\" failed. Exiting.",desired_cwd);
    exit(1);
  }

  nullfd = tor_open_cloexec("/dev/null", O_RDWR, 0);
  if (nullfd < 0) {
    log_err(LD_GENERAL,"/dev/null can't be opened. Exiting.");
    exit(1);
  }
  /* close fds linking to invoking terminal, but
   * close usual incoming fds, but redirect them somewhere
   * useful so the fds don't get reallocated elsewhere.
   */
  if (dup2(nullfd,0) < 0 ||
      dup2(nullfd,1) < 0 ||
      dup2(nullfd,2) < 0) {
    log_err(LD_GENERAL,"dup2 failed. Exiting.");
    exit(1);
  }
  if (nullfd > 2)
    close(nullfd);
  /* signal success */
  if (write(daemon_filedes[1], &c, sizeof(char)) != sizeof(char)) {
    log_err(LD_GENERAL,"write failed. Exiting.");
  }
  close(daemon_filedes[1]);
}
#else
/* defined(MS_WINDOWS) */
void
start_daemon(void)
{
}
void
finish_daemon(const char *cp)
{
  (void)cp;
}
#endif

/** Write the current process ID, followed by NL, into <b>filename</b>.
 */
void
write_pidfile(char *filename)
{
  FILE *pidfile;

  if ((pidfile = fopen(filename, "w")) == NULL) {
    log_warn(LD_FS, "Unable to open \"%s\" for writing: %s", filename,
             strerror(errno));
  } else {
#ifdef MS_WINDOWS
    fprintf(pidfile, "%d\n", (int)_getpid());
#else
    fprintf(pidfile, "%d\n", (int)getpid());
#endif
    fclose(pidfile);
  }
}

#ifdef MS_WINDOWS
HANDLE
load_windows_system_library(const TCHAR *library_name)
{
  TCHAR path[MAX_PATH];
  unsigned n;
  n = GetSystemDirectory(path, MAX_PATH);
  if (n == 0 || n + _tcslen(library_name) + 2 >= MAX_PATH)
    return 0;
  _tcscat(path, TEXT("\\"));
  _tcscat(path, library_name);
  return LoadLibrary(path);
}
#endif

/** Format <b>child_state</b> and <b>saved_errno</b> as a hex string placed in
 * <b>hex_errno</b>.  Called between fork and _exit, so must be signal-handler
 * safe.
 *
 * <b>hex_errno</b> must have at least HEX_ERRNO_SIZE bytes available.
 *
 * The format of <b>hex_errno</b> is: "CHILD_STATE/ERRNO\n", left-padded
 * with spaces. Note that there is no trailing \0. CHILD_STATE indicates where
 * in the processs of starting the child process did the failure occur (see
 * CHILD_STATE_* macros for definition), and SAVED_ERRNO is the value of
 * errno when the failure occurred.
 */

void
format_helper_exit_status(unsigned char child_state, int saved_errno,
                          char *hex_errno)
{
  unsigned int unsigned_errno;
  char *cur;
  size_t i;

  /* Fill hex_errno with spaces, and a trailing newline (memset may
     not be signal handler safe, so we can't use it) */
  for (i = 0; i < (HEX_ERRNO_SIZE - 1); i++)
    hex_errno[i] = ' ';
  hex_errno[HEX_ERRNO_SIZE - 1] = '\n';

  /* Convert errno to be unsigned for hex conversion */
  if (saved_errno < 0) {
    unsigned_errno = (unsigned int) -saved_errno;
  } else {
    unsigned_errno = (unsigned int) saved_errno;
  }

  /* Convert errno to hex (start before \n) */
  cur = hex_errno + HEX_ERRNO_SIZE - 2;

  /* Check for overflow on first iteration of the loop */
  if (cur < hex_errno)
    return;

  do {
    *cur-- = "0123456789ABCDEF"[unsigned_errno % 16];
    unsigned_errno /= 16;
  } while (unsigned_errno != 0 && cur >= hex_errno);

  /* Prepend the minus sign if errno was negative */
  if (saved_errno < 0 && cur >= hex_errno)
    *cur-- = '-';

  /* Leave a gap */
  if (cur >= hex_errno)
    *cur-- = '/';

  /* Check for overflow on first iteration of the loop */
  if (cur < hex_errno)
    return;

  /* Convert child_state to hex */
  do {
    *cur-- = "0123456789ABCDEF"[child_state % 16];
    child_state /= 16;
  } while (child_state != 0 && cur >= hex_errno);
}

/* Maximum number of file descriptors, if we cannot get it via sysconf() */
#define DEFAULT_MAX_FD 256

#define CHILD_STATE_INIT 0
#define CHILD_STATE_PIPE 1
#define CHILD_STATE_MAXFD 2
#define CHILD_STATE_FORK 3
#define CHILD_STATE_DUPOUT 4
#define CHILD_STATE_DUPERR 5
#define CHILD_STATE_REDIRECT 6
#define CHILD_STATE_CLOSEFD 7
#define CHILD_STATE_EXEC 8
#define CHILD_STATE_FAILEXEC 9

#define SPAWN_ERROR_MESSAGE "ERR: Failed to spawn background process - code "

/** Start a program in the background. If <b>filename</b> contains a '/',
 * then it will be treated as an absolute or relative path.  Otherwise the
 * system path will be searched for <b>filename</b>. The strings in
 * <b>argv</b> will be passed as the command line arguments of the child
 * program (following convention, argv[0] should normally be the filename of
 * the executable). The last element of argv must be NULL. If the child
 * program is launched, the PID will be returned and <b>stdout_read</b> and
 * <b>stdout_err</b> will be set to file descriptors from which the stdout
 * and stderr, respectively, output of the child program can be read, and the
 * stdin of the child process shall be set to /dev/null.  Otherwise returns
 * -1.  Some parts of this code are based on the POSIX subprocess module from
 * Python.
 */
int
tor_spawn_background(const char *const filename, int *stdout_read,
                     int *stderr_read, const char **argv)
{
#ifdef MS_WINDOWS
  (void) filename; (void) stdout_read; (void) stderr_read; (void) argv;
  log_warn(LD_BUG, "not yet implemented on Windows.");
  return -1;
#else
  pid_t pid;
  int stdout_pipe[2];
  int stderr_pipe[2];
  int fd, retval;
  ssize_t nbytes;

  const char *error_message = SPAWN_ERROR_MESSAGE;
  size_t error_message_length;

  /* Represents where in the process of spawning the program is;
     this is used for printing out the error message */
  unsigned char child_state = CHILD_STATE_INIT;

  char hex_errno[HEX_ERRNO_SIZE];

  static int max_fd = -1;

  /* We do the strlen here because strlen() is not signal handler safe,
     and we are not allowed to use unsafe functions between fork and exec */
  error_message_length = strlen(error_message);

  child_state = CHILD_STATE_PIPE;

  /* Set up pipe for redirecting stdout and stderr of child */
  retval = pipe(stdout_pipe);
  if (-1 == retval) {
    log_warn(LD_GENERAL,
      "Failed to set up pipe for stdout communication with child process: %s",
       strerror(errno));
    return -1;
  }

  retval = pipe(stderr_pipe);
  if (-1 == retval) {
    log_warn(LD_GENERAL,
      "Failed to set up pipe for stderr communication with child process: %s",
      strerror(errno));
    return -1;
  }

  child_state = CHILD_STATE_MAXFD;

#ifdef _SC_OPEN_MAX
  if (-1 != max_fd) {
    max_fd = (int) sysconf(_SC_OPEN_MAX);
    if (max_fd == -1)
      max_fd = DEFAULT_MAX_FD;
      log_warn(LD_GENERAL,
               "Cannot find maximum file descriptor, assuming %d", max_fd);
  }
#else
  max_fd = DEFAULT_MAX_FD;
#endif

  child_state = CHILD_STATE_FORK;

  pid = fork();
  if (0 == pid) {
    /* In child */

    child_state = CHILD_STATE_DUPOUT;

    /* Link child stdout to the write end of the pipe */
    retval = dup2(stdout_pipe[1], STDOUT_FILENO);
    if (-1 == retval)
        goto error;

    child_state = CHILD_STATE_DUPERR;

    /* Link child stderr to the write end of the pipe */
    retval = dup2(stderr_pipe[1], STDERR_FILENO);
    if (-1 == retval)
        goto error;

    child_state = CHILD_STATE_REDIRECT;

    /* Link stdin to /dev/null */
    fd = open("/dev/null", O_RDONLY); /* NOT cloexec, obviously. */
    if (fd != -1)
      dup2(fd, STDIN_FILENO);
    else
      goto error;

    child_state = CHILD_STATE_CLOSEFD;

    close(stderr_pipe[0]);
    close(stderr_pipe[1]);
    close(stdout_pipe[0]);
    close(stdout_pipe[1]);
    close(fd);

    /* Close all other fds, including the read end of the pipe */
    /* XXX: We should now be doing enough FD_CLOEXEC setting to make
     * this needless. */
    for (fd = STDERR_FILENO + 1; fd < max_fd; fd++) {
      close(fd);
    }

    child_state = CHILD_STATE_EXEC;

    /* Call the requested program. We need the cast because
       execvp doesn't define argv as const, even though it
       does not modify the arguments */
    execvp(filename, (char *const *) argv);

    /* If we got here, the exec or open(/dev/null) failed */

    child_state = CHILD_STATE_FAILEXEC;

  error:
    /* XXX: are we leaking fds from the pipe? */

    format_helper_exit_status(child_state, errno, hex_errno);

    /* Write the error message. GCC requires that we check the return
       value, but there is nothing we can do if it fails */
    nbytes = write(STDOUT_FILENO, error_message, error_message_length);
    nbytes = write(STDOUT_FILENO, hex_errno, sizeof(hex_errno));

    _exit(255);
    return -1; /* Never reached, but avoids compiler warning */
  }

  /* In parent */

  if (-1 == pid) {
    log_warn(LD_GENERAL, "Failed to fork child process: %s", strerror(errno));
    close(stdout_pipe[0]);
    close(stdout_pipe[1]);
    close(stderr_pipe[0]);
    close(stderr_pipe[1]);
    return -1;
  }

  /* Return read end of the pipes to caller, and close write end */
  *stdout_read = stdout_pipe[0];
  retval = close(stdout_pipe[1]);

  if (-1 == retval) {
    log_warn(LD_GENERAL,
            "Failed to close write end of stdout pipe in parent process: %s",
            strerror(errno));
    /* Do not return -1, because the child is running, so the parent
       needs to know about the pid in order to reap it later */
  }

  *stderr_read = stderr_pipe[0];
  retval = close(stderr_pipe[1]);

  if (-1 == retval) {
    log_warn(LD_GENERAL,
            "Failed to close write end of stderr pipe in parent process: %s",
            strerror(errno));
    /* Do not return -1, because the child is running, so the parent
       needs to know about the pid in order to reap it later */
  }

  return pid;
#endif
}

/** Read from stream, and send lines to log at the specified log level.
 * Returns 1 if stream is closed normally, -1 if there is a error reading, and
 * 0 otherwise. Handles lines from tor-fw-helper and
 * tor_spawn_background() specially.
 */
static int
log_from_pipe(FILE *stream, int severity, const char *executable,
              int *child_status)
{
  char buf[256];

  for (;;) {
    char *retval;
    retval = fgets(buf, sizeof(buf), stream);

    if (NULL == retval) {
      if (feof(stream)) {
        /* Program has closed stream (probably it exited) */
        /* TODO: check error */
        fclose(stream);
        return 1;
      } else {
        if (EAGAIN == errno) {
          /* Nothing more to read, try again next time */
          return 0;
        } else {
          /* There was a problem, abandon this child process */
          fclose(stream);
          return -1;
        }
      }
    } else {
      /* We have some data, log it and keep asking for more */
      size_t len;

      len = strlen(buf);
      if (buf[len - 1] == '\n') {
        /* Remove the trailing newline */
        buf[len - 1] = '\0';
      } else {
        /* No newline; check whether we overflowed the buffer */
        if (!feof(stream))
          log_warn(LD_GENERAL,
                  "Line from port forwarding helper was truncated: %s", buf);
          /* TODO: What to do with this error? */
      }

      /* Check if buf starts with SPAWN_ERROR_MESSAGE */
      if (strcmpstart(buf, SPAWN_ERROR_MESSAGE) == 0) {
          /* Parse error message */
          int retval, child_state, saved_errno;
          retval = tor_sscanf(buf, SPAWN_ERROR_MESSAGE "%x/%x",
                              &child_state, &saved_errno);
          if (retval == 2) {
              log_warn(LD_GENERAL,
                "Failed to start child process \"%s\" in state %d: %s",
                executable, child_state, strerror(saved_errno));
              if (child_status)
                  *child_status = 1;
          } else {
              /* Failed to parse message from child process, log it as a
                 warning */
              log_warn(LD_GENERAL,
                "Unexpected message from port forwarding helper \"%s\": %s",
                executable, buf);
          }
      } else {
          log_fn(severity, LD_GENERAL, "Port forwarding helper says: %s", buf);
      }
    }
  }

  /* We should never get here */
  return -1;
}

void
tor_check_port_forwarding(const char *filename, int dir_port, int or_port,
                          time_t now)
{
#ifdef MS_WINDOWS
  (void) filename; (void) dir_port; (void) or_port; (void) now;
  (void) tor_spawn_background;
  (void) log_from_pipe;
  log_warn(LD_GENERAL, "Sorry, port forwarding is not yet supported "
           "on windows.");
#else
/* When fw-helper succeeds, how long do we wait until running it again */
#define TIME_TO_EXEC_FWHELPER_SUCCESS 300
/* When fw-helper fails, how long do we wait until running it again */
#define TIME_TO_EXEC_FWHELPER_FAIL 60

  static int child_pid = -1;
  static FILE *stdout_read = NULL;
  static FILE *stderr_read = NULL;
  static time_t time_to_run_helper = 0;
  int stdout_status, stderr_status, retval;
  const char *argv[10];
  char s_dirport[6], s_orport[6];

  tor_assert(filename);

  /* Set up command line for tor-fw-helper */
  snprintf(s_dirport, sizeof s_dirport, "%d", dir_port);
  snprintf(s_orport, sizeof s_orport, "%d", or_port);

  /* TODO: Allow different internal and external ports */
  argv[0] = filename;
  argv[1] = "--internal-or-port";
  argv[2] = s_orport;
  argv[3] = "--external-or-port";
  argv[4] = s_orport;
  argv[5] = "--internal-dir-port";
  argv[6] = s_dirport;
  argv[7] = "--external-dir-port";
  argv[8] = s_dirport;
  argv[9] = NULL;

  /* Start the child, if it is not already running */
  if (-1 == child_pid &&
      time_to_run_helper < now) {
    int fd_out=-1, fd_err=-1;

    /* Assume tor-fw-helper will succeed, start it later*/
    time_to_run_helper = now + TIME_TO_EXEC_FWHELPER_SUCCESS;

    child_pid = tor_spawn_background(filename, &fd_out, &fd_err, argv);
    if (child_pid < 0) {
      log_warn(LD_GENERAL, "Failed to start port forwarding helper %s",
              filename);
      child_pid = -1;
      return;
    }
    /* Set stdout/stderr pipes to be non-blocking */
    fcntl(fd_out, F_SETFL, O_NONBLOCK);
    fcntl(fd_err, F_SETFL, O_NONBLOCK);
    /* Open the buffered IO streams */
    stdout_read = fdopen(fd_out, "r");
    stderr_read = fdopen(fd_err, "r");

    log_info(LD_GENERAL,
      "Started port forwarding helper (%s) with pid %d", filename, child_pid);
  }

  /* If child is running, read from its stdout and stderr) */
  if (child_pid > 0) {
    /* Read from stdout/stderr and log result */
    retval = 0;
    stdout_status = log_from_pipe(stdout_read, LOG_INFO, filename, &retval);
    stderr_status = log_from_pipe(stderr_read, LOG_WARN, filename, &retval);
    if (retval) {
      /* There was a problem in the child process */
      time_to_run_helper = now + TIME_TO_EXEC_FWHELPER_FAIL;
    }

    /* Combine the two statuses in order of severity */
    if (-1 == stdout_status || -1 == stderr_status)
      /* There was a failure */
      retval = -1;
    else if (1 == stdout_status || 1 == stderr_status)
      /* stdout or stderr was closed */
      retval = 1;
    else
      /* Both are fine */
      retval = 0;

    /* If either pipe indicates a failure, act on it */
    if (0 != retval) {
      if (1 == retval) {
        log_info(LD_GENERAL, "Port forwarding helper terminated");
      } else {
        log_warn(LD_GENERAL, "Failed to read from port forwarding helper");
      }

      /* TODO: The child might not actually be finished (maybe it failed or
         closed stdout/stderr), so maybe we shouldn't start another? */
      child_pid = -1;
    }
  }
#endif
}