File: zmog.tex

package info (click to toggle)
zmailer 2.99.51.52pre3-2
  • links: PTS
  • area: main
  • in suites: potato
  • size: 16,596 kB
  • ctags: 7,422
  • sloc: ansic: 90,470; sh: 3,608; makefile: 2,784; perl: 1,585; python: 115; awk: 22
file content (3697 lines) | stat: -rw-r--r-- 159,617 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
3678
3679
3680
3681
3682
3683
3684
3685
3686
3687
3688
3689
3690
3691
3692
3693
3694
3695
3696
3697
\input texinfo   @c -*-texinfo-*-
@setfilename zmog
@settitle ZMailer Operations Guide

@titlepage
@center @titlefont{ZMOG}
@sp 1
@center The ZMailer Operations Guide
@center Beta Release
@center Version 2.99.21
@sp 2
@center Matti Aarnio
@center mea@@nic.funet.fi
@sp 1
@center Finnish University and Research Network
@center FUNET
@end titlepage

@unnumbered Distribution

Copyright @copyright{} 1988 Rayan S. Zachariassen.
Copyright @copyright{} 1995 Matti Aarnio

This manual is freely distributable for use with ZMailer sources,
which are available at various sites around the network.

@node top, introduction, ,
@menu
* introduction::    An introduction to ZMailer.
* overview::        General overview of how ZMailer works.
* router::          All about the Router process.
* scheduler::       All about the Scheduler process.
* transports::      All about the Transport Agents.
* miscellaneous::   Miscellany topics (compatibility, etc.)
* how-to::          A How-To guide.
@end menu

@node introduction, overview, top, top
@chapter Introduction

ZMailer@iftex
@footnote{@dag}{Evoking references to The Mailer, That Mailer,
or even Le Mailer, depending on reader's origin.}
@end iftex
is a mailer subsystem for the UNIX@iftex
@footnote{@ddag}{UNIX was a Trademark of AT&T Bell Laboratories,
but since then resold several times. Who owns it today, is for
the reader to find out -- if they care.}
@end iftex
operating system.

A mailer is in charge of handling all mail messages that
are created on a system (typically a single host), from
their creation until final disposition locally or by transfer
to another system.

As such, the mailer subsystem (the Message Transfer Agent, a.k.a. MTA)
must interface to local mail reading and composing programs
(User Agents, a.k.a. MUAs), to the various transport methods
that can be used to reach other mailers, and to a variety of
databases describing the mailer's environment.

ZMailer provides this functionality in a package and with a philosophy
that has benefited from experiences with earlier mailers.

ZMailer provides a capable, robust, efficient subsystem to do the job,
which will excel in demanding environments, but is simple enough to fit
easily everywhere.

While the ZMailer does take more system resources, than
sendmail, it controls that resource consumption a bit better;
SMTP input, and @file{sendmail}-compability module are rather
lightweight processes that just feed a job spool file into
local filesystem for other system components to tackle with.

Message routing decissions are done with a fixed set of processes,
and final delivery is achieved with rather fine-grain controlled
scheduler system.


@section Motivation and Heritage

Many of my@footnote{``I'' being usually Rayan Zachariassen
around late 1980es;  Matti Aarnio came in latter.}
reasons for trying to improve the state of the art in message
handling systems are based on the fact that the capabilities
of available software hasn't changed for a long time, whereas
the demands being placed on the software have steadily risen.

A few years ago, people were still dreaming of a world with
consistent standards for addressing electronic mail and moving
it around.
Even though the consistency is constantly improving, it now
seems apparent that there will always be needs that conflict
with the ideal situation.
This is very obvious to sites that interact with two or more
types of networks.

For those sites that are directly attached to just one network,
any degree of transparency in communicating with sites on other
networks, must be provided by the software on a mail gateway.
Most such software was not designed to perform the task of
gatewaying messages between networks with heterogenous addressing
and message standards.

ZMailer was primarily a reaction to Sendmail's disadvantages
(which I shall mention), but also to the bad (and good!) points
of several other mailers.

The design of ZMailer was often guided by my view of the poor
choices or provisions of other mailers, as opposed to things
they had done well.

This allows the design to draw from experience without limiting
its creativity and use of new solutions.

To clarify my opinions somewhat, a commentary on the various
mailers I know of should prove helpful:


@subsection Heritage: sendmail

The best available system to accomplish address manipulation in
a heterogenous environment has previously been Sendmail, which
was carried along on its relative strength for address manipulation.

Unfortunately, Sendmail@footnote{We must agree that Sendmail v.8
has improved a lot from previously abundant v.5@dots{}} has many
design flaws that lessen its usefulness even in typical environments
of our time.

Sendmail's major contribution to the art was the use of production
rules to manipulate addresses and to guide the operations performed
on a message.

It also popularized, and in certain environments pioneered, other
functionality that turned out to be quite useful; for example the
use of system-wide aliases file, as well as personal forwarding file,
a widely available SMTP implementation, external and consistently
treated delivery programs, etc.

Sendmail, in the right hands, can be quite a flexible tool to
translate between the different conventions of various networks.
Unfortunately this is accomplished by programming in an unfamiliar
production language containing many magic features.

The learning time for doing this is very long, the effort involved
is that of learning a completely new language and environment.
Moreover, Sendmail has all major components built into a single
large program.

Both of these design decisions have been acknowledged as mistakes
by the author of Sendmail.

One of the big problems with Sendmail is its style of running all of
the processing from submission/reception of the message to its final
delivery in one process.

It makes no problem when the number of emails in processing is
small, but when there are a lot of messages flowing, there are
a lot of memory consuming, and load generating processes active.
(With Sendmail v.8 there are reasonable load limiting functions
 available though@dots{})

Its major shortcoming in comparison to the MMDF mailer is its
primitive database facility and lack of caching.
(Again the sendmail v.8 has since then improved things a bit.)


@subsection Heritage: MMDF, PMDF

MMDF is a comprehensive mail environment, including its own mail
composition program, and of course a mailer.

There are too many parts to it (as a friend would say,
it is a system, not a subsystem), and the address manipulation
is only sufficient for a relatively homogenous environment.

It does have reasonable database facilities and caching,
as opposed to Sendmail, and the concept of Channels.

However, knowledge about address semantics is distributed in
several programs instead of being centralized.

PMDF is a smaller version of MMDF with correspondingly reduced
features and flexibility. (Commercial PMDF is different story@dots{})


@subsection Heritage: Upas

Upas is a curious approach to the problem.
It lets the user do half the work of message routing, in
a manner similar to PMDF on VMS systems.

It is entirely concerned with the message envelope, and
leaves all message header munging to auxiliary programs,
if appropriate.
In fairness one should note this mailer was developed in
an environment where most message headers were scorned,
thus making this a reasonable approach (``optimize the normal case'').

The Eighth Edition Upas had no database capability at all,
but it did exhibit one useful characteristic: the routing
decisions were made by passing the recipient envelope address
through a set of regular expressions.
This production rule approach is similar to what Sendmail does,
but uses a more familiar mechanism and environment.

@subsection Heritage: Smail 3.0

The final, and most recently developed, mailer worth mentioning
here is Smail3.0.@footnote{Around 1988, since then Smail3.1 has
been released with many differences@dots{}}.

It is intended as a program capable of replacing Sendmail in
many situations.
To a large extent it succeeds as this, and there are some nice
ideas involved as well.

Its two major drawbacks are that it is not as easy to adapt to
local needs as Sendmail is (compiled instead of interpreted
rules and algorithms), and retaining Sendmail's single-program
design.

It addresses database and caching issues, and seems generally
like a nicer design in many respects, a bit like PMDF's
configuration options in a Sendmail package.

@bold{XXX: Smail 3.1 ??}

@subsection Result: ZMailer

Until the recent increase in the demand for inter-network mail
gatewaying, Sendmail's flexibility had quite adequately served
to implement a gateway function between selected networks.

With increased variety of the normal address syntax and mail
capabilities of connected networks, and more complex kinds of
routing decisions becoming necessary, the existing mailers have
been showing their age and their limits.

Also recent nearly exponential growth of the Internet has put
serious demands on system performance, and there is readily a
market for mailers processing more than million messages a day!

ZMailer is intended to give the mail administrator a software
tool that fits the times.

@group
@section Goals

Apart from the generic goals of robustness and efficiency,
the following is a list of the specific goals of ZMailer:

@itemize @bullet
@item
Fully RFC822/RFC976 compatible syntax and semantics.
@item
At a minimum provide Sendmail functionality from point of
view of users and the system administrator.
@item
Make routing decisions based on original sender and path
of the message.
@item
Not have hardcoded address rewriting and routing algorithms.
@item
A better user interface for the mail administrator than Sendmail.
@item
Interact properly with Internet Nameservers.
@item
Easily extensible to make use of new sources of data.
@item
Efficient enough to handle a large message volume, and to
not significantly degrade its own or system performance
when many messages are queued.
@item
Schedule delivery based on destination channel, destination host,
or a combination of both.
@item
Able to do a better job, than Sendmail in our high load environments.
@item
Be small enough to be usable even on a smallest of the workstations
as local program -- well, ``small'' is rather flexible, but Linux
version takes circa 1.0 MB disk, and 0.5 MB RAM when running in minimum
configuration (Linux-1.2).  For Sendmail the same figures are circa
0.3/0.3 MB@dots{}
@end itemize
@end group

For a while it has been apparent that Sendmail's approach to its
task is not well-suited from several perspectives.

In particular, having a single program embody several conceptually
independent functions is recognized as poor design.

In practice, merging queueing and delivery in one program causes
a bottleneck for all messages when a particular delivery mechanism
is slow. (Actually this is not an easy issue on any system.)


@section Design Summary

ZMailer is a multi-process mailer, using two daemon processes
to manipulate messages.
One of these processes is a router (or a set of routers), and it
makes all decisions about what should happen to a message.
The other daemon is a message queue manager, used to schedule
delivery of messages.

The Router uses a configuration file that closely follows Bourne
shell script syntax and semantics, with minimal magic.

Message files are moved around in a series of directories,
and the Scheduler and its Transport Agents run off of control
files created by the Router.

The Router will process messages one at a time, as it finds them
in a directory where User Agents submit their outgoing messages.
(There can be parallel Router processes speeding up the processing.)

Envelope and Message Header information is all kept in the same
message file along with the message body, and this file is never
modified by any ZMailer program.

After parsing the envelope and RFC822 header information,
the Router validates the information extracted, and calls
functions defined in the configuration file to decide exactly
how to deliver the message and how to transform the embedded
addresses.
The algorithms that do this are easily reconfigurable, since
the control flow and address manipulation is specified by familiar
shell script statements.

When the Router is finished, it will produce a message control
file for use by the delivery processing stage of ZMailer, and
move the original message file to another location.

Once the Router has decided what to do with each of the addresses
in a message, the Scheduler builds a summary of this information
by reading the control file created by the Router.
This knowledge is merged with a data structure it maintains that
stores which messages are supposed to be sent where, and how.
According to a pre-arranged agenda, the Scheduler will execute
delivery programs to properly move the message envelope, header,
and body, to the immediate destination.

These delivery programs are called Transport Agents, and communicate
with the Scheduler using a simple protocol that tells them which
messages to process and returns status reports to the Scheduler.

There are several standard Transport Agents included with
the ZMailer distribution.
The collection currently includes a local delivery program, an
SMTP client implementation, and a Transport Agent that can run
Sendmail-compatible delivery programs.

The Scheduler also manages status reports, taking appropriate action
on delivery errors and when all delivery instructions for a message
have been processed.

A separate utility allows querying the Scheduler for the state
of its mail queues.

For existing Sendmail installations, a replacement program is
included that simulates most of the Sendmail functionality in
the ZMailer environment.
This allows ZMailer to replace a Sendmail installation
without requiring changes in standard User Agents.

@node overview, router, introduction, top
@chapter Overview

This chapter deals with the life of a message, and what
will happen to a message in the course of being processed.
The processing activity is divided into four major phases
that will be dealt with here.

These phases are: injection of a message into the mailer
subsystem, message routing, message transport queueing
and scheduling, and actual delivery of a message.

The phases communicate through the filesystem, by moving
files from one directory to another.

All directories taking part in this communication are clustered
under the @file{POSTOFFICE} directory (@file{/usr/spool/postoffice}),
which is intended to also hold other maintenance information or
directories for use by the system Postmaster.

For this reason, we shall refer to files within this hierarchy
using the tilde-abbreviation (e.g. @file{~/file} referring to
the Postmaster's @file{$HOME/file} which is normally
@file{/usr/spool/postoffice/file}).

@section Message Submission

A mail message is submitted to the mailer subsystem by depositing
a ``message file'' in a particular @file{ROUTER} directory
(@file{~/router}).
There is a Sendmail replacement program which submits messages this way.

The messages are picked up by a daemon process scanning this
directory, and processing all message files it finds in it.

To avoid problems with the daemon processing an incomplete
or inconsistent message file, the message files are created
in a separate @file{PUBLIC} directory (@file{~/public}), and
then renamed into the @file{ROUTER} directory.

A message file has 3 parts to it, the first part contains
the envelope information for the message (if any).
It consists of all the lines from the start of the file to
the start of part 2 (exclusive) which are in the format of
RFC822 Message Header lines @emph{except} that there is no
colon after the header field name.

An RFC822 Message Header is easily converted to an envelope
header line by simply deleting the colon after the field name.
As with message headers, various field names have specific
semantics, and these will be discussed in detail later.

The second part of a message file is its RFC822 Message Header.
The third part is the message body. Either of the envelope portion
(part 1) or the message header portion (part 2) may be null.
They are separated from the message body by an empty line,
according to RFC822.

The standard UNIX conventions for files are obeyed (i.e. lines
are terminated by a Newline character (@key{LF})).
The message body is never examined by the mailer itself, although
transport/delivery programs must of course filter the message body
appropriately for the destination.
In other words, the message body may contain arbitrary binary data.
The only restrictions are that the envelope and message header must
obey the RFC822 lexical/syntax rules.


Once the message file has been written into the @file{ROUTER}
directory, its content will never change until the system removes
it after successful delivery.


A subroutine interface exists, which should be used by application
programs or User Agents to submit messages.

The subroutine interface is properly part of the system C library
(@file{/lib/libc.a}), and will be documented as such.

It is quite possible to submit messages by using the standard
utilities to copy or move a file into the @file{ROUTER} directory.

Indeed, some maintenance functions the Postmaster should perform,
and automatic resubmission of deferred messages, are most easily
accomplished in this manner.

Note that the format of a message file allows a user to simply
create a file that obeys RFC822 conventions, in order to submit
a message.
It also allows simple resubmission of a message which includes
the UNIX standard @samp{From } envelope header line (found in
@code{Mail} format mailbox files), as this syntax will indeed
be interpreted to represent envelope information.

@section Router

The daemon mentioned above, the one that processes message
files appearing in the @file{ROUTER} directory, is called
the Router process.
It is the responsibility of this process to decide what to
do with the message file, and pass this information on to
the next stage (message transport/delivery queueing and scheduling).
It does this by creating a Control File attached to the message
file, which contains all the necessary information for the next
stages to accomplish their work without detailed reference to
the message file (i.e.  without having to reparse it).
When the Router process is finished processing a message file,
it will relink it into a @file{QUEUE} (@file{~/queue}) directory
(it is flat directory, but it should not be a problem), and deposit
the control file in a @file{TRANSPORT} directory (@file{~/transport})
for processing by the next stage.

The Router will parse the message file contents, determine
the boundaries between the various parts of the file, extract
addresses and other information from the RFC822 format fields,
and manipulate this information to determine the proper action
for each destination address.
The lexical and syntax analysis is carried out as a basic function
within this process, but the semantics are determined by the
contents of a configuration file for the Router.
This configuration file is required to properly initialize the
Router when it starts up, and furthermore defines functions that
analyse an address, determine how to route it (given context
information), and that can rewrite a message header address
appropriately based on various context information.

The configuration file looks like a Bourne Shell script at first glance.
There are minor syntax changes from standard @code{sh}, but the aim
is to be as close to the Bourne Shell language as is practical.
The contents of the file are compiled into a parse tree, which
can then be interpreted by the Router.
The configuration file is usually self-contained, although an easy
mechanism exists to make use of external UNIX programs when so desired.
Together with a very flexible database lookup mechanism, functions,
and address manipulation based on string-matching, and token-matching
regular expressions, the configuration file language is an extremely
flexible substrate to accomplish its purpose.
When the language is inadequate, or if speed becomes an issue,
it is possible to call built in (C coded) functions.
The interface to these functions is mostly identical to what
a standalone program would expect (modulo symbol name clashes
and return values), to ease migration of external programs to
inclusion in the Router process.

The Router makes use of environmental information to augment
the information that may be contained in the message file.
For example, the owner of the message file is the local user
who submitted the message (which fact is used to check believability
of some of the message header information), the message file
modification time is used for local submission time, and the
message file name is part of the synthesized message identification.
Other envelope information, apart from the standard sender and
recipient addresses, may be specified to augment the behavior
of the mailer.
For example, the standard library routines used to submit messages,
may include code to pass along information from the submitting user's
environment variables.

If something goes wrong when the Router processes a message
file, its action depends on the severity and type of error.
If for example there is a protocol violation of some kind,
the Router may generate a rejection message sent to the originator
of the offending message.
The Router supplies the addressee and specific diagnostic
messages corresponding to the error, and uses one of the
canned files in the @file{FORMS} directory (@file{~/forms})
for @emph{everything else}.
In particular this means such headers as @samp{From:},
@samp{Subject:} and @samp{Cc:} lines, and a generic comment
on the class of error, are all taken from a standard form.
This easily allows certain kinds of errors to be brought to
the attention of the Postmaster or other maintenance person
(by judicious use of the Carbon Copy field), and indeed
different errors may be directed to different people.

For more serious problems, the message file is filed away
in yet another @file{POSTMAN} directory (@file{~/postman}).
This directory is where the Router will put any files that
need manual attention by the Postmaster.
The Postmaster may take corrective action (usually editing
the message file), and resubmit the message file by simply
moving it (using @code{mv}) to the @file{ROUTER} directory.

If something went wrong that may correct itself at a later
time (for example if a database access indicates a temporary
failure), the message file will be relinked into a @file{DEFERRED}
directory (@file{~/deferred}).
At some later time, these deferred message files may be
resubmitted by moving them back to the @file{ROUTER} directory.
This may be accomplished by a simple cron job.
As indicated, the only problems that would cause this would be
the lack of a resource needed by the Router.
This may include out-of-space conditions on the disk, a database
access timing out or returning a server failure reply, etc.

When everything does work properly, once a control file has been
created in the @file{TRANSPORT} directory, and the message file
moved to the @file{QUEUE} directory, the job of the Router is
done for that message and it continues scanning its @file{ROUTER}
directory for more work.

@section Queue Manager and Scheduler

The process that picks up control files from the @file{TRANSPORT}
directory, is called the Scheduler.
It is a daemon that orchestrates the flow of messages out from
the mail subsystem.
To do this, it maintains an internal model of which messages need
to go where, and how, and passes the relevant information to
the transport/delivery programs that it starts up.
Various parameters associated with each transport/delivery program
are controlled by a configuration file for the Scheduler.
This configuration file is much simpler than the one for the Router,
indeed it is a simple table format.
A set of messages is selected using a channel/host specification
pattern, and associated with each pattern one must specify a startup
interval, command, and some related information, that will be used
to deliver a selected message to the appropriate addresses.
Specifying startup intervals for programs is the function which gives
the Scheduler its name.


[XX: to do: rendezvous with Scheduler from unrelated program, e.g. uucico ]

@section Transport Agents

A Transport Agent is responsible for doing the actual transport/delivery
of a given message to a selected set of addresses.
The selection of addresses is determined by the Transport Agent itself,
perhaps using information about the name of the delivery channel or
next host, passed on the command line.
The messages each Transport Agent is asked to examine are determined
by the Scheduler.
A very simple protocol is run on the standard input and standard output
of the Transport Agent, with a supervisor program (the Scheduler)
choosing which control files the Transport Agent should know about.
In turn, the Transport Agent returns status information about each
of the addresses it processes in each control file, so the Scheduler
can update its internal model of the collection of queued messages.
As well, the Transport Agent is in charge of enforcing locking of a
destination address while it is being processed, and the subsequent
status update (success, deferral, error) in the message control file.
These operations are performed in-place (and synchronously) on the
contents of a control file.

All the actions and decisions made by a Transport Agent are driven
entirely based on the contents of the control file.
There is enough information about the original message file, that
most Transport Agents will not need to reparse it.
Standard Transport Agents exist for local mail delivery, SMTP/TCP,
an error processing function, and interfacing with standard Sendmail
``mailer''s.

@section The System Environment

Simplicity is an important thread in the design and implementation
of ZMailer.
This is with the hope that a simple (not simple-minded) design will
encourage flexibility, elegance, and efficiency in the end product.
If the design is done right, the result should fit in naturally with
the UNIX environment, and will have some desirable side-effects:
code portability (to UNIX variants and to other operating systems),
and a smaller conceptual load for the person(s) maintaining the mail
subsystem.
This section is about the external interfaces to ZMailer; how it
depends on the underlying system, how it interacts with it, and
how the maintainers (the System Administrator and the Postmaster)
communicate and interact with ZMailer.

All ZMailer activity is (largely by convention) confined to two
directory hierarchies.
One is used to keep program binaries and various databases,
the other is a work area that is used when ZMailer does its job.
For various reasons, this latter hierarchy is set up to mimic
the various sections of a real postoffice, and indeed this
analogy will reappear in a few user interface situations.
The program/database locations may be spread out arbitrarily
on your system.
Unless there are good local reasons not to collect these files
in one place, the following few conventions should be kept in mind:

The program/database directory is kept in @file{/usr/lib/mail}
(first choice), or @file{/usr/lib/zmail} (in case the first choice
is already taken).
The program binaries of the Router and Scheduler portions of
ZMailer are kept here, along with all program configuration
files, and utility scripts.
The databases, including the system aliases database, are kept
in a @file{db} subdirectory.
The program binaries of all the Transport Agents are kept in
a @file{ta} subdirectory.
If you like longer names, use @file{databases} and @file{transports}
respectively.

In a mail/file server environment, mail clients only need a view
of the @file{POSTOFFICE} directory hierarchy, so that User Agents
can submit messages for processing and perhaps for the mail queue
querying program to be able to read control files.
The programs, configuration files, and databases stored under
@file{/usr/lib} are only used by the mail server machine.

[XX: did I miss anything? Is this logical? should config files
     and programs be separate?]

An upcoming section (@pxref{postoffice, , The Postoffice}) will
deal with the @file{POSTOFFICE} hierarchy in some detail.
To motivate the issues dealt with there, we will first deal with
the mechanics of sending a message.

@section User Agent support

To ease the task of interfacing directly to the ZMailer MTA,
a C library routines are provided with ZMailer.
The most important of these routines are used when submitting a message.
They implement the message file name collision avoidance protocol,
outlined in the previous subsection.
An independent routine is provided to encourage proper quoting
of the full names of users, as used in RFC822 message headers.
This is an attempt at removing any excuse for ``poetic license''
on the part of User Agents, MTAs, and other systems (e.g. USENET News),
where violation of the RFC822 specification in this regard is
a frequent irritation.  @xref{uasupport, , User Agent support}, for
details.

As mentioned earlier, information may be passed to ZMailer using
envelope header lines in the message file.
This includes a method for overriding the full name of the originating
user, as found in the GECOS field of the password file entry for the user.
It also includes a method for requesting an alternate login name
(or rather @i{local-part}, in RFC822 terminology), which is of course
subject to approval by security mechanisms in ZMailer.
In order to promote a standard way of specifying these optional values,
the message submission interface routines will seed the file with
the appropriate envelope information to be interpreted by ZMailer.
The user interface consists of the environment variables
@strong{FULLNAME} and @strong{PRETTYLOGIN}, which are accessed through
the standard @code{getenv()} routine.
These environment variables need only be set by a user for all mail
submitted through these interface routines to make use of the features.

@section Compatibility

Because ZMailer will often not be the first mailer installed on
a computer, utility programs are provided to ease the transition
between the different mail subsystems.
The programs allow the change of MTA to be largely transparent
to the User Agents, and other programs that interacted with
the previously installed MTA.
This allows conversion of such programs to be deferred to a more
convenient time.
If a slight performance penalty is acceptable, conversion may not
be necessary at all.

In a Sendmail environment, there is just one critical program that
needs to be replaced, namely the Sendmail binary itself.
The only program, that is not a User Agent, which executes Sendmail
directly, is the Rmail program (usually @file{/bin/rmail}) which
is conventionally used to transfer mail using UUCP.
To avoid certain limitations of the standard Rmail programs, and
at the same time gain performance by interacting directly with ZMailer,
a new version of Rmail comes with the ZMailer distribution.

Most of the options on Sendmail are supported -- when sensible to do so.
Among them are a way to do a Verbose mode, and do other things that
are common on Sendmail oriented environments.

@node postoffice, , ,overview
@section The Postoffice

All of the message manipulation activity of ZMailer is confined to a
directory hierarchy conventionally placed under @file{/usr/spool/postoffice}.
This name reflects the kinds of activity carried out by ZMailer under the
postoffice directory.  The subdirectories under @file{~} are:

@table @file
@item ~/deferred
a parking area for message files that cannot be processed due
to temporary absence of resources needed by the Router.
Such a situation would typically be due to a nameserver failure,
or in case of unexpected I/O errors.
Such message files can be resubmitted by simply relinking them
(use @code{mv}) to the directory scanned by the Router.
This might be done periodically by a @code{find} command.
Since the time granularity of @code{find} is rather coarse,
a utility called @code{resubmit} is included with ZMailer
to carry out exactly this task.
@item ~/forms
contains canned error and warning messages used by the ZMailer programs.
Each file contains a prototype message, the only missing information
is a destination address, and perhaps specific information that will
be supplied by whichever programs make use of the form.
In particular, specifying carbon-copy headers in these forms, allows the
postmaster to get copies of mail automatically sent to users.
Different types of messages may be carbon-copied to different people,
if there are specialized postmasters on the system.
@item ~/postman
is where ZMailer puts messages files that should be examined by
the postmaster.
Usually the postmaster is expected to take some corrective action,
and resubmit the message.
The actual reason ZMailer took this action will be found in
the Router logs.
As always, messages files can be resubmitted simply by relinking
them into the directory scanned by the Router.
@item ~/public
is the publically writable directory used by the standard message
submission routines to create a new message file.
When a message file has been properly created, it is relinked into
the directory scanned by the Router.
Empty files in this directory are often caused by improper handling
of interrupts in a User Agent.
@item ~/queue
is the final resting place of message files, after the Router has
processed them.
This is where Transport Agents finds the message files when necessary.
These files are eventually unlinked by the Scheduler.
@item ~/router
is the directory scanned by the Router for new message files.
From here, the message file goes to one of the @file{~/queue}
(nominally), @file{~/deferred}, or @file{~/postman} directories.
@item ~/transport
is the collection of pending control files.
The files here are unlinked by the Scheduler when all destinations
have been processed.
The Scheduler scans this directory periodically to detect new files
that are not yet in the internal processing, and scans them in.
@end table

In a server-client machine environment, only the server machine
needs to have this directory hierarchy.
All the clients just need a view of the postoffice, for benefit
of the User Agents.
The message file name collision avoidance protocol should work
properly across any remote filesystem.
With this setup, the various ZMailer processes should only be
run on the server machine.

@node router, scheduler, overview, top
@chapter Router

The Router is the smart half of ZMailer.
All the other parts of ZMailer essentially just carry out
instructions, as determined by the Router.
Therefore, the Router is by far the most complex part of ZMailer.
It must understand, in great detail, the structure of messages.
It has to contain logic to manipulate portions of this structure,
and, since many sites have different requirements and are in
different network and mail environments, the logic used must be
easily customizable.
At the same time it should be efficient, since it is a bottleneck
for message processing, and should cater to various services expected
by System Administrators, Postmasters, and the users of a machine.

This description of the Router will begin by explaining how
a message is submitted by User Agents and why a standard
submission interface is a good idea.
The structure of message files, and how that structure is
analysed and used, is treated next.
Then follows exposure of the mechanisms used to manipulate
this information, and especially the tools available for the
person configuring ZMailer to customize its behaviour.
A final review of the details of the control logic will explain
the reasons for various embedded behaviours of the ZMailer Router.

@section Message Submission

In the parlance of mail and message systems, ZMailer is an MTA,
a Message Transfer Agent.
It exists to process mail, similar to the function performed by
a post office.
As in the real world analogy, an MTA does not participate in
the process of composing messages and getting them into the system.
To do so, would correspond to your neighborhood postman taking
dictation of your letters, and taking them along when leaving
your house.
In reality of course, people use a variety of tools to compose
letters (quill pens, word processors, etc.), and to send them off.
This functionality is embodied in a front end to the MTA, called
a User Agent (UA for short).
The choice of UA is a very personal one, and is usually not critical
to the basic process of composing a message, sending it off, and
getting it delivered properly.

Even though there may be many UA's in use on a computer system,
there is usually only one MTA.
The exceptions to this rule usually have to do with limitations
in an MTA's capabilities.
For example, a computer that can transfer mail using both
the X.400 protocols, and the Internet protocols, may need
two different MTAs to cover both protocol suites.
Oftentimes, one of the two MTAs is a primary mailer, and
takes care of all decision-making.
The other mailers would then be treated by the primary MTA
as a means for delivering messages to particular destinations,
and these secondary mailers would be configured to punt any
non-trivial traffic to the primary MTA.
Both in the case of a User Agent, and in the cases of alternate
MTAs, there must be a way to inject messages into the mail subsystem.

When you want to mail a letter, what do you do?
Well, you drop it in a mailbox somewhere.
ZMailer accepts messages the same way: you drop a file,
containing your message, into a special submission directory.
Like a postman (although rather more frequently), ZMailer scans
this directory to pick up the new messages, and processes them.
What happens to a message from then on is interesting in its
own right of course, but presently we shall focus on how the
message gets from a user into the mail subsystem.

The simplest way to ``drop a file into a directory'' is
of course to actually edit a file in that directory.
However, a program scanning such a directory would not be
able to tell when you had finished writing your message
and stopped editing the file.
To do so, would require cooperation between the scanning
program and all the programs that could conceivably create
a file a portion at a time.
The next most obvious method is to edit the message file
in another location, and then simply copy or link it into
the special message submission directory.
Indeed, this is almost exactly the mechanism used.
Actually, the copying program happens to be among those
programs that may construct a new file piece by piece
(a disk block at a time).
For large files there is a vulnerable window between when
the copy starts and when it finishes.
If the submitter is unlucky, the message may be processed
by the scanning program (the ZMailer Router) before it is
completely written out.
In fact this is not a problem due to the implementation
policy of relinking message files.
The real problem is if a partial message is completely processed
and delivered and removed, with a corrupt message body.

To avoid problems, the only acceptable method is to make
the complete message available to the Router at once.
This is done by linking the message file into the directory
being scanned.
Due to the semantics of hard links and the UNIX filesystem,
doing this requires that the message file is created on the
same filesystem as the submission directory.
ZMailer provides a publically writable directory, specifically
for the purpose of creating message files before they are
relinked into the submission directory.
In fact, the submission directory itself is publically writable, so
users can do this relinking themselves (e.g. with the @code{mv} command).

There are some security concerns with this approach.
Because both these directories are writable, it is conceivable
that a malicious user can cause problems for other users,
for example remove their message files, or read or alter
them if the file permissions allow.
The solution to the latter problem, is obviously to ensure that
file permissions do not allow people other than the originating
user to access a message file.
There are two solutions to the former problem; one is to maintain
ignorance of the contents of the various directories, the other,
better, method is to use a feature introduced in 4.3BSD -- setting
the sticky bit on the directories.
The semantics of the sticky bit on a directory is to only allow
the owner of a file to unlink it from a generally writable directory.
If this function is not available, read permissions to the submission
directory can safely be removed if only the standard message submission
routines are used by the User Agents.
This leaves us having to find a way to secure the directory used for
creating files.
Read permission to it cannot be withdrawn if the aforementioned
standard routines are used (they are described later).
Other things can be done, but obviously ``ignorance'' is not
a reliable way of enforcing security.
Perhaps if you notice the analogy with the common usage of @file{/tmp}
for various intermediate files, you will not consider security any more
of a problem in this case.
My recommendation would be to treat these directories the same way you
treat @file{/tmp}.
That is, if the directory sticky-bit semantics are available, use
that feature.
If not, try trusting your users enough to not cover up the directories.
You should note that the possible danger is confined to removing
message files.
There is no way for a user to forge the origin of a mail message,
since all validation of message origin is based on the ownership
of the message file.
Trusted user id's may of course supply a different origin address.

There is only one other significant problem with this approach, which
is the potential for name clashes of files in each of the two directories.
This problem can be only be solved if all User Agents cooperate in using
the exact same collision avoidance or collision resolution technique.
What is really needed, by all the various ZMailer components operating
on a message, is a way to get and hold a lock on the message.
There are various ways to achieve this, and a kernel-based locking
mechanism may seem appropriate in certain situations.
However, given the realization afforded by an overview of the structure
of ZMailer, it becomes clear that each message goes through states
corresponding to the current processing stage.
Instead of storing the state in the message file (or some other
location associated with a particular message file), the state is
encoded in the current location of a message file.
For example, if the message file is in the submission directory,
it means it is waiting for the ZMailer Router to process it.
With this solution in hand, there still remains the original
matter of avoiding file name clashes.

The best way of avoiding name clashes is to generate names that
cannot clash.
The only obvious unique property of a file is its inode number.
Therefore, if one uses the inode number in the file name itself,
name clashes will be completely avoided.
This is a truism if the files are created on the same filesystem,
and of course breaks down if that assumption is invalid.
There is just one minor problem: the inode number of a file is
not known until the file is actually created.

To resolve that Catch-22 situation, the message file must be
created under one name, and then immediately renamed to its
guaranteed unique name, using its inode number.
The first name for a message file can safely be chosen by
the same method used to find names for temporary files in @file{/tmp}.
Now all the problems are solved with only two important assumptions:
all message files are created by this same mechanism, and all message
files are created on the same filesystem.

ZMailer will work with files that have been manually created and moved
around, although this should only be done routinely by mail system
maintainers.

@section Message File Format

As mentioned, the Router picks up message files from a specific directory.
Normally, message file names can be arbitrary valid file names, and indeed
this is convenient when debugging.
However, because the Router daemon scans its own current directory,
miscellaneous output from the Router process may show up in this
directory (e.g. profiling data, or core dumps (unthinkable as that is)).
Furthermore, it is useful to be able to hide files from the Router
scanning (indeed the Router may wish to do so itself).

When the Router process is scanning for message files then, it
only considers at file names that have a certain format.
Specifically, the message file name must start with a digit.
This method was chosen to accomodate the message file names, as generated
by the standard submission interface library routines, which will be strings
of digits representing the message file's inode number.

A message file contains three sections: the message envelope, the message
header, and the message body (in that order).
The message body is separated from the previous sections by a blank line.
The message body may be empty, and either of the message envelope or message
header may be empty.
The restriction on the latter situation, is that one of those sections
must contain destination information for the message.

The message envelope and the message header have very similar syntax.
The only difference is that while the message header must adhere
to RFC822, the message envelope header fields are terminated by
whitespace (@samp{ }) instead of a colon (@samp{:}).
The semantics of the two message file sections is quite different,
and will be covered later.

The header fields recognized by ZMailer in the message envelope are:

@table @code
@item channel @i{word}
@ifinfo
sets the channel corresponding to the message origin(*)
@end ifinfo
@iftex
sets the channel corresponding to the message origin@dag{}
@end iftex
@item from @i{address}
@ifinfo
a source address(*)
@end ifinfo
@iftex
a source address@dag{}
@end iftex
@item fullname @i{phrase}
sets the full name of the local sender
@item loginname @i{local-part}
requests using this mail id for the local sender
@item rcvdfrom @i{domain}
@ifinfo
sets the host the message was received from(*)
@end ifinfo
@iftex
sets the host the message was received from@dag{}
@end iftex
@item to @i{address-list}
a destination address list
@item user @i{local-part}
@ifinfo
sets the user the message was received from(*)
@end ifinfo
@iftex
sets the user the message was received from@footnote{@dag}{This
is a privileged field.
That is, the action will only happen if ZMailer trusts the owner
of the message file (@pxref{security, , Security}).}
@end iftex
@item via @i{word}
for RFC822 Received: header to be generated
@item with @i{word}
for RFC822 Received: header to be generated
@end table

@ifinfo
The (*)'s beside the descriptions indicate this is a privileged field.
That is, the action will only happen if ZMailer trusts the owner of
the message file (@pxref{security, , Security}).
@end ifinfo
As with a normal RFC822 header, other fields are allowed (though they
will be ignored), and case is not significant in the field name.
The Router will do appropriate checks for the fields that require it.

With this knowledge, we can now appreciate the minimal message file:

@group
@example
--------------------
to bond

--------------------
@end example
@end group

This will cause an empty message to be sent to @samp{bond}.
A slightly more sophisticated version is:

@group
@example
--------------------
from m
to bond
via courier
From: M
To: Bond
Subject: do get a receipt, 007!

You are working for the Government, remember?
--------------------
@end example
@end group

Notice that there is no delimiter between the message envelope and the
message header.  A more sophisticated example in the same vein:

@group
@example
--------------------
from ps/d-ops
to <007@@sis.mod.uk>
From: M <d-ops@@sis.mod.uk>
Sender: Moneypenny <ps/d-ops@@sis.mod.uk>
To: James Bond <007@@sis.mod.uk>
Subject: where are you???!
Classification: Top Secret
Priority: Flash

We have another madman on the loose.  Contact "Q" for usual routine.
--------------------
@end example
@end group

If the @samp{Classification} header is paid attention to in ZMailer,
this requires that the Router recognize it in the message header,
and take appropriate action.
In general the Router can extract most of the information in the message
header, and make use of it if the information is lacking in the envelope.
The envelope headers in the above message are superfluous, since the same
information is contained in the message header.
Using the following envelope headers would be exactly equivalent to using
the ones shown above (assuming the local host is @samp{sis.mod.uk}):

@group
@example
--------------------
From Moneypenny <ps/d-ops@@sis.mod.uk>
To James Bond <007@@sis.mod.uk>
...
--------------------
@end example
@end group

ZMailer will extract the appropriate address information from whatever
the field values are, as long as they obey the defined syntax (indicated
in the list of recognized envelope fields above).
ZMailer will complain in case of unexpected errors in the envelope headers.

The message body is not interpreted by ZMailer itself.
As far as the Router is concerned, it can be arbitrary data.
However, certain Transport Agents may require limitations on
the message body data.
For example, the SMTP only deals with ASCII data with a small
guaranteed line length.

@section Header Scanning and Parsing

Message header and envelope is scanned according to the lexical rules of
RFC822 (and RFC976), and parsed according to the grammar rules of RFC822.
RFC976 compatibility requires that the @samp{!} and @samp{%} characters
be treated as specials (just like @samp{.} and @samp{@@}).
This behavior is enabled at compile time (by defining the @code{RFC976}
preprocessor symbol), and is indeed enabled by default.

XXX: The only divergence from RFC822/RFC976 syntax is that comments are not
allowed in certain locations within addresses, and comments and quoted
strings may not span line boundaries.
Neither of these are design limitations, they will disappear before final
release.
All other RFC822 constructs are properly recognized and supported.

The mentioned RFC documents serve to describe ZMailer behavior with respect
to lexical scanning, tokenization, and parsing.
In summary, based on the class of each character, a token stream is
synthesized for each header.
Various headers have defined semantics (e.g. the @samp{To:} header contains
an address list), which drive the parse of the token stream for that header.
The headers that have specific semantics to the Router are:

@group
@example
Field name       RFC822 Syntax description      Class
-------------------------------------------------------------------
channel                 word                    Envelope
fullname                phrase                  Envelope
loginname               addr-spec               Envelope
rcvdfrom                domain                  Envelope
user                    mailbox                 Envelope
via                     word                    Envelope
with                    word                    Envelope

bcc                     #address                Recipient
cc                      1#address               Recipient
date                    date-time
encrypted               1#2word
errors-to               1#address               Sender
from                    1#mailbox               Sender/Envelope
in-reply-to             *(phrase | msg-id)
keywords                1#phrase
message-id              msg-id
received                received
references              *(phrase | msg-id)
reply-to                1#address               Sender
return-path             route-addr              Sender
return-receipt-to       1#address               Sender
sender                  mailbox                 Sender
to                      1#address               Recipient/Envelope
@end example
@end group

All the fields mentioned above are parsed by the Router.
Some (e.g. @samp{In-Reply-To:}) are just parsed and not interpreted.
Since the Router will complain about format violations, this is
a way of enlightening people about what a particular field is not
supposed to contain.

The @samp{Date:} and @samp{Received:} fields are interesting in that
it is rather unusual for an RFC822 mailer to parse these fields.
Indeed, whether or not they are parsed depends on the definition of
compile-time preprocessor symbols (@code{CANON_DATE} and
@code{CANON_RECEIVED} respectively).
If a @samp{date} header is parsed successfully, it will be printed using
proper RFC822 date-time syntax when the message is delivered.
For this to be useful, the date string parse in the Router must be
rather flexible to recognize the endless variety of formats that exist,
and it is.

The intention with parsing @samp{Received:} headers is to prepare for
the possibility of using the information in the trace headers to aid
the routing algorithm.
For now, parsed trace headers are output in a canonical format that
follows RFC822, similar to what happens with parsed @samp{date} headers.
As long as the information is not used, there is no common reason to
enable this feature.
[XX: It should perhaps be possible to select these features on
    a per-message basis. any thoughts on this?]

Some of the header field names are tagged with what kind of addresses
that header field contains.
This information is used when searching for destination addresses when
there are none specified in the envelope, and to know which headers
contain addresses that must be sent through the address manipulation
mechanisms of the Router.

@section Router Activities

The ZMailer Router has three basic functions that it must carry out on each
message:

@itemize @bullet
@item
Determining how to deliver a message to its destinations given in the
message envelope.
@item
Rewriting message header and envelope addresses to accommodate the standards
imposed by the method of delivery and the destination.
@item
Ensuring only properly formatted and standard-conforming (RFC822) messages
leave the local system.
@end itemize

For everything but the syntax and semantics of addresses, the last goal is
achieved by mechanisms internal to the Router.
This is a reasonable approach since a standard is not something that adapts
to local conditions.
However, when pursuing the first two goals, many sites have found
it invaluable to be able to modify the behavior of the routing function,
and of the address rewriting function, to take local idiosyncrasies
into account.
The importance of this ability is very apparent to sites in complicated
environments.
Since ZMailer was partially motivated by the inadequacies of other
mailers in such an environment, much effort has gone into the design
of the configurable parts of the Router behavior.
The wired logic of the Router is treated in a later subsection
(@xref{sequencer, , Router Control Flow}).
Presently, we shall examine how routing and address manipulation
is carried out, and the Router facilities which support these activities.

@section Routing Model

For routing purposes, one wants to derive three pieces of information
from an address: where to send the message, how to send the message,
and what to tell the immediate destination of the message about it.
This is the information needed to properly transmit or deliver
a message to its next destination.

The mechanism used to transmit a message may be regarded as a conduit
(pipe, channel, circuit, etc.) between the local MTA and a remote MTA.
In ZMailer terminology, such a conduit is called a Channel.
A Channel is just a tag associated with a destination address for
the message, and is used by the Scheduler to manage delivery of the message.
Thus, a Channel is a concept (i.e. not associated with any particular
program), and may be serviced by one or more Transport Agents.
As far as the Scheduler is concerned, it is an uninterpreted
classification of the message.
For example, if there are different physical links to a remote MTA,
different Transport Agent programs may serve the same Channel.

The Channel, or rather the Transport Agents serving a Channel,
may need to know which remote MTA to deliver the message to.
This is most often a hostname of a neighbouring host on a common network.
If the Channel can only have one destination host (for example the local
delivery Channel), a destination is superfluous.
By convention, the Router will translate null destinations into
the symbol @samp{-} in a message control file.

The remote MTA will need to know what to do with the message, in
the form of some envelope information.
In RFC822, this information is embodied in an address for further
delivery with respect to the remote host.

The Router must determine this triple (@i{channel}, @i{next-host},
@i{next-address}) for every address in the envelope, including the
(single) origin address to be able to verify origin.
If not for security, then to make sure that a proper RFC822 address
was specified for the sender, and that a bogus address form is not
passed on.
To do this, the Router will call a function that takes an address as
its argument and returns a triple.
This function may be completely specified in a configuration file read
by the Router, and its task is termed address resolution or routing.

While the @code{router} function rewrites envelope address as appropriate,
there must also be a way to rewrite message header addresses.
In Sendmail, this was done based entirely on which ``mailer''
(similar to a ZMailer Channel) the message was sent through.
To do more sophisticated rewriting was not possible due to a complete
lack of other information.
If one wished to do different manipulations depending on the final
destination of a message for example, it was almost impossible to do so
(no variables or control flow in Sendmail rulesets).
It was also impossible to do address manipulation or validity checking
based on the origin of the message, since no such information was available.

The ZMailer Router remedies these and other shortcomings in several ways:
the configuration language has control flow and variables, and the decision
of how to rewrite each address is carried out with access to all the needed
sender and recipient information.
The word ``decision'' is used on purpose to indicate that the choice
of rewriting method is divorced from the actual message header address
rewriting process.
What happens is that for each recipient, the Router calls a function
passing the triples derived from the sender and the recipient address
as arguments.
The return value from this @code{crossbar} function (so named because
a crossbar switch is the closest image, that came to mind, of what it
does) includes the name of a function that is to be used for rewriting
the message header addresses.
This returned function is then called separately with all the addresses
in the message header, and the results will be incorporated in the message
header for the destination corresponding to the recipient triple.
At the same time, while the routing function does generic resolution of
an address into its corresponding triple, the crossbar function may modify
the sender and recipient triples if necessary, and so serves as a cleanup
or filtering function for the routing information.
The crossbar function can also be completely specified in a configuration
file read by the Router.

The names (determined at compile-time) and interface specifications for
the routing and crossbar functions, are the only crucial ``magical''
things one needs to contend with in a proper Router configuration.
The syntax and semantics of the configuration file's contents are dealt
with in the following subsection.
The details of the two functions introduced here are specified after that,
once the necessary background information has been given.

@section Configuration File Programming Language

Whenever the Router process starts up, its first action is to read its
configuration file.
The configuration file is a text file which contains statements
interpreted immediately when the file is read.
Some statements are functions, in which case the function is defined
at that point in reading the configuration file.
The purpose of the configuration file is to provide a simple way
to customize the behavior of the mailer, and this is primarily
achieved by defining the @code{router} and @code{crossbar} functions.
For these to work properly, some initialization code and auxiliary
functions will usually be needed.

At first sight, a configuration file looks like a Bourne shell script.
Indeed, the ideal is to duplicate the functionality, syntax, and to a large
degree the semantics, of a shell script.
Therefore, the configuration file programming language is defined in terms
of its deviation from standard Bourne shell syntax and semantics.
The present differences are:

@itemize @bullet
@item
No @code{for}, @code{while}, and @code{repeat} statements, no pipes,
or I/O redirection.
@item
Case statement labels have no following @samp{)}, i.e. use

@group
@example
case foo in
pattern         action ;;
esac
@end example
@end group

@noindent
instead of

@group
@example
case foo in
pattern)        action ;;
esac
@end example
@end group
@item
Case label patterns use V8 (Eighth Edition UNIX) regular
expression syntax (@code{egrep}-like).
@item
Functions are allowed, parameter lists are allowed.
If not enough arguments are present in a function call to exhaust
the parameter list, the so-far unbound parameter variables are bound
to @samp{} (the empty string) as local variables.
For example, this is the identity address rewriting function:

@group
@example
null (address) @{
        return $address         # surprise!
@}
@end example
@end group
@item
Multiple-value returns are allowed.
The @code{return} statement can be used to return a non-@samp{}
value from a function.
The following are all legal @code{return} statements:

@group
@example
return
return $address
return $channel $@{next_host@} $@{next_address@}
@end example
@end group
@item
Variables are dynamically scoped, the only local variables are the ones
in a function's parameter list.
Only the first value of a multiple-value return may be assigned
to a variable.
All values are strings, so no type information, checking, or declaration,
is necessary.
@item
Quoting is a bit stilted.
All quotes (double-, single-, back-), must appear in matching pairs
at the beginning and end of a @i{word}.
Single quotes are not stripped, double quotes cause the enclosed
character sequence to be collected into a quoted-string RFC822 token.
For example, the statement:

@example
foo `bar "`baz`"`
@end example

is evaluated as @code{(apply 'foo (apply 'bar (baz)))}.
@item
The forms @code{$@{variable:=value@}}, @code{$@{variable:-value@}}, and
@code{$@{variable:+value@}} are supported.
The special form @code{$@{string:relation@}} returns the value of
@code{relation(string)}, implementing a database lookup function.
@item
Patterns (in case labels) are evaluated once, the first time they are
encountered.
@item
At the end of a case label, the sequentially next case labels of the same
case statement will be tried for successful pattern matching (and the
corresponding case label body executed).
The only exceptions (apart from encountering a return statement) are:
@table @code
@item again
a function which retries the current case label for a match.
@item break
continues execution after the current case statement.
@end table
@item
Various standard Bourne shell functions do not exist built in.

@c @item
@c The function @code{import} must be used to declare a unix program to be
@c accessible to the config file code.
@c This allows development using an existing utility, and integration into the
@c router of the same functionality can be delayed until the need is proven.
@c For example use the statement:
@c 
@c @example
@c import hostname /bin/hostname
@c @end example
@c 
@c to do the obvious.
@c Programs defined in this manner will have the message file on their
@c standard input when they are executed.
@end itemize

There are currently only two entry-points (i.e. magic names known to the
Router code) in the configuration file, namely the @code{router} and the
@code{crossbar} functions.

The @code{router} function is called with an address as argument, and
returns a triple of (@i{channel}, @i{host}, @i{user}) as three separate
values, corresponding to the channel the message should be sent out on
(or, the router function can also be called to check on who sent a message),
the host or node name for that channel (null if local delivery), and the
address the receiving agent should transmit to.

The @code{crossbar} function is in charge of rewriting envelope addresses,
selecting message header address munging type (a function to be called
with each message header address), and possibly doing per-message logging
or enforcing restrictions deemed necessary.
It takes a sender-triple and a receiver-triple as arguments (six parameters
all together).
It returns the new values for each element of the two triples, and in
addition a function name corresponding to the function to be used to
rewrite header addresses for the specific destination.
If the destination is to be ignored, returning a null function name will
accomplish this.

There is one more magic symbol the Router knows about, which is
(optionally) defined by the configuration file.
That is the name of the definition of the alias database, a protocol
which will be dealt with in the subsection explaining the database
lookup mechanism hinted at above.

The Router has several built in (C coded) functions.
Their calling sequence and interface specification is exactly the same
as for the functions defined in the configuration file.
Some of these functions have special semantics, and they fall into three
classes, as follows:

Functions that are critical to the proper functioning of the configuration
file interpreter:

@group
@table @code
@item return
returns its argument(s) as the value of a function call
@item again
repeats the current case label
@item break
exits a case statement
@end table
@end group

Functions that are necessary to complete the capabilities of the interpreter:

@group
@table @code
@item import
defines a function name that refers to an external program
@item relation
defines a database to the database lookup mechanism
@item sh
an internal function which runs its arguments as @file{/bin/sh} would
@end table
@end group

Non-critical but recommended functions:

@group
@table @code
@item getzenv
retrieves global ZMailer configuration values
@item echo
emulates @samp{/bin/echo}
@item exit
aborts the Router with the specified status code
@item hostname
internal function to get and set the system name
@item trace
turns on selected debugging output
@item untrace
turns off selected debugging output
@item [
emulates a subset of @samp{/bin/test} (a.k.a. @samp{/bin/[}) functionality
@end table
@end group

The @code{relation} function is described in a later section
(@xref{databases, , Database Interface}), and the @code{trace} and
@code{untrace} functions are described in connection with debugging
(@xref{routerlog, , Logging}).

The @code{hostname} function requires some further explanation.
It is intended to emulate the BSD UNIX @file{/bin/hostname}
functionality, except that setting the hostname will only set
the Router's idea of the hostname, not the system's.
Doing so will enable generation of @samp{Message-Id} and @samp{Received}
``trace'' headers on all messages processed by the Router.
It is done this way, since the Router needs to know the official
domain name of the local host in order to properly generate these
headers, and this method is cleaner than reserving a magic variable
for the purpose.
The Router cannot assume the hostname reported by the system is a properly
qualified domain name, so the configuration file may generate it using
whichever method it chooses.
If the hostname indeed is a fully qualified domain name, then:

@example
hostname `hostname`
@end example

@noindent
will enable generation of trace headers.

Finally, note that a symbol can have both a function-value and a
string-value.  The string value is of course accessed using the $-prefix
convention of the Bourne shell language.

@section Address Manipulation

Most of the flexibility of Sendmail derives from its production-rule
model for address rewriting.
Very loosely, the concepts of rulesets in Sendmail correspond to
the functions of the ZMailer Router configuration file programming
language, and the rules themselves correspond to the case label
bodies of the case statements in our language.

Addresses are represented as string values in this language, no different
from any other strings.
Therefore, addresses can be assigned as the value of a variable, or passed
as an argument to a function.
The way to do address rewriting is to modify the value of a variable chosen
to contain the @emph{current} address (in the sense of the Sendmail rewriting
process).
In keeping with the production rule model, much of the address rewriting
is typically done within case statements, whose semantics have been tailored
for this activity.

A @code{case} statement in the configuration file language has
almost the same syntax as the Bourne shell case statement.
However, its semantics are different, in that it is similar to
the philosophy of (Sendmail) rulesets.
That is, the normal action is for an address to ``enter'' at
the top, and for each case label (rule) that matches, the case
label body (action) is executed.
This is carried on sequentially for each case label (rule-action
pair) in the case statement (rule set), unless the normal continuation
action is modified by a control statement.

As opposed to Sendmail, where a rule is retested until it fails,
a case label pattern is only retested if the case label body calls
the special function @code{again}.
This change was made because it is frequently a waste of time to retest
a pattern match when one has just modified the string to be matched against.
Sendmail does provide a way to continue to the next rule-action pair,
but since it is not the default behavior, it is often not used in many
of the places it should be used.
As a way of reducing the consequent waste of time, the default behavior
has been changed.

The other special function that is specific to case statements,
is @code{break}.
It is used with the same semantics as if within a C language looping
construct or switch statement, i.e. to exit the case statement and
continue with the statement after it.
Of course, a @code{return} statement will completely return from
its enclosing function at any time.

Case conditions usually are not just a simple constant string; they will
usually contain a variable expansion and perhaps a function call.
The value of such a condition changes as the variable(s) it depends
on changes.
When doing repeated case label pattern matching with the condition
string value, it would be rather unsavory to reevaluate the condition
expression every time.
If no antecedent variable has changed value, obviously the expression
will not change its value either.
To avoid this unnecessary effort, the case condition is only reevaluated
when any variable it depends on has been assigned to, and then of course
only when the current expression value is actually needed.

Some final, and very important points:  Even though the case label patterns
look like normal regular expressions that one can find in editors and other
system utilities, the pattern matching in the Router is token-based, rather
than character-based.
The tokens are of course the RFC822 tokens scanned from the value of
the condition expression.
This is done to avoid surprises from simplistic patterns, and to cut
down on unnecessary verbosity in describing an address when using
the normal regular expression semantics.
Another thing that helps the matter, is that all case label patterns
are anchored at the beginning and end of the string.
An anchored pattern easily simulates an unanchored pattern, but not
vice versa.
In patterns, parentheses are used to group a number of alternates,
and are also used to bracket portions of the pattern, so the corresponding
tokens in the matched string can later be referred to.
To avoid introducing another special character (backslash, conventionally
used to refer to selected portions of the matched string), the semantics
of the $-prefix notation are extended to handle this need.
If a @samp{$} is followed by a digit N, this is expanded as the value of
the portion of the matched string selected by the N'th group of parentheses
in the pattern.

To give an idea of how a case statement looks, here is a code fragment:

@group
@example
case $hostname in
.+\.(edu|gov|mil|oth|org|net|ca|dk|uk)      # add toplevels as you please
        break ;;                            # do nothing
.*      hostname = $hostname.$orgdomain ;;  # default domain
esac
@end example
@end group

@node databases, security, , router
@section Database Interface

Many of the decisions and actions taken by configuration file code
depend on the specifics of the environment the MTA finds itself in.
So, not just the facts that the local host is attached to (say)
the UUCP network and a Local Area Net are important, but it is
also essential to know the specific hosts that are reachable by
this method.
Hardcoding large amounts of such information into the configuration
file is not practical.
It is also undesirable to change what is really a program
(the configuration file), when the information (the data)
changes.

The desirable solution to this data abstraction problem is to provide
a way for the configuration file programmer to manage such information
externally to ZMailer, and access it from within the Router.
The logical way to do this is to have an interface to externally
maintained databases.
These databases need not be terribly complicated; after all the simplest
kind of information needed is that a string is a member of some collection.
This could simply correspond to finding that string as a word in
a list of words.

However, there are many ways to organize databases, and the necessary
interfaces cannot be known in advance.
The Router therefore implements a framework that allows flexible interfacing
to databases, and easy extension to cover new types of databases.

To use a database, two things are needed: the name of the database, and
a way of retrieving the data associated with a particular key from that
database.
In addition to this knowledge, the needs of an MTA do include some
special processing pertinent to its activities and the kind of keys
to be looked up.

Specifically, the result of the data lookup can take different forms:
one may be interested only in the existence of a datum, not its value,
or one may be looking up paths in a pathalias database and need to
substitute the proper thing in place of @samp{%s} in the string returned
from the database lookup.
It should be possible to specify that this kind of postprocessing should
be carried out in association with a specific data access.
Similarly, there may be a need for search routines that depend on
the semantics of keys or the retrieved data.
These possibilities have all been taken into consideration in
the definition of a @i{relation}.
A relation maps a key to a value obtained by applying the appropriate
lookup and search routines, and perhaps a postprocessing step,
applied to a specified database that has a specified access method.

The various attributes that define a relation are largely independent.
There will of course be dependencies due to the contents or other
semantics of a database.
In addition to the features mentioned, each relation may optionally
have associated with it a subtype, which is a string value used to
communicate to the lookup routine which table of several in a database
one is interested in.

There are no predefined relations in the Router.
They must all be specified in the configuration file, before first use.
This is done by calling the special function @code{relation} with
various options, as indicated by the usage string printed by the relation
function when called the wrong way:

@example
Usage: relation -t dbtype [-f file -s# -b|n -l/u -d driver] name
@end example

The @samp{t} option specifies one of several predefined database types, each
with their specific lookup routine.  It determines a template for the set of
attributes associated with a particular relation.  The predefined database
types are:@refill

@group
@table @code
@item hostsfile
@file{/etc/hosts} lookup using @code{gethostbyname()}.
@item unordered
the database is a text file with key-datum pairs on each line,
keys are looked up using a sequential search.@refill
@item ordered
the database is a text file with key-datum pairs on each line, keys are
looked up using a binary search in the sorted file.@refill
@item dbm
the database is in DBM format (strongly discouraged).
@item ndbm
the database is in NDBM (new DBM) format.
@item bind
the database is the BIND nameserver, accessed through the standard resolver
routines.@refill
@end table
@end group

A subtype is specified by appending it to the database type name
separated by a slash.
For example, specifying @code{bind/mx} as the argument to
the @samp{t} option will store away @samp{mx} for reference by
the access routines whenever a query to that relation is processed.
The subtypes must therefore be recognized by either the database-specific
access routines (for translation into some other form), or by
the database interface itself.

For @code{unordered} and @code{ordered} database types, the datum
corresponding to a particular key may be null.
This situation arises if the database is a simple list, with one
key per line and nothing else.
In this situation, the use of an appropriate post-processor option
(e.g. @samp{b}) is recommended to be able to detect whether or
not the lookup succeeded.

The @samp{f} option specifies the name of the database.
This is typically a path that either names the actual (and single)
database file, or gives the root path for a number of files comprising
the database (e.g. @code{foo} may refer to the NDBM files @file{foo.pag}
and @file{foo.dir}).
For some types of the databases there is no practical access to any
particular database file, these include @code{hostfile}, and @code{bind}.
The use of the @code{dbm} format is strongly discouraged, since
a portable program can only have a single DBM database associated
with it.
For YP (nee, ``NIS'') database the optional file defines NIS-domain.

The @samp{s} option specifies the size of the cache.
If this value is non-zero (by default it is 10), then an LRU cache
of this size is maintained for previous queries to this relation,
including both positive and negative results.

The @samp{b} option asks that a postprocessor is applied to the database
lookup result, so the empty string is returned from the relation query
if the database search failed, and the key itself it returned if the search
succeeded.
In the latter case, any retrieved data is discarded.
The option  letter is short for Boolean.

The @samp{n} option asks that a postprocessor is applied to the database
lookup result, so the key string is returned from the relation query if
the database search failed, and the retrieved datum string is returned if
the search succeeded.
The option letter is short for Non-Null.

The @samp{l} option asks that all keys are converted to lowercase before
lookup in the database.
This is mutually exclusive with the @samp{u} option.

The @samp{u} option asks that all keys are converted to uppercase before
lookup in the database.
This is mutually exclusive with the @samp{l} option.

The @samp{d} option specifies a search routine.
Currently the only legal argument to this option is @code{pathalias},
specifying a driver that searches for the key using domain name lookup rules.

The final argument is not preceeded by an option letter.
It specifies the name the relation is known under.
Note that it is quite possible for different relations to use
the same database.

Some sample relation definitions follow:

@group
@example
if [ -f /etc/named.boot ]; then
  # relation -nt bind,cname -s 100 canon  # T_CNAME canonicalize hostname
    relation -nt bind,any -s 100 canon    # T_ANY canonicalize hostname
    relation -nt bind,uname uname         # T_UNAME UUCP name
    relation -bt bind,mx neighbour        # T_MX/T_WKS/T_A reachability
    relation -t  bind,mp pathalias         # T_MP pathalias lookup
else
    relation -nt hostsfile -s 100 canon   # canonicalize hostname
    relation -t unordered -f $MAILBIN/db/hosts.uucp uname
    relation -bt hostsfile neighbour
    relation -t unordered -f /dev/null pathalias
fi
@end example
@end group

The above fragment defines a set of relations that can be accessed in
the same way, using the same names, independent of their actual definition.

@group
@example
# We maintain an aliases database in the following format. Note: the
# 'aliases' db name is magic to the internal alias expansion routines.
if [ -f $MAILBIN/db/aliases.dat ]; then
    relation -t ndbm -f $MAILBIN/db/aliases aliases
else
    relation -t ordered -f $MAILBIN/db/aliases.idx aliases
fi
@end example
@end group

As the comment says, the relation name @code{aliases} has special
significance to the Router.
Although the relation is not special in any other way (i.e. it can
be used in the normal fashion), the semantics of the data retrieved
are bound by assumptions in the aliasing mechanism.
These assumptions are that key strings are local-name's, and
the corresponding datum gives a byte offset into another file
(the root name of the aliases file, with a @file{.dat} extention),
which contains the actual addresses associated with that alias.
The reason for this indirection is that the number of addresses
associated with a particular alias can be very large, and this
makes the traditional simple database formats inadequate.
For example, quick lookup in a text file is only practical
if it is sorted and has a regular structure.
A large number of addresses associated with an alias makes
the structuring a problem.
The situation for DBM files and variations have problems too,
due to the intrinsic limits of the storage method.
The chosen indirection scheme avoids such problems without
loss of efficiency.

Finally, some miscellaneous definitions that illustrate various
possibilities:

@group
@example
relation -t unordered -f /usr/lib/news/active -b newsgroup
relation -lmt ndbm -f $MAILVAR/db/active -b newsgroup
relation -t unordered -f /usr/lib/uucp/L.sys -b ldotsys
relation -t ordered -f $MAILBIN/db/hosts.transport -d pathalias transport
@end example
@end group

Here, the first two illustrate convenient coincidences of format, and the
last definition shows what might be used if outgoing channel information
is maintained in a pathalias-format database (e.g. @samp{bar     smtp!bar}
means to send mail to @samp{bar} via the SMTP channel).

@subsection Using a Pathalias Database

Accessing route databases is a rather essential capability for a mailer.
At the University of Toronto, all hosts access a centrally stored database
through a slightly modified nameserver program.
If such a setup is not practical at your site, other methods are available.
The most widespread kind of route database is produced by
the @code{pathalias} program.
It generates key-value pairs of the forms:

@group
@example
uunet                ai.toronto.edu!uunet!%s
.css.gov             ai.toronto.edu!uunet!seismo!%s
@end example
@end group

@noindent
which when queried about @samp{uunet} and @samp{beno.css.gov} correspond
to the routes:

@group
@example
ai.toronto.edu!uunet
ai.toronto.edu!uunet!seismo!beno.css.gov
@end example
@end group

Notice that there are two basic forms of routes listed:
routes to UUCP node names and routes to subdomain gateways.
Depending on the type of route query, the value returned from
a pathalias database lookup needs to be treated differently.
For now, this may be accomplished by a configuration file relation
definition and interface function as shown:

@group
@example
relation -t ndbm -f $MAILBIN/uuDB -d pathalias padb

# pathalias database lookup function
padblookup (name, path) @{             # path is a local variable
        path = $@{$name:padb@}
        case "$path" in
        ((.+)!)?([^!]+)!%s
                if [ $3 == $name ]; then
                        path = $2!$3
                else
                        path = $2!$3!$name
                fi
                ;;
        .*%s.*  echo illegal route in pathalias db: $path
                ;;
        esac
        return $path
@}
@end example
@end group

This is actually a simplistic algorithm, but it does illustrate the method.
The lookup algorithm used when the @samp{-d} flag is specified in the
relation definition command is rather simple; it doesn't test various case
combinations for the keys it tries.
Therefore, the keys in the pathalias output data should probably
be converted to a single case, and the @samp{-l} or @samp{-u} flag
given in the relation definition.

@section Mail Forwarding

Although more interesting and useful models exist, the mail forwarding
functionality of ZMailer has been designed to generally emulate the
interface and behaviour of Sendmail.
The mechanisms that accomplish this are likely to be generalized
in a future version.

If a relation named @code{aliases} is defined by the configuration file,
then the data returned by a lookup in that database is assumed to be
a printed decimal representation of the byte offset of the definition
of the alias in a separate file.
In other words, the @code{aliases} relation associates a particular
local-part, with an index into another file that contains the actual
alias definition.
The name of this other data file is constructed from the name of
the file associated with the @code{aliases} relation, typically
it will be @file{aliases.dat}.

The file containing the actual aliasing data is automatically created
by the Router when asked to reconstruct the aliases database.
It does this based on a text file containing the alias definitions.
This text file, which corresponds to the Sendmail aliases file,
consists of individual alias definitions, possibly separated by
blank lines or commentary.
Comments are introduced by a sharp sign (octothorp: @samp{#}) at
any point where a token might start (for example the beginning of
a line, but not in the middle of an address), and extend to the end
of the line.
Each alias definition has the exact syntax of an RFC822 message header,
containing an address-list, except for comments.
The header field name is the local-part being aliased to the address-list
that is the header value.

The fact that an alias definition follows the syntax for an RFC822 message
header, introduces an incompatibility with Sendmail.
The string @samp{:include:} at the start of a local-part
(a legacy of RFC733) has special semantics.
Sendmail would strip this prefix, and regard the rest of the local-part
as a path to a file containing a list of addresses to be included in
the alias expansion.
Indeed, the Router behaves in the same manner, but because some of
the characters in the prefix are RFC822 specials, the entire local-part
must be quoted.
Thus, whereas Sendmail allowed:

@group
@example
people: :include:/usr/lib/mail/lists/people
@end example
@noindent
the proper syntax with ZMailer is:
@example
people: ":include:/usr/lib/mail/lists/people"
@end example
@end group

Like Sendmail, if a local-part is not found in the aliases database,
the Router also checks @file{~@i{local-part}/.forward} (if such exists)
for any address expansion.
The @file{.forward} file format is also an RFC822 address-list,
similar to what Sendmail expects.

There exists a special mechanism to do address expanding aliasing
on a mailing-lists, however it doesn't really take into account all
needs that mailinglist might have -- like reader rewrite, moderation,
etc. non-trivialities.  If you need more, than simple alias-like
address expansions, do look at various Mailing-List-Manager softwares.

The mechanism to do list-aliases is to look up for a file in directory
@samp{$MAILVAR/lists/}.  If there exists a file with looked up name,
then the address expansion is done by reading that file for recipient
addresses in RFC822 format (just like with @samp{:include:}d file).
Furthermore, all recipients produced from that address are sent with
@samp{listname-owner} as the sender, and error catcher@dots

Related to the list-alias mechanism is test for @code{*-owner}, and
@code{*-request} local-parts; if @code{basename somelist-owner -owner}
resolves to a file in the @samp{$MAILVAR/lists/}-directory, the owner
of the file is taken to do handle the list.

As special cases, a local-part starting with a pipe character (@samp{|}) is
treated as mail destined for a program (the rest of the local-part is any
valid argument to a @code{sh -c} command), and a local-part starting with
a slash character (@samp{/}) is treated as mail destined for the file named
by the local-part.

@node security, sequencer, databases, router
@section Security

Having local-parts that allow delivery to arbitary files, or can trigger
execution of arbitrary programs, can clearly lead to a huge security problem.
Sendmail does address this problem, but in a restrictive and unintuitive
manner.
This aspect of ZMailer security has been designed to allow the privileges
expected by common sense.

The responsibility for implementing this kind of security is split between
the Router and the Transport Agent that delivers a message to an address.
Since it is the Transport Agent that must enforce the security, it needs
some information to guide it.
Specifically, for each address it delivers to, some information about
the ``trustworthyness'' of that address is necessary so the Transport Agent
can determine which privileges it can assume when delivering for that
destination.
This information is determined by the Router, and passed to
the Transport Agent in the message control file.
The specific measure of trustworthyness chosen by
[XX: the present incarnation of]
ZMailer, is simply a user id (uid) value representing the source of
the address.

When a message comes in from a non-local host, the destination
addresses should obviously have no privileges on the local host
(when mailing to a file or a program).
Similarly, common sense would indicate that locally originated
mail should have the same privileges as the originator.
Based on an initial user id assigned from such considerations,
the privilege attached to each address is modified by the attributes
of the various alias files that contain expansions of it.
The algorithm to determine the appropriate privilege is to use
the user id of the owner of the alias file if and only if that
file is not group or world writable, and the directory containing
the file is owned by the same user and is likewise neither group
nor world writable.
If any of these conditions do not hold, an unprivileged user id
will be assigned as the privilege level of the address.

It is entirely up to the Transport Agent whether it will honour
the privilege assignment of an address, and indeed in many cases
it might not make sense (for example for outbound mail).
However, it is strongly recommended that appropriate measures are
taken when a Transport Agent has no control over some action that
may affect local files, security, or resources.

@group
The described algorithm is far from perfect.  The obvious dangers are:

@itemize @bullet
@item
The grandparent directories, to the Nth degree, are ignored, and
may not be secure. In that case all security loses anyway.
@item
There is a window of vulnerability between when the permissions
are checked, and the delivery is actually made. This is the
best argument I have heard so far for embedding the local delivery
program (currently a separate Transport Agent) in the Router.
@end itemize
@end group

There is also another kind of security that must be addressed.
That is the mechanism by which the Router is told about the origin
of a message.
This is something that must be possible for the message receiving
programs (@file{/bin/rmail} and the SMTP server are examples of these)
to specify to ZMailer.
The Router knows of a list of trusted accounts on the system.
If a message file is owned by one of these user id's, any sender
specification within the message file will be believed by ZMailer.
If the message file is not owned by such a trusted account, the Router
will cross-check the message file owner with any stated @samp{From:}
or @samp{Sender:} address in the message header, or any origin specified
in the envelope.
If a discrepancy is discovered, appropriate action will be taken.
This means that there is no way to forge the origin of a message
without access to a trusted account.

@node sequencer, routerlog, security, router
@section Router Control Flow

The following few pages use pseudo-code to describe the algorithm that
produces a control file (containing delivery instructions and the new
message headers) from a message file.
This algorithm is implemented in a C function called @code{sequencer()},
an apt description of how it orchestrates the various parts of
the ZMailer Router to implement the semantics of RFC822 message processing.

@display
@t{sequencer(}message file name@t{)}
@t{@{}
    Parse envelope and message header from the message file

    @t{if (}hostname has been set@t{)}
        Stamp the message with a trace header

    Determine if message contains Resent-* headers or not
    (from here on, only pay attention to the appropriate group of headers)

    Determine if the owner of the message file is trusted user

    @t{if (}there is no sender specified in the envelope@t{) @{}
        @t{if (}there is a Sender or From field in the message header
            @t{&&} the owner of the message file is trusted@t{)}
            Use the header value from it as the message sender
        @t{else}
            Generate a sender based on the owner of the file
        @t{if (}there is still no sender@t{)}
            Generate a sender referring to the local Postmaster
    @t{@} else @{}
        @t{if (}the owner of the message file is not trusted@t{) @{}
            Save message file for Postmaster to see
            Generate a sender based on the owner of the file
        @t{@}}
    @t{@}}
    @t{if (}an error occurred during parsing of the message envelope@t{) @{}
        Save the message file for the Postmaster to see and correct
        @t{return;}
    @t{@}}
    @t{if (}an error occurred during parsing of the message header@t{) @{}
        Save the message file for the amusement of the Postmaster
        @t{header_error = TRUE;}
    @t{@} else}
        @t{header_error = FALSE;}
    default address delivery uid@t{ = nobody;}
    @t{if (}the owner of the message file is trusted@t{) @{}
        @t{if (}an incoming channel is specified in the envelope@t{)}
            set the trusted channel origin accordingly
        @t{if (}an incoming host is specified in the envelope@t{)}
            set the trusted host origin accordingly
        @t{if (}an incoming user is specified in the envelope@t{)}
            set the trusted user origin accordingly
        @t{if (}any element of the trusted origin triple is null@t{) @{}
            set the resolved origin triple by
                routing the sender address
        @t{@}}
        @t{if (}the message origin is a local user@t{)}
            default address delivery uid@t{ = uid of that user;}
    @t{@} else @{}
        /* We know sender is local */
        default address delivery uid@t{ = uid of owner of message file;}
        @t{if (}message header contains a Sender, but no From field@t{) @{}
            Rename the ``Sender'' field into a ``From'' field
        @t{@}}
        @t{if (}the message header specifies a Sender@t{) @{}
            @t{if (}the specified Sender address does not correspond
                to the resolved origin of the message@t{) @{}
                Rename the ``Sender'' field a ``Fake-Sender'' field
                Set a flag to generate a Sender header
            @t{@}}
        @t{@} else if (}the message header only specifies a From field@t{) @{}
            @t{if (}the specified From address does not correspond
                to the resolved origin of the message@t{) @{}
                Set a flag to generate a Sender header
            @t{@}}
        @t{@}}
        @t{if (}flag is set that we need to generate a Sender header@t{)}
            Do so based on the owner of the message file
    @t{@}}
    @t{if (}default address delivery uid != nobody
        @t{&&} there is no From message header@t{) @{}
        Generate one based on the envelope origin address information
    @t{@}}
    /* Recipient determination */
    @t{if (}there are no recipients specified in the envelope@t{) @{}
        @t{if (}header_error@t{) @{}
            Reject the message with a ``bad header'' error
            @t{return;}
        @t{@}}
        Add all the message header recipient addresses (from To, Cc,
            and Bcc headers@t{)} to the message envelope recipient list
        @t{if (}there are still no recipients specified in the envelope@t{) @{}
            Reject the message with a ``no recipients'' error
            @t{return;}
        @t{@}}
    @t{@}}
    @t{if (}header_error@t{) @{}
        Return the message with a ``bad header'' warning
        Add Illegal-Object warning headers to the message header
    @t{@}}
    @t{if (}there is no To message header@t{) @{}
        /* Insert the To: header lines */
        Add the list of message recipients from the message envelope
            in the message header in To headers
    @t{@}}
@t{#ifdef notdef}
    rewrite all addresses in the message according to the incoming-rewriting
        rules for the originating channel.
@t{#endif notdef}
    @t{if (}hostname has been set@t{) @{}
        /* Make sure Message-Id exists, for loop control */
        @t{if (}there is no Message-Id message header@t{)}
            Generate a message id and add it to the message header
        @t{else}
            extract a message id from the existing header
        Log the message id
    @t{@} else}
        there is no message id
    /* Route recipient addresses */
    @t{for (}every recipient address in the message envelope@t{)}
        delivery privilege of address = default address delivery uid
    @t{for (}every recipient address in the message envelope@t{) @{}
        @t{router()}    /* Route the address */
        @t{if (}the returned triple is null@t{)}
            @t{continue;}    /* ignore this recipient */
        /* Rewrite this envelope address */
        @t{crossbar(source triple, recipient triple)}
        @t{if (}the return value from @code{crossbar()} is null@t{)}
            @t{continue;}    /* ignore this recipient */
        @t{else if (}the message header rewriting function name is null@t{) @{}
            Save the message for the Postmaster to see
            @t{return;}
        @t{@}}
        /* Don't send message to the same address twice */
        @t{if (}we have already seen this destination triple@t{) @{}
            @t{if (}the message is going to be sent to that destination@t{)}
                @t{continue;}    /* suppress duplicates */
        @t{@}}
        @t{if (}this destination triple has not been alias expanded
            @t{&&} it represents a local destination
            @t{&&} it has an alias expansion@t{) @{}
            Add the list of expanded addresses to the list of
                addresses processed by this loop, each with
                a delivery privilege determined by the source
                of the alias expansion
            @t{continue;}    /* ignore this address */
        @t{@}}
        Flag that this destination triple will be sent out
        Add the address destination triple to a list for each channel
        @t{if (}the message header address rewriting function name
            is new for this message@t{) @{}
            Add the name to a collection of the kinds of message
                header address rewritings that need be done
        @t{@}}
    @t{@}}
    @t{for (}every kind of message header address rewriting we need to do@t{) @{}
        Call the indicated function with every address in the header
        Store the transformed headers for later use
    @t{@}}
    @t{if (}there is no Date message header@t{)}
        Generate one based on the modification time of the message file
    /* Emit specification to the transport system */
    @t{for (}every recipient address@t{) @{}
        @t{if (}control file has not been created@t{) @{}
            Create message control file
            Write a standard preamble consisting of
                the corresponding message file name
                the offset of the start of the message body
                the message id (if any) for log identification
            @t{if (}this message did not come from the error channel@t{)}
                Write an error return address
        @t{@}}
        @t{if (}the envelope sender address form for this recipient
            is different from the previous sender address form@t{)}
            Write the sender address origin triple
        Write the recipient address destination triple, and the
            corresponding address delivery privilege
        @t{if (}message header for this address is different than
            the message header for the next recipient address@t{) @{}
            Write the complete message header for this destination,
                as reconstructed from the original message
                header and the stored headers transformed by
                message header address rewriting
        @t{@}}
    @t{@}}
    @t{if (}we created a message control file@t{) @{}
        relink the message file itself to the @samp{QUEUE} directory
        relink the control file to the @samp{TRANSPORT} directory
    @t{@}}
    @t{return;}
@t{@}}
@end display

@node routerlog, addresstest, sequencer, router
@section Logging

When the Router starts up as a daemon, it will attach its standard
output and standard error streams to a log file.
All messages from the Router will appear on one of these streams,
and will therefore show up in a central location for perusal by
the Postmaster or other interested parties.
Usually only abnormal occurrences will be logged in this manner,
but any messages printed will show up here.
In particular, many of the components of the Router contain trace
print statements that can be enabled at run-time.
In fact, interactive debugging of the configuration file is performed
this way, since when the Router is run in the foreground, the standard
output and error streams are attached to the terminal in the normal fashion.
Thus, all messages will appear in front of the person testing the
configuration.

The tracing functionality is controlled either on the command line,
or by calling the @code{trace} and @code{untrace} functions from within
the configuration file, or interactively.
The interactive behaviour of the Router, is to read and execute its
configuration file (as normal), and then sit in an infinite loop
reading commands from its standard input stream.
This allows a person executing the Router interactively, to execute
arbitrary statements in the configuration file programming language.
The statements typed in are buffered until an End Of File indication,
and then executed by the configuration file interpreter.
This cycle is repeated until a syntax error occurs, or the process
is interrupted.  [XX: yes, this is a very rough mode of
interaction. do you have any suggestions for improving it? ]

The @code{trace} and @code{untrace} functions take one or more words
as arguments, and turn on (off) flags that enable tracing in a component
of the Router corresponding to each word.
The current list of words, and the corresponding actions traced, are:

@table @code
@item alias
alias expansion
@item all
turns all trace flags on
@item assign
variable assignment
@item bind
the BIND nameserver responses
@item compare
case label pattern matching
@item db
database lookups
@item final
print message information after sequencer returns
@item functions
function calls and returns
@item matched
successful case label matches
@item memory
memory allocation statistics
@item off
turns all trace flags off
@item on
same as @code{functions}
@item parsetree
the configuration file parse tree
@item regexp
regular expression execution
@item resolv
the BIND resolver library @code{RES_DEBUG} option
@item rewrite
message header rewriting
@item router
envelope recipient address routing
@item sequencer
control flow in the sequencer function
@end table

In addition to this, each message processed is logged via the standard
system logging facility (syslog) if it is available, and when
not especially disabled at the router invocation.
(If normal message logging to syslog is disabled, important error
 reporting messages are still logged.)


@node addresstest, , routerlog, router
@subsection Address Testing

For example, if you wish to see how an address is routed, you can run the
command:

@example
echo "trace on ; router $address" | router -I
@end example

@noindent
which, with @code{$address} bound to @samp{bond@@sis.mod.uk} might produce
something like:

@group
@example
GNU Mailer router (Zmailer alpha.1 #0: Sun Jan 31 17:38:53 EST 1988)
    rayan@@ephemeral.ai:/usr/src/zmailer/router
Copyright 1988 Rayan S. Zachariassen

router: parameters: 'bond@@sis.mod.uk'
    echo: parameters: 'router:' 'bond@@sis.mod.uk'
router: bond@@sis.mod.uk 
    echo: returns: ''
    canonicalize: parameters: 'bond@@sis.mod.uk'
        focus: parameters: 'bond<@@sis.mod.uk>'
            [: parameters: 'sis.mod.uk' ']'
            [: returns: 'true'
        focus: returns: 'bond<@@sis.mod.uk>'
    canonicalize: returns: 'bond<@@sis.mod.uk>'
    [: parameters: '' ']'
    [: returns: ''
    [: parameters: 'sis.mod.uk' '==' 'ephemeral.ai.toronto.edu' ']'
    [: returns: ''
    [: parameters: 'sis.mod.uk' ']'
    [: returns: 'true'
router: returns: 'smtp' 'sis.mod.uk' 'bond@@sis.mod.uk'
@end example
@end group

@node scheduler, transports, router, top
@chapter Scheduler

The Scheduler complements the Router as the other major process in ZMailer.
The decisions it makes involve how to manage and time delivery of messages
to their destination, and its name arises from this scheduling function.
While the Router interprets message files, the Scheduler interprets only
the control files corresponding to the message files.

The control files are usually produced by the Router, and appear
in a directory scanned by the Scheduler daemon.
Whenever a new control file does appears in that directory, its
contents are used to update a data structure, maintained by
the Scheduler, that describes which addresses in which messages
are destined for which hosts and channels.
The information stored along with each channel/host combination
is a set of byte offsets into the control file, giving the location
of address specifications corresponding to that combination.
This information can later be passed to a transport/delivery program,
and is updated based on feedback from these programs.


A Transport Agent is a program that the Scheduler executes to deliver
messages.
The Scheduler determines the correspondence between channels, hosts,
or channel/host destinations, and a specific Transport Agent, by
interpreting a simple table from a configuration file.
The Transport Agent process is told which control files it should
inspect for work, and tells the Scheduler the status of the destination
addresses it tried to process.
The Scheduler then updates the model it maintains of work that needs
to be done, and will eventually remove the last link to a control file
and its corresponding message file.
At that point, ZMailer has done its job with regard to each message.

Note that the communication between the Scheduler and other programs
is mostly via the message control file, instead of by direct interaction
(when the Scheduler converses with Transport Agents).

@section Message Control File

A message control file is a file created by the Router to contain all
the information necessary for delivery of a message submitted in
a corresponding message file.
It is interpreted by the Scheduler, which needs to know at all times
which messages are pending to go where, and how.
It is also interpreted by one or more Transport Agents, possibly
concurrently, that extract the delivery information relevant to
their purpose.

The concurrency aspect means that the Transport Agents must cooperate
on a locking protocol to ensure that delivery to a particular destination
is attempted by only one Transport Agent at a time, and a status protocol
to ensure unique success or failure of delivery for each destination.
There are potentially many ways to implement such protocols, but, in
the spirit of simplicity, ZMailer uses a control file as a form of shared
memory.
Specific locations within each control file are reserved for flags that
indicate a specific state for their associated destination address.
The rest is taken care of by the I/O semantics when multiple processes
update the same file.

Apart from necessary envelope and control information, a control file also
contains the new message header for the message, which contains the header
addresses as rewritten by the Router.
Since a message may have several destinations with incompatible address
format requirements, there may be several corresponding groups of message
headers.
This will be illustrated by the sample control file shown in the following
subsection.

@subsection Format

A control file consists of a sequence of fields.
Each field starts at the beginning of a line (i.e. at byte 0 or
after a Newline), and is identified by the appearance of a specific
character in that location.
This id character is normally followed by a byte containing a tag value
(semaphore flag), followed by optional semaphore data (transporter process
id, for example), followed by the field value.

Here is a simple control file produced by a test message, just before it was
removed by the Scheduler:

@group
@example
--------------------
i 24700
o 72
l <88Jan10.003129est.24700@@bay.csri.toronto.edu>
e Rayan Zachariassen <rayan>
s local - rayan
r+      local - rayan 2003
m
Received: by bay.csri.toronto.edu id 24700; Sun, 10 Jan 88 00:31:29 EST
From:   Rayan Zachariassen <rayan>
To:     rayan, rayan@@ephemeral
Subject: a test
Message-Id: <88Jan10.003129est.24700@@bay.csri.toronto.edu>
Date:   Sun, 10 Jan 88 00:31:24 EST

s local - rayan@@bay.csri.toronto.edu
r+      smtp ephemeral.ai.toronto.edu rayan@@ephemeral.ai.toronto.edu 2003
m
Received: by bay.csri.toronto.edu id 24700; Sun, 10 Jan 88 00:31:29 EST
From:   Rayan Zachariassen <rayan@@csri.toronto.edu>
To:     rayan@@csri.toronto.edu, rayan@@ephemeral.ai.toronto.edu
Subject: a test
Message-Id: <88Jan10.003129est.24700@@bay.csri.toronto.edu>
Date:   Sun, 10 Jan 88 00:31:24 EST

--------------------
@end example
@end group

The id character values are defined in the @file{mail.h} system header file,
which currently contains:

@group
@example
#define _CF_MESSAGEID  'i' /* inode number of file containing message */
#define _CF_BODYOFFSET 'o' /* byte offset into message file of body */
#define _CF_BODYFILE   'b' /* alternate message file for new body */
#define _CF_SENDER     's' /* sender triple (channel, host, user) */
#define _CF_RECIPIENT  'r' /* recipient n-tuple, n >= 3 */
#define _CF_ERRORADDR  'e' /* return address for error messages */
#define _CF_XORECIPIENT 'R'/* one of XOR set of recipient n-tuples */
#define _CF_RCPTNOTARY 'N' /* DSN parameters for previous recipient */
#define _CF_DSNENVID   'n' /* DNS 'MAIL FROM<> ENVID=XXXX' data */
#define _CF_ERRORADDR  'e' /* return address for error messages */
#define _CF_DIAGNOSTIC 'd' /* diagnostic message for ctlfile offset */
#define _CF_MSGHEADERS 'm' /* message header for preceeding recipients */
#define _CF_LOGIDENT   'l' /* identification string for log entries */
#define _CF_OBSOLETES  'x' /* message id of message obsoleted by this */
#define _CF_VERBOSE    'v' /* log file name for verbose log (mail -v) */
#define _CF_TURNME     'T' /* trigger scheduler to attempt delivery now */
@end example
@end group

There is one field per line, except for @code{_CF_MSGHEADERS} which
has some special semantics described below.
The following describes the fields in detail:

@table @code
@item i
This field identifies the message file corresponding to this control file.
It is the name of the message file in the @file{QUEUE} directory
(@file{~/queue}).  This is typically the same as the inode number for that
file, but need not be.  It is used by Transport Agents when copying the
message body, and by the Scheduler when unlinking the file after all the
destination addresses have been processed.  For example:@refill

@example
i 21456
@end example

@item o
Specifies the byte offset of the message body in the message file.  It is
used by Transport Agents in order to copy the message body quickly, without
parsing the message file.  For example:

@example
o 466
@end example

@item e
Gives an address to which delivery errors should be sent.  The address
must be an RFC822 @i{mailbox}.  For example:

@example
e "Operations Directorate" <d-ops@@sis.mod.uk>
@end example

@item l
The field value is an uninterpreted string which should prefix all log
messages and accounting records associated with this message.  This value
is typically the message id string.  For example:

@example
l <88Jan6.103158gmt.24694@@sis.mod.uk>
@end example

@item s
This field specifies an originator (sender) address triple, in the sequence:
previous channel, previous host, return address.
It remains the current sender address until the next instance of this field.
Since there can only be one sender of a message, multiple instances of
the field will correspond to different return address formats as produced
by the @code{crossbar} algorithm in the Router.
For example:

@group
@example
s smtp sis.mod.uk @@lab.sis.mod.uk:q@@deadly-sun.lab.sis.mod.uk
s uucp sisops lab.sis.mod.uk!deadly-sun.lab.sis.mod.uk!q
@end example
@end group

@item r
This field specifies a destination (recipient) address triple, in the
sequence: next channel, next host, address for next host.  Optional
information to be passed to the Transport Agent may be placed after the
mandatory fields; this currently refers to the delivery privilege of the
destination address.
Since the optional values of this field are only interpreted by
the Transport Agent, changes in what the Router writes must be
coordinated with the code of the Transport Agents that might interpret
this field.
For example:

@group
@example
r local - bond 0
r uucp uunet sisops!bond -2
@end example
@end group

@item m
Apart from a message body, a Transport Agent needs the message headers
to construct the message it delivers.
These message headers are stored as the value of this field.
Since message headers obviously can span lines, the syntax for
this field is somewhat different than for the others.
The field id is immediately followed by a newline, which is followed
by a complete set of message headers.
These are terminated (in the usual fashion) by an empty line, which
also terminates this field.
In the following example, the last line of text is followed by an empty
line, after which another field may start:

@group
@example
m
From: M
To: Bond
Subject: do get a receipt, 007!

@end example
@end group

@item d
This field is @emph{not} written by the Router.
It is written by the Scheduler to remember errors associated
with specific addresses.
The field  value has two parts, the first being the byte offset
in the control file of the destination (recipient) address causing
the error, and the rest of the line being an error message.
The Transport Agents discover these errors and report them to the Scheduler.
The Scheduler will collect them and report them to the error return address
(if any) after all the destinations have been processed [XX:or at other times].
For example:

@example
d 878 No such local user: 'bond'.
@end example

@end table

It should be noted, that in sender and recipient fields the first two
field values (channel and host) cannot contain embedded spaces, but the
third field value (the address) may.
Therefore, in the presence of extra fields, parsing within Transport Agents
must be cautious and not assume that an address does not contain spaces.

As mentioned, the second byte of most fields are used for concurrency control
and status indication.
This tag byte can contain several values that indicate current or previous
activity.
The fields where this is relevant are the destination (recipient) address
and diagnostic fields.
The tag values are defined in the @file{mail.h} file mentioned previously,
as follows:

@group
@example
#define _CFTAG_NORMAL ' ' /* what the router sets it to be */
#define _CFTAG_LOCK   '~' /* that line is being processed, lock it */
#define _CFTAG_OK     '+' /* positive outcome of processing */
#define _CFTAG_NOTOK  '-' /* something went wrong */
#define _CFTAG_DEFER  _CFTAG_NORMAL /* try again later */
@end example
@end group

The extract above is self-explanatory.

A message control file will normally contain a preamble that specifies
information about the associated message file, the message body offset, an
error return address, and a log entry tag.  After this comes a repeated
sequence of: sender address field, recipient address fields, and the message
header corresponding to these recipients.  After as many of these groups as
are necessary, any diagnostic fields will be appended to the end of the
control file.  The restrictions on the sequence of addresses and message
headers, are that a sender address field must precede any recipient address
field, and a recipient address field must (immediately) precede any message
header field, and no sender or recipient addresses may follow the last
message header field.@refill

@section Scheduler Configuration File

The major action of the Scheduler is to periodically start up Transport
Agents and tell them what to do.  This is controlled by a table in a
configuration file that is read by the Scheduler when it starts.  A typical
configuration file would look something like:@refill

@group
@example
# pattern    intvl   ch/ho/* uid     gid     command
local/*      10s      2 0 0  root    daemon  mailbox local
smtp/*       1m      10 2 0  root    daemon  smtp -l /tmp/smtp.log $host
error/*      5m      10 0 0  root    daemon  errormail
uucp/*       10m     10 0 0  root    daemon  sm -c $channel uucp
@end example
@end group

Any line starting with a @samp{#} character is assumed to be a comment
line, and is ignored, as are empty lines.  All other lines must follow a
rigid format.  Each line consists of eight white-space separated fields.
The fields, in sequence, are:@refill

A pattern, that selects which channel/host combinations are relevant to the
current line.  The pattern has the form: @i{channel}/@i{host}, with the slash
being mandatory.  The subpatterns (i.e. each side of the slash) may contain a
@code{glob} (or @code{sh}) style pattern.  These patterns are tested in the
order they appear, with the channel and host values for destination addresses
in a message.  When both patterns match, the line with the matching pattern
describes the Transport Agent that should be used to deliver the message to
that destination.  It is important that the Transport Agent recognizes at
least the set of addresses the in message control file, that the Scheduler
configuration table assumes it does.  Otherwise, some addresses may never
get delivered to, and the message will stay in the Scheduler
indefinitely.@refill

An interval specification that says how often the Scheduler should check for
work pending for the Transport Agent described by that line.  The time
specification must use an appropriate suffix: @samp{s} for seconds, @samp{m}
for minutes, @samp{h} for hours, or in combinations, e.g. @samp{1h30m}.
The minimum value specified in the configuration file will be the directory
scanning interval used by the Scheduler.@refill

A maximum number of Transport Agents simultaneously active for the channel
matched by the pattern for that entry.  If 0, no upper limit is enforced.@refill

A maximum number of Transport Agents simultaneously active for the host
matched by the pattern for that entry.  If 0, no upper limit is enforced.@refill

A total maximum number of Transport Agents simultaneously active due to that
entry.  If 0, no upper limit is enforced.@refill

A user id to set as the real and effective user id when executing the command
associated with that entry.  Either a symbolic (login name) or numeric value
may be specified.@refill

A group id to set as the real and effective group id when executing the
command associated with that entry.  Either a symbolic (login name) or
numeric value may be specified.@refill

Finally, the Transport Agent invocation command itself, as it would appear
on a normal command line.  Note however that the Scheduler executes the
command directly without all the command line interpretation afforded by
a shell.  The only special action is to replace instances of the word
@samp{$channel} with the name of the channel matched by the pattern, and
instances of @samp{$host} with the name of the host matched by the
pattern.@refill

Note that the command must have enough privileges specified to write into
the control file, in addition to whatever is necessary to perform its
delivery duties and logging.@refill

@section Transport Agent protocol

Once the Scheduler starts up a Transport Agent by executing one of the
commands specified in the configuration file, it needs to pass information to
the Transport Agent about which messages and addresses it should process.
The Transport Agent in return needs to report to the Scheduler about
the success or failure of its delivery activity, so that issues related to file
management and error reporting can all be centralized in the Scheduler
process.@refill

To accomplish this, the Scheduler engages in a simple exchange with each
Transport Agent it has started.  For this reason, the Scheduler creates two
pipes attached to the standard input and standard output of the Transport
Agent processes it executes.  The standard error descriptor is shared with the
Scheduler process, and usually refers to the Scheduler log file.@refill

Just before the Scheduler starts up a Transport Agent, it scans through its
model of the pending message control files, and determines which are relevant
to the impending invocation of the Transport Agent.  Once the subprocess is
running, the Scheduler will write to the standard input of its child, the
names of the control files it should process, one to each line.  This list is
terminated by an empty line, to indicate to the Transport Agent that the
Scheduler finished its business normally.@refill

In turn, the delivery process will open each named control file, scan it for
destination addresses relevant to its specific invocation, and attempt
delivery to those addresses.  For each destination address, it will print to
its standard output a line that describes the address, and that contains
a status indication.  The syntax is:@refill

@display
@i{id}/@i{offset}/@i{status} @i{comment}
@end display

The @i{id} is the message file id contained in the @samp{i} field of the
control file.  The @i{offset} is the byte offset into the pertinent control
file of the destination address field.  The @i{status} is one of the
following list of keywords understood by the Scheduler:@refill

@group
@table @code
@item ok
Delivery was successful.
@item error
Delivery was unsuccessful.
@item deferred
Delivery was attempted, but is deferred.
@end table
@end group

The optional @i{comment} is an arbitrary string that clarifies the @i{status}
code.  It is separated from the @i{status} code by a single space.
For example, the following is a possible sequence of reports:@refill

@group
@example
18453/3527/ok
18453/3565/deferred Unable to contact sis.mod.uk!
18453/4211/error No such local user: 'bond'.
@end example
@end group

After each message file has been processed, an empty line is output to
indicate this.  The Transport Agent will continue to the next message control
file (if any) that has been written to it by the Scheduler.@refill

A complete exchange between a Scheduler and a Transport Agent might proceed
as shown:@refill

@group
@example
Scheduler          Transport Agent
---------------------------------------------------------------------
smtp/21456
                   18453/878/ok
                   18453/1013/error Illegal hostname: 'spectre'
                   _
local/21456
                   18453/945/deferred Cannot lock mailbox: 'bond'
                   _
_
@end example
@end group

The underscores indicate empty lines emitted by either side in the
synchronous protocol.  After such a conversation, the Transport Agent process
will exit gracefully.  Whenever the status of a destination is updated, the
Scheduler will check its internal data on whether or not a link to the
control file should be removed, or if indeed delivery has been completed and
both the message file and the last links to its control file should be
removed.@refill

@section Mail Queue printing

It is possible to read through the directory hierarchy under
the @file{TRANSPORT} directory to synthesize a model of which messages
are queued to go where.  However, this method does not guarantee an accurate
image of the model within the Scheduler process, nor can it provide any
status information (as given by the status commentary of Transport Agents)
other than success or failure.  The ideal solution would be a way of
interrogating the Scheduler itself about the current state, and then perhaps
use this as a basis for verbose embellishments.@refill

Such a facility is incorporated in the Scheduler.  The exact interrogation
mechanism depends on the facilities of the host operating system: a system
with TCP/IP would use a socket rendezvous, a system with named pipes would
use a prearranged special file and signalling mechanism, a system without
either would rely on normal files.@refill

In all cases, the result of the interrogation is a terse list of of messages,
their destinations (channel/host combination), and the offsets of
the addresses corresponding to each destination.  For example, a sample
state dump from the Scheduler is:@refill

@group
@example
29198:  smtp/csri.toronto.edu, 2 addresses [196,247]
        smtp/ephemeral.ai.toronto.edu, 1 address [128]
@end example
@end group

This shows one message (id 29198) queued for transmission via SMTP to two
different destination hosts.  One destination has two associated addresses,
referred to by byte offsets (196 and 247) into the control file for the
message (@file{~/transport/29198}).  The other destination has only one
address associated with it.@refill

The above dump corresponds to the state just after the Scheduler has parsed
the control files.  After the Transport Agent corresponding to the
@code{smtp/ephemeral.ai.toronto.edu} destination has exited, the state might
become:@refill

@group
@example
29198:  smtp/csri.toronto.edu, 2 addresses [196,247]
        smtp/ephemeral.ai.toronto.edu, 1 address [128]
                connect: Connection refused (will retry)
@end example
@end group

In this case, the host @code{ephemeral.ai.toronto.edu} is alive, but no
SMTP server is running.  After the Transport Agent for the other destination
has exited, the state might be:@refill

@group
@example
29198:  smtp/ephemeral.ai.toronto.edu, 1 address [128]
                connect: Connection refused (will retry)
@end example
@end group

Finally, once this remaining destination is processed successfully, the
Scheduler reports:@refill

@example
Mail queue is empty
@end example

These reports from the Scheduler succinctly express the state of the
queues in a format that is human-readable, and that is also easy to parse
automatically.  The only information not provided are the actual addresses
referred to by the state dump.  The program that queries the Scheduler for
this information is capable of finding these addresses if it needs
to (assuming the control files are readable), and presenting it in a different
format.  The Scheduler does not remember the actual address information, and
so cannot easily include it in the dump.  Since the Scheduler must spend a
minimum of time servicing requests from Transport Agents and mail queue
queries, it leaves nontrivial work to the querying program.@refill

Some advantages arise from this mechanism: in environments with host-to-host
interprocess communication (e.g. TCP/IP) it becomes possible to query
Schedulers on remote hosts about their state, and such remote queries can
only get verbose information if the querying process has access to the
control files of the remote ZMailer installation.  This makes it possible for
an environment making use of distributed filesystems to have a single ZMailer
installation on a mail server host, and for all the other local machines to
access its services transparently.  At the same time, no private information
can be divulged without direct access to the @file{POSTOFFICE} directory.@refill

The ZMailer distribution contains a utility program @code{mailq} that is used
to query Schedulers.  It supports the transparency paradigm in an
NFS environment, by arranging to query the Scheduler running on the NFS
server host for the @file{POSTOFFICE} directory visible on the local
host.@refill

@section Logging

As with the Router, the Scheduler daemon will attach its standard output and
standard error streams to a log file.  The standard error stream of each
Transport Agent invocation is inherited from the Scheduler, and so is
attached to the same log file.  The Scheduler does have an option to produce
a debugging log, but otherwise only extraordinary occurrances are logged
(for example, a delivery failure, missing Transport Agents, etc.).@refill

@node transports, miscellaneous, scheduler, top
@chapter Transport Agents

A Transport Agent is a program that delivers mail to a particular
destination.  The destination paradigm in ZMailer involves the concept of a
channel, a next-host, and a next-address.  The first two are used by the
Scheduler to select a Transport Agent, and by a Transport Agent to identify
which destinations it should process.  Any necessary information to
accomplish this selection, is either contained within the Transport Agent, or
supplied by the Scheduler on the command line when invoking a Transport
Agent.  The message control files examined by a Transport Agent instance are
passed by the Scheduler in a simple protocol designed for the purpose, and
status reports on the actions of the Transport Agent are returned by the same
protocol.@refill

When a Transport Agent starts up, it expects to read message control file
path names on its standard input stream, and will print status reports on
its standard output stream.  Unexpected errors are sent to the standard error
stream for logging.  A Transport Agent can be invoked interactively for test
purposes, but usually it is started as a child of the Scheduler daemon, with
its input and output streams attached to the Scheduler using pipes, and
sharing the error stream with the Scheduler itself (and other concurrent
Transport Agent processes).@refill

The following sections describe the Transport Agent programs that come with
the ZMailer distribution.@refill

@section Local delivery (mailbox)

The delivery of local mail is of paramount importance in a mailer.  Of all
the things that might go wrong during mail processing, a mistake by the
local delivery process can be the most critical.  Since it is also a very
frequent operation, this Transport Agent must be both robust and efficient.
Perfection is elusive, but the local delivery program included with ZMailer
has proven itself in the original version used with Sendmail.@refill

This program will look for destinations with a channel of @samp{local},
and will ignore the next-host specification.  The next-address specification is
either a local account id, a full path to a file, or a pipe (@samp{|})
followed by a valid argument to @code{sh -c}.  The following are examples
of @i{legal} values:@refill

@group
@example
bond
/usr/arch/lists/info-widget
/etc/passwd
|sed -e '1,/^$/d' >> /etc/passwd
|/bin/mail badhost!badguy </etc/passwd
@end example
@end group

@noindent
Here are some @i{illegal} variations:

@group
@example
<bond>                            @r{(angle brackets invalid in local-part)}
"bond"                            @r{(double quotes unlikely in login id)}
bond@@sis.mod.uk                   @r{(local-parts do not contain @samp{@@})}
james bond                        @r{(whitespace unlikely in login id)}
lists/info-widget                 @r{(not an absolute pathname)}
sed -e '1,/^$/d' >> /etc/passwd   @r{(does not start with a @samp{|})}
@end example
@end group

Note that the effect of these addresses depends on whether the local delivery
program actually honours the request, and if it does which privileges are
used while executing the indicated action.@refill

Specifically, if the next-address does not start with either @samp{|} or
@samp{/}, it is assumed to be a user name.  This is checked by lookup in
the system account database (@file{/etc/passwd}), to determine which user id
should own the mailbox file.  If the indicated account exists, mail is
delivered to its corresponding mail spool file, in the standard format
(return address and delivery date in a @samp{From } line preceding the
actual message, etc.).  Note that even though account names should ideally
be case-independent there is no portable way of doing a case-independent
lookup in the account database.  The Router deals with this by first trying
the name lookup without doing a case translation, and if that fails the
uppercase letters in the @i{local-part} will be mapped to lowercase and the
lookup retried. The local delivery program does @emph{no} aliasing on
its own.@refill

If delivery to a file or command is indicated, the actual delivery is done
using the user id listed as the destination address privilege in the control
file.  What this actual privilege allows, is up to the security mechanism
in the Router.  Since addresses specified from a remote host start out with
minimal privileges, they will usually not cause any harm on the local
system.@refill

Programs executed by this Transport Agent will be given an environment
containing the @code{$PATH}, @code{$SHELL}, @code{$HOME}, @code{$USER},
@code{$UID}, and @code{$SENDER} environment variables.  The first two are
constant, the next three depend on the delivery privilege of the address, and
the value of the last environment variable is set to be the return address
of the message being delivered.  The current directory is set to be
@code{$HOME} when possible.@refill

The local delivery program contains code that may be enabled at compile time
to honour the @code{comsat} protocol.  There are separate symbols to enable
this for local users (@code{BIFF}) and for remote users (@code{RBIFF}).
In the former case, users would enable the feature individually by executing
@code{biff y}, while in the latter case a @file{.rbiff} file in the user's
home directory triggers the remote notification of new mail.

@section Error mail delivery (errormail)

The error messages of ZMailer are stored as message file forms in a
specific @file{FORMS} directory.  The various ZMailer programs will access
the appropriate forms directly, but errors detected in the Router
configuration file must be handled in a different way.  By convention,
any problem found by configuration file code is handled by changing the
message destination to be a triple of the form:@refill

@example
(error, @i{form}, @i{address})
@end example

The @samp{error} channel is serviced by this Transport Agent, which expects
the @i{form} listed to be the name of a file in the @file{FORMS} directory.
This should be a prototype message file, containing all generic information
associated with the error (i.e. the message header lines and an appropriate
explanation to the user).  The @i{address} is the address rejected by the
configuration file code, for whatever reason is given in the @i{form}
file.@refill

By convention, the names of the @i{form} files indicate the class of error
that occurred.  The following describes the standard forms that come with
ZMailer:@refill

@table @file
@item err.badheader
Syntax error in the message header.
@item err.delivery
Delivery problem, used by the Scheduler on behalf of Transport Agents.@refill
@item err.nonewsgroup
A non-existent USENET Newsgroup was addressed.
@item err.norecipients
The message has no recipients listed.
@item err.unresolvable
The routing code in the Router configuration file cannot determine a
destination for the message.@refill
@item warn.badheader
Used to chastise a user who sends improperly formatted mail.@refill
@end table

Of these, only @file{err.nonewsgroup} and @file{err.unresolvable} are
referred to by the Router configuration file, the rest are used internally by
the Router or Scheduler.  Therefore these forms @emph{must} be available for
proper operation of ZMailer.@refill

To illustrate, here is the default @file{err.badheader} form:

@group
@example
--------------------
From:   The Post Office <postmaster>
Subject: Invalid message header
Cc:     The Postmaster <postmaster>

The following message arrived with an illegal header according to the
RFC822/976 protocol specification. If you do not recognize the source
of the bad header, perhaps you should ask a postmaster at your site.

The following annotated headers illustrate where the error(s) occurred:

--------------------
@end example
@end group

@section SMTP client (smtp)

The SMTP Transport Agent implements this message transfer protocol
according to RFC821.  It scans message control files for a channel called
@samp{smtp}, and a next-host as specified on the command line.  Only a single
virtual circuit (VC) is established to the remote SMTP server, and all
transactions are carried out in sequence across this VC.  By contrast,
Sendmail opens a new VC for every mail message.

This program does not enforce the line length limits of the SMTP protocol,
nor does it check that the message file data is 7 bit ASCII.  However, the
CRLF line termination rule is followed, as are all other aspects of the SMTP
protocol.  When connected to a ZMailer SMTP server program, message bodies
containing arbitrary binary data may be transferred (since the SMTP DATA
encoding is reversible, and there are no line length limits on either end).

A log file may be specified for recording the SMTP transaction.

@section Sendmail compatible delivery programs (sm)

Because Sendmail already has many ``mailer''s written for it, and to ease the
transition from Sendmail to ZMailer, this Transport Agent was written to
interface with such programs from the ZMailer environment.  The basic
characteristic of a Sendmail ``mailer'' is that its command line specifies
what must be done with the message available on its standard input stream.

Because of the generic interface, this Transport Agent requires a small
configuration file which it reads on startup.  The configuration file declares
which programs are available, how to invoke them, and what channel each
program corresponds to.  Here is a sample configuration file:

@group
@example
# M     F =    P =                    A =
local   mS     /usr/lib/mail/localm   localm -r $g $u
prog    -      /bin/sh                sh -c $u
tty     rs     /usr/local/to          to $u
uucp    U      /usr/bin/uux           uux - -r -a$g -gC $h!rmail ($u)
news    m      /usr/lib/mail/pnews    post.news $h $u
@end example
@end group

The configuration file is a table with each line containing four fields:
a channel name, Sendmail ``mailer'' flags, the full path name of the program
to execute, and the command line that program should see.

The flags field contains the flags that are appropriate to the ZMailer
environment, for example the presently recognized flags are:

@table @samp
@item f
Include a @samp{-f @i{sender}} in the command line.
@item r
Include a @samp{-r @i{sender}} in the command line.
@item S
Do not reset the uid to the real uid of the Transport Agent process.
@item n
Do @emph{not} prepend a @samp{From } line to the message.
@item s
Strip quotes on addresses [XX:todo].
@item m
Many recipients may be handled by a single instance of the command.
@item P
Add a ``Return-Path'' message header.
@item U
Prepend a @samp{From ... remote from ...} line to the message.
@item X
Use the SMTP hidden dot algorithm (i.e. escape periods on a line by
themselves).@refill
@item E
Replace occurrences of @samp{From } at the start of a line in the
message body with @samp{>From }.@refill
@item 7
Pass 7-bit ASCII by stripping 8th bit of bytes in the message [XX:todo].@refill
@item -
No-op flag.
@end table

Mailer flags that are not mentioned in the above table have been excluded due
to their lack of semantics in this situation.  Typically their functionality
should be accomplished in the Router instead [XX: if it isn't, and it is
needed, please let me know].

The command line specification may contain anything valid in the same field
in a Sendmail ``mailer'' definition.  In particular, any argument containing
@samp{$u} is expanded as many times as there are recipients that can be dealt
with at once by that command.  The @samp{$g} macro expands to the return
address of the message, and @samp{$h} to the next-host in the destination.

At present, no special environment is set up for programs executed by this
Transport Agent.  The standard output and standard error of such processes
are caught by the Transport Agent, and the first line read (if any) is
passed on to the Scheduler using the normal status reporting mechanism.

@node miscellaneous, how-to, transports, top
@chapter Miscellaneous
@section Sendmail compatibility

After installing the Sendmail compatible ZMailer interface programs, the
present user-visible incompatibilities with Sendmail proper are:

@itemize @bullet
@item
Verbose mode (@samp{-v} flag to Sendmail) is not implemented.@refill
@item
Occurrences of @samp{:include:} specifications in the aliases database
must be quoted.@refill
@item
The ``Return-Receipt-To'' message header is not yet honoured.@refill
@item
The mailing-list management features of Sendmail are not implemented,
avaiting consultation.
@end itemize

@section SMTP server

The ZMailer distribution contains an SMTP server program for the BSD socket
implementation of TCP/IP.  It is an asynchronous implementation, in that
address semantics are not checked in real time, nor are other (optional in
the SMTP standard) functions that require Router functionality.  The server
simply says ``Yes yes, sure!'' to everything, and passes the information to
the Router for verification.  The program may also be used in non-daemon
mode to unpack BSMTP format messages on the standard input stream.  For
compatibility with the Sendmail variation on the SMTP protocol, it accepts
the @code{VERB} and @code{ONEX} commands as No-Ops.  The @code{VRFY},
@code{EXPN}, @code{HELP}, and @code{TURN} commands are presently
unimplemented, as is the case for the interactive @code{SEND}, @code{SAML},
and @code{SOML} commands.@refill

@node how-to, uasupport, miscellaneous, top
@chapter How-To Guide

This chapter is intended to give practical tips on topics related to the
maintenance and customization of ZMailer.  If you want to see something
covered here, let me know.@refill

@section How to install ZMailer

Thie @file{README} file in the distribution contains specific instructions
for installing ZMailer.  The following goes into slightly more depth than the
@file{README} file does:@refill

The documentation for ZMailer (part of which you are reading right now) is
maintained in @emph{texinfo} format.  To format this for a high-quality output
device requires that you already have @TeX{} running, and that you have the
Texinfo macro package installed in the @TeX{} macro library.  If not, these
macros are part of every GNU Emacs distribution (and included with the
current ZMailer distribution).  Generating a line printer or screen version
of the documentation requires the aid of GNU Emacs (see @file{doc/Makefile}).
If you have neither @TeX{} nor GNU Emacs, ask me for a preformatted version
of the documentation.@refill

There is very little hardcoded configuration information in the ZMailer
programs.  The @file{conf.c} files in the @file{router} and @file{scheduler}
subdirectories of the distribution are the primary locations of static
configuration information.  You should check these files, but there is no
need to change them unless you know what you are doing, and insist.@refill

The only other static global information is kept in the @file{mail.h} header
file in the @file{include} subdirectory.  In an operating system environment
that integrates ZMailer, this file is intended to go in
@file{/usr/include}.@refill

There is some dynamic global information, and other compile time information, 
that needs to be specified somewhere.  The way it is done with ZMailer, is
that you (the installer) edits a global configuration file (@file{Config}),
which contains variable definitions that will propagate to all the makefiles
in the distribution.  These definitions will also appear in a file
@file{/etc/zmailer.conf} that ZMailer programs refer to for global
information.  This information includes for example the locations of the
@file{POSTOFFICE} directory hierarchy, so this facility allows easy
dynamic reconfiguration of some installation parameters.  The file is in
@file{/etc} to increase reconfiguration flexibility for diskless mail
clients.@refill

Canned error and warning messages are kept in the @file{proto/forms}
directory of the distribution.  They should be modified to suit local
preferences.  By default, all errors will be carbon-copied to the postmaster,
a local address that is hopefully defined in the aliases database.  Until
you are comfortable with the ZMailer system, you should probably use the
default forms.@refill

Once the above preliminaries have been taken care of, the time has come for
your computer to earn its keep.  If you run the command:@refill

@example
make it so
@end example

@noindent
the following will happen:

@itemize @bullet
@item
A recursive @code{make clean} is run to scrub the distribution
hierarchy.@refill
@item
The global @code{make} file (@file{Makefile}) is edited to update it with
rules for updating all the @code{make} files in the distribution when the
@file{Config} is modified.@refill
@item
The @file{Config} file is processed into a @code{sed} script, which is then
applied to all the @code{make} files in the directory tree.@refill
@item
All programs are compiled.@refill
@item
Another @code{sed} script constructed from the @file{Config} file is applied
to update the @file{proto/zmailer} shell script.@refill
@item
The @file{POSTOFFICE} directory hierarchy is created, and the canned error
messages from the @file{proto/forms} directory are copied to
@file{~/forms}.@refill
@item
The ZMailer directory hierarchy under @file{/usr/lib} is created, and the
standard configuration and control files and shell scripts from the
@file{proto} directory are copied to that location (referred to as
@file{MAILBIN} in the @file{Config} file).@refill
@item
All the program binaries are installed under the @file{$MAILBIN}
directory.@refill
@item
Finally, another @code{sed} script is applied to the @file{Config} file, to
produce @file{/etc/zmailer.conf}.@refill
@end itemize

Then it is time to get your aliases database working.  If you don't already
have a central aliases database, you should create one.  The minimum
requirement is that the @code{postmaster} address expands to a real account
id.  If you already have a central aliases database, this is typically
because you are currently running Sendmail.  In that case, start by copying
the Sendmail aliases file to @file{$MAILBIN/db/aliases}.  The ZMailer Router
is used to build the aliases database from the aliases file.  To do this
correctly, the Router must know what kind of aliases database to access and,
in this situation, create.  The distribution Router configuration file
will check for the existence of a @file{$MAILBIN/db/aliases.dir} file to
indicate that NDBM or DBM is being used.  If this file is absent, the Router
will access the database using a binary search algorithm on an index
file.@refill

If you are using NDBM, prepare the way for the Router by creating a null
@file{aliases.dir} file in the @file{db} subdirectory.  Then run the Router
to initialize the aliases database (@code{router -i}).  If you get syntax
errors, correct them in the @file{aliases} file.  Eventually the Router will
report some simple counts (a la Sendmail) of defined aliases, indicating it
was successful in initializing the aliases database.@refill

You should now arrange for host-specific information to be made available to
ZMailer.  This is obviously a very site-specific customization.  Although
the method of access and location of such information is defined in the
Router configuration file (which incidentally is @file{$MAILBIN/router.cf}),
certain Transport Agents need to know the hosts' UUCP node name.  This is
read from the file @file{/etc/uucpname} if it exists, and secondarily
obtained from the @code{uname} system call in certain environments.  The
convention of using @file{/etc/uucpname} is due to 4.3BSD UUCP which allows
this as a configuration option.  I recommend this method, since it greatly
increases portability of the UUCP binaries between your machines.@refill

The sample Router configuration file in the distribution, assumes that the
host names is should deliver mail locally for, are listed in the file
@file{$MAILBIN/db/localdelivery}.  For example, in an environment with a mail
server and clients, all hostnames should be listed in this file.  This is
suggested as a convention for how to discover this information, and where.
The sample configuration file should be studied for other guidance of this
sort.@refill

You can finally try running the Router in interactive mode, as illustrated in
the @file{README} file.  This is the stage at which you should start playing
with the configuration file and with the various ZMailer programs.  This is
also an opportune moment (or day) for you to customize or write a Router
configuration file for your host/site.@refill

If you have an @file{/etc/services} file, it should be updated with the
definition of a TCP port used for mail queue querying.  The Scheduler acts as
a server listening on this port, and the @code{mailq} program included with
the distribution will connect with the Scheduler and obtain a dump of the
mail queue by this mechanism.  If your system does not have TCP, a rendezvous
mechanism using named pipes will be used instead.  For systems without either
of these facilities, a prearranged file is used along with a release
protocol when the mail queue dump is completed.@refill

When you are comfortable with the new environment and want to start ZMailer,
there is a shell script provided (@file{$MAILBIN/zmailer}) to carry out the
normal startup functions.  If invoked without arguments, it will start the
Router and Scheduler daemon processes, and the SMTP server process.  The
latter may clash with any running Sendmail daemon if it is also acting as an
SMTP server.  This script may also be invoked with individual arguments like
@code{router} or @code{scheduler} to start up just the specified
process(es).  It may be run from the @file{/etc/rc.local} file to start
ZMailer on reboot.@refill

@section How to write a Router configuration file

Sorry, I don't know what the problem areas will be at this point, so this
section is incomplete.  The following is a quick summary:@refill

The configuration file is read and all statements executed sequentially.
Like any other statement, a function definition is also executed, with the
side effect of defining a function.  All functions must be defined before
use.  Normal statements appearing at the top level in the configuration file
(i.e. not within a function definition), usually have the purpose of setting
up an environment for the rest of the configuration file.  An ``environment''
encompasses global variables (e.g. ``what is my name'') initialized in
assignment statements, and database definitions by the @code{relation}
statement.@refill

There are four instances of magic semantics assumed by the Router:@refill

@itemize @bullet
@item
Setting the hostname by calling the @code{hostname} function, will enable
generation of trace headers (i.e. @samp{Received} and
@samp{Message-Id}).@refill
@item
The aliases database is defined by the @code{aliases} relation.  The value of
a database lookup must be a byte offset into another file that contains the
actual alias definitions.@refill
@item
A @code{router} function must exist.  It takes an address as its one
argument, and returns three values representing the channel, next-host, and
next-address.@refill
@item
A @code{crossbar} function must exist.  It takes two triples (six arguments)
and returns those triples and the name of a rewriting function to be applied
to all the header addresses (seven values).  The argument triples represent
the origin and recipient envelope addresses, and this function is in charge
of rewriting them as appropriate.@refill
@end itemize

Naturally, all function names returned by the @code{crossbar} function must
correspond to a defined function.@refill

Tell me what is missing from this description.  Would a play-by-play of the
sample configuration file be very useful?@refill

@node uasupport, , how-to, top
@appendix User Agent support

@appendixsec Submission Interface

Three C library routines are provided to open (create), abort (remove), and
close (submit) a message file.  Internally, they make use of the stdio package,
and their interface is modelled after it.  The interface definition is:@refill

@group
@example
#include <mail.h>

FILE *mail_open()

int mail_abort(mfp)
FILE *mfp;

int mail_close(mfp)
FILE *mfp;
@end example
@end group

The parameter passed in a @code{mail_abort()} or @code{mail_close()} call is
the value returned by a call to the @code{mail_open()} function.
The routines take care of all the necessary housekeeping.
They are properly used as follows:@refill

@group
@example
...
FILE *mfp;
on exit or interrupt, arrange to call mail_abort(mfp);
if ((mfp = mail_open()) == NULL) {
    ... error handling when message submission is not possible ...
} else {
    ... output the mail message to mfp ...
    if (oops && (mail_abort(mfp) == EOF))
        ... print a message that the abort failed ...
    else if (mail_close(mfp) == EOF)
        ... error handling when message submission fails ...
}
reset behaviour on exit or interrupt
...
char *tmalloc(n) unsigned int n; { return n bytes of memory }
@end example
@end group

Notice the definition of @code{tmalloc()}.  This routine should allocate
memory that will remain usable within the lifetime of the message submission
(i.e. until a @code{mail_abort()} or @code{mail_close()} call).  This allows
a User Agent or other application program that makes many calls to these
routines during its lifetime, to provide an alternate byte allocator that
will not cause them to run out of data space.@refill

Another point to be made is that these routines and all other code in
ZMailer that relinks files, uses @code{link()}/@code{unlink()} combinations
and never the @code{rename()} system call, even if it is available.
Unfortunately, @code{rename()} does not retain the inode number of the
file being renamed.@refill

Finally, although this interface will honour the @strong{FULLNAME} and
@strong{PRETTYLOGIN} environment variables mentioned earlier, a User Agent
can override this mechanism by seeking to byte 0 of the message file and
writing its message data from there.@refill

The system standard header file @file{mail.h}, declares these routines
appropriately.  It contains all the common definitions used in passing
information between the components of ZMailer.  This includes the names
of various directories, the postmaster, and symbolic names for various
keys used in the control file protocol.@refill

@appendixsec Fullname quoting

The library routine that constructs a full user name, does so purely based on
information passed to it.  This means it can be used with the contents of a
GECOS field (everything after a @samp{,} or a @samp{;} is ignored), or some
other arbitrary string, without incurring any unnecessary cost involved in
a password database lookup.  The interface specification is as follows:@refill

@group
@example
char *
fullname(gecos, buf, buflen, login)
        char *gecos;            /* the name we wish to quotify */
        char buf[];             /* place to put the result */
        int buflen;             /* how much space we have */
        char *login;            /* what to use for a login name */
@end example
@end group

The return value from @code{fullname()} is always the value of the
second parameter.  A sample usage would be:@refill

@group
@example
struct passwd *pw;
char buffer[BUFSIZ], *name;
extern char *fullname();

name = fullname(pw->pw_gecos, buffer, sizeof buffer, pw->pw_name);
@end example
@end group

If the fourth parameter is @code{(char *)NULL}, the @code{fullname()} routine
will look for the @strong{USER} and @strong{LOGNAME} environment variables,
in that order, if it needs a login name due to the expansion of a @samp{&} in
the GECOS field.  For example:@refill

@group
@example
fullname("& Kirk", ..., "jim") returns "Jim Kirk".
fullname("James T. &", ..., "kirk") returns "\"James T. Kirk\"".
@end example
@end group

The routine will truncate the text of its return value to fit in the space
available in the buffer.  If there is a leading double-quote, there will also
be a trailing double-quote.  The decision to quote is made according to the
specifications in RFC822 for a @i{phrase}.  In other words, when scanned
according to the lexical rules of RFC822, the return value from
@code{fullname()} will constitute a valid RFC822 @i{phrase}.@refill

@contents
@bye