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
|
Mon Feb 7 15:31:54 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be_include/be_union.h:
* TAO_IDL/be/be_union.cpp:
* TAO_IDL/be_include/be_union_branch.h:
* TAO_IDL/be/be_union_branch.cpp:
* TAO_IDL/be_visitor_union/union_ch.cpp:
* TAO_IDL/be_visitor_union/union_ci.cpp:
* TAO_IDL/be_visitor_union/union_cs.cpp:
* TAO_IDL/be/be_visitor_typecode/typecode_defn.cpp:
The IDL compiler has been generating a
separate member typecode inside the union for each
duplicate case label, but the TypeCode class
contains no state (and therefore no methods using
such state) containing info about multiple case
labels - only about members. Typecode generation was changed
to generate just one typecode for each member with
the value of the first case label encountered for
that member. The private union function _discriminant()
(called by (de)marshaling code) was modified to
return this value no matter what the actual current
discriminant value may be. Thanks to
Christopher Kohlhoff <chris@kohlhoff.com> for sending
in yet another IDL file/test code example that got to
the crux of this matter.
* tests/IDL_Test/union.idl:
* tests/IDL_Test/main.cpp:
Added an expanded version of Chris Kohlhoff's example
to IDL_Test.
Mon Feb 07 12:15:00 2000 Michael Kircher <Michael.Kircher@mchp.siemens.de>
* tests/AMI/run_test.pl: Set the "-x" switch on the client,
this will cause it to shut down properly.
* tests/AMI/simple_client.cpp: Activated the "-x" switch behavior
to shut down the ORB, also added a more descriptive debugging
output for exception handling.
Sat Feb 05 04:24:09 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/Leader_Follower: The leader followers class did not support
nested calls to the event loop, i.e., calling perform_work() or
run() from the servant code. In order to accommodate this
feature the following changes were made:
- server_threads* was renamed to event_loop_threads* since this
name was more appropriate.
- In set_event_loop_thread(), if we are the client thread
running the show, then it is ok and we down have to wait.
- Changes to <leader_> are made more carefully now. Both the
values of <event_loop_thread_> and <client_leader_thread_> are
taken into account since they may signify that we are already
a leader.
Since set_client_thread() and reset_client_thread() both
became more careful, set_client_leader_thread() and
reset_client_leader_thread() became easier. Also, code in the
wait strategies became a bit easier.
- <is_event_loop_thread_> was changed from a flag to
<event_loop_thread_>, a counter that keeps track of how many
nested event loop call we have made.
Thanks to Dave Meyer <dmeyer@lima.alex.std.saic.com> for
suggesting the addition of this missing feature!
* tests/Nested_Event_Loop: A new test for invoking nested event
loop calls from user code, i.e., a call to perform_work() or
run() from the implementation of the servant.
Fri Feb 04 22:28:16 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/GIOP_Server_Request.cpp (parse_header_lite): Recent auto
compile failures to MT_Cubit were caused by the following
change:
Mon Jan 10 12:53:08 2000 Jeff Parsons <parsons@cs.wustl.edu>
Jeff changed TAO_GIOP_ServerRequest::parse_header_std but not
TAO_GIOP_ServerRequest::parse_header_lite. Why this bug which
was there since Jan 10 only surfaced after the following change:
Thu Feb 3 17:47:14 2000 Jeff Parsons <parsons@cs.wustl.edu>
is left as an exercise to the reader ;-)
Fri Feb 4 00:22:51 2000 Pradeep Gore <pradeep@danzon.cs.wustl.edu>
* docs/releasenotes/index.html:
added a section for the Notification Service.
Thu Feb 3 19:06:58 2000 Ossama Othman <ossama@uci.edu>
* tests/DSI_Gateway/test_dsi.cpp (invoke):
Added missing ACE_CHECK after call to ORB::shutdown().
Thu Feb 3 20:55:18 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_union/union_cs.cpp:
Added code generation to initialize the union's
discriminant in the default constructor. This ensures
that, if an uninitialized union is inserted into an
Any, whether by itself or contained in some constructed
type, the Any destructor's call to deep_free() will
execute properly. The discriminator must be set to the
first case label value of the first member declared in
the union. Thanks to Christopher Kohlhoff <chris@kohlhoff.com>
for sending in the example that uncovered this bug.
Thu Feb 3 17:47:14 2000 Jeff Parsons <parsons@cs.wustl.edu>
* tao/DomainC.cpp:
* tao/GIOP.cpp:
* tao/ImplRepoC.cpp:
* tao/InterfaceC.cpp:
* tao/MessagingC.cpp:
* tao/Object.cpp:
* tao/POAC.cpp:
* tao/PolicyC.cpp:
* tao/PollableC.cpp:
* tao/Stub.cpp:
* TAO_IDL/be/be_visitor_operation/operation_cs.cpp:
Another temporary workaround to enable the coexistence of
reliable oneways requests and regular twoway requests until
we have GIOP 1.2 fully in place. There were compatibility
problems with Java servers. Thanks to Lothar Werzinger
<werzinger.lothar@krones.de> for reporting this problem.
Thu Feb 03 16:12:01 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/Leader_Follower.i (~TAO_LF_Server_Thread_Helper): The
charanga (Linux box) Exceptions auto build caught the type
problem described in bug 416. In this case, a client thread
completed its twoway call and while giving up leadership, it
noticed that there were server threads waiting to become the
leaders. Hence it sent a broadcast to the server threads and did
not wake any of the followers. In the meantime, however, the
server threads had timed out, and therefore did not select a new
leader.
So the fix is for the server thread to always select a new
leader if needed, irrespective of whether a server thread was
able to become a leader or it had timed out.
reset_server_thread() was removed and
~TAO_LF_Server_Thread_Helper was given the responsibility.
* tao/ORB_Core.cpp (run): No need to call
server_thread_helper.reset_server_thread() explicitly anymore.
~TAO_LF_Server_Thread_Helper will take care of it.
Thu Feb 3 13:40:42 2000 Nagarajan Surendran <naga@cs.wustl.edu>
* orbsvcs/tests/AVStreams/mpeg/source/server/server.cpp:
* orbsvcs/tests/AVStreams/mpeg/source/server/Machine_Properties.{h,cpp}:
* orbsvcs/tests/AVStreams/mpeg/source/server/Property_Exporter.{h,cpp}:
* orbsvcs/tests/AVStreams/mpeg/source/server/Video_Repository.{h,cpp}:
* orbsvcs/tests/AVStreams/mpeg/source/server/augmented_server.cpp:
Fixed all the TAO_ exception macros to ACE_ exception macros.
* orbsvcs/tests/AVStreams/mpeg/source/server/Makefile:
Added libAVSvc to the LDLIBS and -DTAO_ORBSVCS_HAS_Trader to the
Makefile so that the TAO_String_Hash_Key class doesn't get defined
twice.
Thanks to Alex Luk <fluk7@ie.cuhk.edu.hk> and Kenneth Osenbroch
<kennetho@stud.cs.uit.no> for reporting this.
Thu Feb 3 10:31:05 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Makefile:
Disable the new event channel with old versions of g++, they are
too busted to even try to get them to work.
Wed Feb 2 19:15:09 2000 Darrell Brunsch <brunsch@uci.edu>
* orbsvcs/orbsvcs/orbsvcs.dsp: Added the two filters Carlos
created to the MSVC project file.
Wed Feb 2 17:49:33 2000 Ossama Othman <ossama@uci.edu>
* docs/pluggable_protocols/index.html:
Updated the "Notes from a Real World Pluggable Protocol
Implementation" section so that it no longer states that changes
to ACE and TAO were necessary. Changes to ACE and TAO weren't
necessary for that implementation.
Wed Feb 2 19:17:13 2000 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu>
* docs/pluggable_protocols/index.html: Updated the documentation
to clarify the conditions under which a pluggable protocol can
be developed without affecting ACE+TAO. Thanks to Gilbert
Grosdidier <Gilbert.Grosdidier@in2p3.fr> for motivating this.
Wed Feb 2 17:29:34 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_typecode/typecode_defn.cpp:
One more fix to this file. Nothing was breaking, but
it was pointed out, again by Christopher Kohlhoff
<chris@kohlhoff.com> that, even though the members
with multiple case labels were getting generated only
once, the encapsulated length was calculated as if
they were included more than once.
Wed Feb 2 14:14:52 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/tests/EC_Custom_Marshal/ECM_Supplier.cpp:
Fixed race condition during shutdown. The EC can die before the
supplier has a chance to disconnect. In that case when trying
to disconnect the supplier receives an exception that should be
ignored.
Wed Feb 2 13:12:07 2000 Carlos O'Ryan <coryan@uci.edu>
* docs/releasenotes/ec.html:
Updated.
Wed Feb 2 14:03:00 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_typecode/typecode_defn.cpp
My fix of Feb. 1 regarding multiple case labels and
typecode generation broke two use cases. One, reported
by Christopher Kohlhoff <chris@kohlhoff.com> was when
a previously defined struct is nested inside a union.
The other, when a union is recursive, appeared when
the DII versions of the param tests were failing. This
fix restores generation of correct typecodes for unions
with nested structs, recursive members, and when any of
its members (nested, recursive or vanilla) has multiple
case labels
Wed Feb 2 08:39:29 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_Default_Factory.cpp:
Some platforms do no recursiong when instantiating templates
(that is not ANSI-compliant, btw), I added missing template
instantiations those platforms.
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.cpp:
Fixed warning about signed/unsigned comparisons.
Tue Feb 1 19:56:08 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/fe/idl.yy:
* TAO_IDL/fe/y.tab.cpp:
* TAO_IDL/fe/y.tab.h:
Added code to yacc file to prevent a crash when an
IDL file set a union's discriminator type to octet,
or typedef of octet. It now just outputs the error
message. This closes out bug #121.
Tue Feb 1 17:43:09 2000 Nagarajan Surendran <naga@cs.wustl.edu>
* orbsvcs/tests/AVStreams/mpeg/source/mpeg_client/Command_Handler.cpp:
* orbsvcs/tests/AVStreams/mpeg/source/mpeg_client/ctr.cpp:
Fixed compile errors and warning with egcs-1.1.1 on Sun OS 5.7.
Thanks to Alex Luk <fluk7@ie.cuhk.edu.hk> and Kenneth Osenbroch
<kennetho@stud.cs.uit.no> for reporting this.
Tue Feb 1 16:50:28 2000 Nagarajan Surendran <naga@cs.wustl.edu>
* orbsvcs/tests/AVStreams/mpeg_shared/com.cpp:
* orbsvcs/tests/AVStreams/mpeg_shared/fileio.cpp:
* orbsvcs/tests/AVStreams/mpeg_shared/filters.cpp:
* orbsvcs/tests/AVStreams/mpeg_shared/routine.cpp:
* orbsvcs/tests/AVStreams/mpeg_shared/sendpt.cpp:
* orbsvcs/tests/AVStreams/mpeg_mib/mibDrawArea.cpp:
Fixed warnings for egcs on SunOS5.7.
Tue Feb 1 13:51:21 2000 Carlos O'Ryan <coryan@uci.edu>
* docs/ec_options.html:
Documented the compile time options for the RTEC.
Tue Feb 1 14:14:11 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be_include/be_union_branch.h:
* TAO_IDL/be/be_union_branch.cpp:
* TAO_IDL/be/be_visitor_typecode/typecode_defn.cpp:
Prevented multiple generations of typecode for a
union member with multiple case labels. The CORBA
spec requires that the limit for an index supplied
to member_label, member_type, etc. not exceed the
member count, otherwise a Bounds exception is thrown.
The TAO CDR_Interpreter operates on the same assumption,
and it was crashing on unions with members having
multiple case labels. Thanks to Christopher
Kohlhoff <chris@kohlhoff.com> for reporting this bug.
Tue Feb 1 11:25:00 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_Type_Filter.cpp:
More fixes in the can_match() routine, this is why we need
something better that "0" as a wildcard for source and/or type.
* orbsvcs/tests/EC_Custom_Marshal/ECM_Supplier.cpp:
It was using the wrong object reference to disconnect
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.h:
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.cpp:
* orbsvcs/tests/Event/lib/Counting_Consumer.cpp:
* orbsvcs/tests/Event/lib/Counting_Supplier.cpp:
Fixed several bugs in the connect/disconnect calls and their
use.
* orbsvcs/orbsvcs/Event/EC_Basic_Filter_Builder.cpp:
* orbsvcs/orbsvcs/Event/EC_ConsumerAdmin.cpp:
* orbsvcs/orbsvcs/Event/EC_Defaults.h:
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.cpp:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.cpp:
Added some debugging messages (that are not compiled by
default).
* orbsvcs/tests/EC_Throughput/ec.conf:
* orbsvcs/tests/Event/Basic/svc.conf:
Fixed configuration scripts.
Mon Jan 31 19:44:02 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_Default_Factory.cpp:
More tweaking to get the templates to compile on every
platform.
Mon Jan 31 21:32:14 2000 Carlos O'Ryan <coryan@cs.wustl.edu>
* orbsvcs/tests/Event/lib/ECTest.dsp:
Add the new files to the MSVC project file.
Mon Jan 31 19:32:56 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/tests/Event/Event.dsw:
* orbsvcs/tests/Event/Basic/Basic.dsw:
* orbsvcs/tests/Event/Basic/Bitmask.dsp:
* orbsvcs/tests/Event/Basic/Disconnect.dsp:
* orbsvcs/tests/Event/Basic/Negation.dsp:
* orbsvcs/tests/Event/Basic/Timeout.dsp:
* orbsvcs/tests/Event/Basic/Wildcard.dsp:
Updated the project files and workspaces, included the new
project for the bitmask test.
Mon Jan 31 19:02:34 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_Bitmask_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Bitmask_Filter.i:
* orbsvcs/orbsvcs/Event/EC_Bitmask_Filter.cpp:
This filter quickly rejects events that do not match a given
bitmask.
If the event is not rejected based on the mask then the child is
consulted to finally accept or reject the event.
When composed with the Null_Filter it accepts any events that
satisfy:
(event.header.type & type_mask) != 0
&& (event.header.type & source_mask) != 0
* orbsvcs/orbsvcs/Event/EC_Masked_Type_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Masked_Type_Filter.i:
* orbsvcs/orbsvcs/Event/EC_Masked_Type_Filter.cpp:
This filter only accepts events whose type and/or source have a
given value when a bitmask is applied to them.
In short the filter checks that satisfy:
(event.header.type & type_mask) == type_value
&& (event.header.source & source_mask) == source_value
* orbsvcs/orbsvcs/Makefile:
* orbsvcs/orbsvcs/Event/EC_Basic_Filter_Builder.cpp:
The basic factory can create the new filters.
* orbsvcs/orbsvcs/Event_Service_Constants.h:
* orbsvcs/orbsvcs/Event_Utilities.h:
* orbsvcs/orbsvcs/Event_Utilities.i:
* orbsvcs/orbsvcs/Event_Utilities.cpp:
Add support for the new filters into the ConsumerQoS factory.
* orbsvcs/tests/Event/Basic/Makefile:
* orbsvcs/tests/Event/Basic/Bitmask.cpp:
* orbsvcs/tests/Event/Basic/run_test.pl:
New test for the bitmask filters.
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.h:
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.cpp:
* orbsvcs/tests/Event/Basic/Disconnect.h:
* orbsvcs/tests/Event/Basic/Disconnect.cpp:
* orbsvcs/tests/Event/Basic/MT_Disconnect.h:
* orbsvcs/tests/Event/Basic/MT_Disconnect.cpp:
* orbsvcs/tests/Event/Basic/Negation.h:
* orbsvcs/tests/Event/Basic/Negation.cpp:
* orbsvcs/tests/Event/Basic/Timeout.h:
* orbsvcs/tests/Event/Basic/Timeout.cpp:
* orbsvcs/tests/Event/Basic/Wildcard.h:
* orbsvcs/tests/Event/Basic/Wildcard.cpp:
* orbsvcs/tests/Event/Performance/Makefile:
* orbsvcs/tests/Event/lib/Makefile:
* orbsvcs/tests/Event/lib/Counting_Consumer.h:
* orbsvcs/tests/Event/lib/Counting_Consumer.cpp:
* orbsvcs/tests/Event/lib/Counting_Supplier.h:
* orbsvcs/tests/Event/lib/Counting_Supplier.cpp:
Refactor several tests, they all had some supplier or consumer
that simply counted events. The shared class is now in the
library.
Mon Jan 31 20:21:24 2000 Douglas C. Schmidt <schmidt@tango.cs.wustl.edu>
* Added the following changes for Borland C++Builder 4:
TAO/orbsvcs/orbsvcs/Makefile.bor: Added new files.
TAO/tests/MProfile/test_i.cpp, test_i.h: Added definition of
missing constructor, added required ACE_THROW_SPEC and removed
reference to missing inline file.
TAO/tests/Param_Test/client.bor: Added new file
recursive_union.obj.
Thanks to Christopher Kohlhoff <chris@kohlhoff.com> for
contributing these.
Mon Jan 31 20:29:51 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/fe/idl.ll:
* TAO_IDL/fe/idl.yy:
* TAO_IDL/fe/lex.yy.cpp:
* TAO_IDL/fe/lex.yy.cpp.diff:
* TAO_IDL/fe/y.tab.h:
* TAO_IDL/fe/y.tab.cpp:
Added keyword 'fixed' to lexer and parser.
* TAO_IDL/util/utl_scope.cpp:
Modified lookup_by_name_local() to output an error if it sees
any case variation of 'fixed'. The change to the lexer and
parser will catch the use of 'fixed' as a type. This change will
catch the use of 'fixed' as an identifier. Fixed types are not
implemented in TAO, so we need to output an error message
whenever the word is used in an IDL file. Thanks to J. Scott
Evans <evans@cpi.com> for reporting this bug.
Mon Jan 31 20:15:25 2000 Carlos O'Ryan <coryan@cs.wustl.edu>
* orbsvcs/orbsvcs/Event/EC_Concrete_Proxy_Set.h:
* orbsvcs/orbsvcs/Event/EC_Basic_Factory.cpp:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.cpp:
* orbsvcs/orbsvcs/Event/EC_Null_Factory.cpp:
Had to expand several typedefs to workaround problems with SunCC
4.2 apparently the compiler cannot deal with nested typedefs
inside templates very well.
* orbsvcs/orbsvcs/Event/EC_Proxy_Collection.h:
Fixed a comment.
Mon Jan 31 20:13:58 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_Concrete_Proxy_Set.h:
Added missing ACE_TYPENAME macros, only HP/aCC seems to detect
the problem.
* orbsvcs/orbsvcs/Event/EC_QOS_Info.h:
* orbsvcs/orbsvcs/Event/EC_QOS_Info.i:
Workaround a bug in the HP/aCC compiler, unless the copy
constructor is explicitly defined the compiler crashes when
trying to generate the code for it.
Mon Jan 31 13:21:48 2000 Darrell Brunsch <brunsch@uci.edu>
* orbsvcs/tests/ImplRepo/run_test.pl:
The both_ir test was leaving servers running. Added shutdown
commands for them at the end of the test.
* tao/POA.cpp:
Noticed that servers using the IMR were not shutting down properly
on non-Win32 platforms. This happened because the code was grabbing
locks twice, so it deadlocked when the default lock wasn't a thread
mutex. It now calls the equivalent *_i methods.
Mon Jan 31 14:54:14 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tests/Leader_Followers: The thread-per-connection TAO server has
problems when shutting down since there is no portable way to
wakeup a thread blocked in read(). Therefore, changed the
server to use a thread-pool instead of a thread-per-connection
approach.
Mon Jan 31 11:14:21 2000 Carlos O'Ryan <coryan@cs.wustl.edu>
* orbsvcs/orbsvcs/orbsvcs.dsp:
* orbsvcs/orbsvcs/orbsvcs_static.dsp:
Updated the project files.
Mon Jan 31 08:25:51 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_Concrete_Proxy_Set.i:
Don't use typedefs inside templates with return values, it
confuses g++ 2.7.X
* orbsvcs/orbsvcs/Event/EC_ObserverStrategy.h:
Some typedefs were protected, and thus not accessible for some
classes.
* orbsvcs/orbsvcs/Event/EC_SupplierAdmin.cpp:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.cpp:
Move some template instantiations around.
* orbsvcs/orbsvcs/Event/EC_Worker.h:
* orbsvcs/orbsvcs/Event/EC_Worker.cpp:
Try to work around old g++ bug.
* orbsvcs/tests/EC_Throughput/ECT_Throughput.cpp:
* orbsvcs/tests/EC_Throughput/Makefile:
Remove unnecessary header file and updated dependencies.
* orbsvcs/orbsvcs/Event/EC_Proxy_Collection.h:
* orbsvcs/orbsvcs/Event/EC_Negation_Filter.h:
Fix typos in the comments.
Sun Jan 30 19:59:30 2000 Carlos O'Ryan <coryan@uci.edu>
* Changed the basic collection classes that keep both the consumer
and supplier proxies. The new scheme uses the strategy pattern:
the concrete collection (such as a list or RB-tree) is the
strategy to store/retrieve objects, while different classes
implement the serialization protocol (such as using
copy-on-read, or delayed changes).
This new implementation uses templates, because the code is
generic for both the consumer and supplier proxies. It is
possible that the same implementation can be used for other
Event Services (such as Notification or CosEvents)
* orbsvcs/orbsvcs/Event/EC_Proxy_Collection.h:
* orbsvcs/orbsvcs/Event/EC_Proxy_Collection.i:
* orbsvcs/orbsvcs/Event/EC_Proxy_Collection.cpp:
Implement the different serialization protocols for the proxy
collections.
In this new implementation there are no iterators exposed to the
rest of the EC, instead a for_each() method is implemented, each
component that needs to iterate over a consumer or supplier
collection implements a different Worker object for the task.
* orbsvcs/orbsvcs/Event/EC_Concrete_Proxy_Set.h:
* orbsvcs/orbsvcs/Event/EC_Concrete_Proxy_Set.i:
* orbsvcs/orbsvcs/Event/EC_Concrete_Proxy_Set.cpp:
Implement the concrete collections, currently we have two of
them, one based on RB-trees and another based on an ordered
list.
* orbsvcs/orbsvcs/Event/EC_Worker.h:
* orbsvcs/orbsvcs/Event/EC_Worker.i:
* orbsvcs/orbsvcs/Event/EC_Worker.cpp:
Defines the interfaces for the Worker objects.
* orbsvcs/orbsvcs/Event/EC_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Event_Channel.h:
* orbsvcs/orbsvcs/Event/EC_Event_Channel.i:
Modified to create the new collection classes.
Removed the methods to create consumer and supplier admin locks,
because they have been superseeded by the collection classes.
* orbsvcs/orbsvcs/Event/EC_Defaults.h:
Added support for the new collection scheme, and removed old
defaults for the previous style.
* orbsvcs/orbsvcs/Event/EC_Default_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.i:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.cpp:
Added support for the new proxy collections.
The factory can create both list-based or RB-tree-based
collections, decorated with any of the three basic algorithms
(i.e. simple locking, aka immediate changes, delayed changes or
copy-on-read)
The file contains most of the template instantiations, must fix
that.
* docs/ec_options.html:
The options in the default resource factory have changed, to
handle the new collection types.
* orbsvcs/orbsvcs/Event/EC_ConsumerAdmin.h:
* orbsvcs/orbsvcs/Event/EC_ConsumerAdmin.i:
* orbsvcs/orbsvcs/Event/EC_ConsumerAdmin.cpp:
* orbsvcs/orbsvcs/Event/EC_SupplierAdmin.h:
* orbsvcs/orbsvcs/Event/EC_SupplierAdmin.i:
* orbsvcs/orbsvcs/Event/EC_SupplierAdmin.cpp:
* orbsvcs/orbsvcs/Event/EC_ObserverStrategy.h:
* orbsvcs/orbsvcs/Event/EC_ObserverStrategy.i:
* orbsvcs/orbsvcs/Event/EC_ObserverStrategy.cpp:
* orbsvcs/orbsvcs/Event/EC_SupplierControl.h:
* orbsvcs/orbsvcs/Event/EC_SupplierControl.cpp:
* orbsvcs/orbsvcs/Event/EC_Reactive_ConsumerControl.h:
* orbsvcs/orbsvcs/Event/EC_Reactive_ConsumerControl.i:
* orbsvcs/orbsvcs/Event/EC_Reactive_ConsumerControl.cpp:
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.h:
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.i:
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.cpp:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.i:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.cpp:
* orbsvcs/orbsvcs/Event/EC_Trivial_Supplier_Filter.cpp:
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.i:
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.cpp:
Changed all the iterations in the Event Service implementation
to use Workers instead of iterators.
* orbsvcs/orbsvcs/Event/EC_Command.h:
* orbsvcs/orbsvcs/Event/EC_Command.i:
* orbsvcs/orbsvcs/Event/EC_Command.cpp:
Added new Reconnected command for delayed reconnections.
Normalized the name of the Shutdown command.
* orbsvcs/orbsvcs/Event/EC_Basic_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Basic_Factory.cpp:
* orbsvcs/orbsvcs/Event/EC_Null_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Null_Factory.cpp:
Create the equivalent proxy collections corresponding to the old
strategies.
* docs/tutorials/Quoter/Event_Service/ec.conf:
* orbsvcs/CosEvent_Service/svc.conf:
* orbsvcs/Event_Service/svc.conf:
* orbsvcs/examples/RtEC/Schedule/svc.conf:
* orbsvcs/examples/RtEC/Simple/ec.conf:
* orbsvcs/tests/EC_Mcast/svc.conf:
* orbsvcs/tests/EC_Throughput/ec.conf:
* orbsvcs/tests/EC_Throughput/ec.mt.conf:
* orbsvcs/tests/Event/Basic/mt.svc.conf:
* orbsvcs/tests/Event/Basic/observer.conf:
* orbsvcs/tests/Event/Basic/sched.conf:
* orbsvcs/tests/Event/Basic/svc.conf:
* orbsvcs/tests/Event/Performance/ec.mt.conf:
* orbsvcs/tests/Event/Performance/ec.st.conf:
* orbsvcs/tests/Event/Performance/ec.rb_tree.conf:
* orbsvcs/tests/Event/Performance/ec.list.conf:
All the configuration files were updated to avoid any obsolete
flags.
* orbsvcs/orbsvcs/Event/EC_Dispatching_Task.h:
* orbsvcs/orbsvcs/Event/EC_Dispatching_Task.i:
* orbsvcs/orbsvcs/Event/EC_Dispatching_Task.cpp:
* orbsvcs/orbsvcs/Event/EC_MT_Dispatching.cpp:
* orbsvcs/orbsvcs/Event/EC_Priority_Dispatching.cpp:
Normalized the name of the Shutdown_Command, the one to shutdown
a task is now called Shutdown_Task.
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set.h:
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set.i:
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set_T.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set_T.h:
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set_T.i:
Removed, the EC_Proxy_Collection and EC_Concreate_Proxy_Set
classes have superseeded this implementation.
* orbsvcs/tests/Event/lib/Consumer.h:
* orbsvcs/tests/Event/lib/Consumer.cpp:
* orbsvcs/tests/Event/lib/Supplier.h:
* orbsvcs/tests/Event/lib/Supplier.cpp:
Optimized activation and de-activation of consumers and
suppliers.
Sun Jan 30 18:52:45 2000 Darrell Brunsch <brunsch@uci.edu>
* orbsvcs/ImplRepo_Service/ImplRepo_i.cpp:
* orbsvcs/ImplRepo_Service/ImplRepo_i.h:
* orbsvcs/ImplRepo_Service/Options.cpp:
* orbsvcs/ImplRepo_Service/Options.h:
Added a timeout to the startup code so if the startup
server does not respond within the timeout, then kill
it and return an exception. The timeout period can
be controlled by the -t option.
Also replaced the ACE_Process code with the ACE_Process_Manager.
This should automatically clean up the child processes
when they exit.
* orbsvcs/ImplRepo_Service/tao_imr_i.cpp:
Prints out more information when a CannotActivate exception
occurs.
Sat Jan 29 12:57:53 2000 bala <bala@cs.wustl.edu>
* TAO version 1.0.13 released.
Fri Jan 28 20:11:23 2000 Carlos O'Ryan <coryan@uci.edu>
* tao/ORB_Core.cpp:
When the last ORB is detroyed it must reset the first_orb_
variable in TAO_ORB_Table to 0. Thanks to Alex for reporting
this bug and providing a fix.
Fri Jan 28 18:35:15 2000 Jeff Parsons <parsons@cs.wustl.edu>
* tao/Request.cpp:
Removed the check on work_pending() in method
get_repsonse(), used in deferred synchronous requests.
Recent changes to perform_work() and/or work_pending()
have made it possible for work_pending() to return 0
even if the DII request has not completed (received its
response). This was causing get_response() to return
TRUE erroneously, and the resulting null message
block pointer would eventually cause a client crash.
Now get_response() returns TRUE only if the
repsonse has been received.
Fri Jan 28 17:27:19 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_sequence/cdr_op_cs.cpp:
Fix sent in by Alex Arulanthu <Alex.Arulanthu@sylantro.com> that
plugs a memory leak in generated code when decoding a sequence
of length 0. Thanks to Jianfei Xu <jxu@yahoo.com> for reporting
this.
Fri Jan 28 14:05:32 2000 Jeff Parsons <parsons@cs.wustl.edu>
* performance-tests/Latency/st_client.cpp:
Added 'x' to the string passed to get_opts(). It was in
the switch statement and in the usage message, as it is
in the other client types in this test suite.
Fri Jan 28 12:51:26 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tests/Leader_Followers: Minor changes to make it work with
different compilers.
Thu Jan 27 20:25:37 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/Leader_Follower:
wait_for_client_leader_to_complete: If a client thread (a thread
making a remote request) is currently running the event loop, an
event loop thread (a thread running ORB->run or
ORB->perform_work) must wait for that client thread to complete.
Additional state: <client_thread_is_leader_> keep track of
whether a client thread the current leader.
<server_threads_waiting_> keeps track of whether server threads
waiting for the client leader to complete.
<server_threads_condition_> is the condition variable used by
event loop threads waiting for the client leader to complete.
TAO_LF_Server_Thread_Helper: Auto-pointer like class similar to
TAO_LF_Client_Thread_Helper and TAO_LF_Leader_Thread_Helper that
helps manage the leader follower object. This class helps with
grabbing the lock and calling set_server_thread and grabbing the
lock, calling reset_server_thread, and electing a new leader.
set_server_thread: This method now calls
wait_for_client_leader_to_complete if needed. Also it now has a
relevant return value for timeouts and errors.
set_leader_thread and reset_leader_thread: Added the setting and
resetting <client_thread_is_leader_>.
elect_new_leader: Election of new leader must first consider
event loop threads before considering waiting clients. If
server threads are waiting, we wake them all up. Note that it
is necessary to give precedence to server threads over client
threads since only one client runs at any given time; with a tp
reactor, multiple server threads can execute simultaneously.
* tao/ORB_Core.cpp (run): Use of TAO_LF_Server_Thread_Helper
helped with abstracting and simplifying some of the leader
follower code. In additions, timeouts are determined by
examining the <timeout> variable. Therefore, we don't need
<break_on_timeouts> anymore.
* tao/ORB (run and perform_work): The new ORB_Core::run simplified
ORB::run and ORB::perform_work.
* tao/Wait_Strategy.cpp (TAO_Wait_On_Reactor::wait,
TAO_Exclusive_Wait_On_Leader_Follower::wait, and
TAO_Muxed_Wait_On_Leader_Follower::wait): The while loop in
TAO_Wait_On_Reactor::wait and
TAO_Exclusive_Wait_On_Leader_Follower::wait was incorrect.
This:
*max_wait_time != ACE_Time_Value::zero
should have been:
*max_wait_time == ACE_Time_Value::zero
While fixing this, I simplified the while loop making it easier
to read and understand.
Also fixed TAO_Muxed_Wait_On_Leader_Follower::wait which did not
handle timeouts correctly.
* tests/Leader_Followers: This is a test for server applications
that have client threads (threads making remote calls) starting
before server threads (threads running the event loop).
The event loop threads should wait for the client threads that
become leaders to give up leadership before continuing.
* version_tests.dsw:
* tests/Makefile (DIRS):
* run_all_list.pm:
Added new test Leader_Followers.
Thu Jan 27 08:36:17 2000 Darrell Brunsch <brunsch@uci.edu>
* orbsvcs/ImplRepo_Service/tao_imr_i.cpp: The help messages
were still displaying "tao_ir" instead of "tao_imr". Also,
if an Implementation Repository wasn't running, tao_imr
wouldn't get a chance to parse the command line arguments -
so it wouldn't even display the help messages unless an
ImplRepo was running. Thanks to Dave Zumbro
<zumbro_d@ociweb.com> for reporting these. [Bugs 412, 413]
Wed Jan 26 16:17:24 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_scope.cpp:
* TAO_IDL/be/be_visitor_interface_fwd/interface_fwd_ch.cpp:
* TAO_IDL/be/be_visitor_interface_fwd/interface_fwd_ci.cpp:
* TAO_IDL/be/be_visitor_interface_fwd/cdr_op_ch.cpp:
A fix for what I broke yesterday. At least we should now
again be able to compile files where a forward declared
interface is defined later in the file.
* TAO_IDL/be/be_visitor_interface/cdr_op_ch.cpp:
Cosmetic changes.
* tests/IDL_Test/interface.idl:
Commented out the forward declared interface that's not
defined in the same file for now, until it is working on
more platforms.
Tue Jan 25 19:42:54 2000 Balachandran Natarajan <bala@cs.wustl.edu>
* tao/UIOP_Profile.cpp (decode):
* tao/IIOP_Profile.cpp : A problem when TAO clients were
interacting with GIOP 1.2 servers. The profile received by the
TAO client is parsed in decode (). In the code for IIOP_Profile,
a check was made for a higher minor version but no action was
taken. In the process, the version information that the
IIOP_Profile class was holding had the minor version information
of the received profile and not what the TAO client can
send. So, changes have been made to check for the minor version
number. If the minor version is less than or equal to the minor
version that TAO supports, we save that information. If it is
greater than the one we support, the {IIOP, UIOP}_Profile
classes would highest minor version that we support. This
information gets passed on through the invocation classes to the
transport and then to the messaging classes. Thanks to Klemen
Zagar <klemen.zagar@irj.si> for reporting this problem.
Tue Jan 25 19:27:24 2000 Jeff Parsons <parsons@cs.wustl.edu>
* docs/releasenotes/index.html:
Added entry for forward declared interface change to
IDL compiler.
Tue Jan 25 19:01:33 2000 Carlos O'Ryan <coryan@cs.wustl.edu>
* orbsvcs/orbsvcs/Event/EC_Event_Channel.cpp:
Fixed shutdown problem for the Timer Module.
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.cpp:
Fixed memory leak, if the supplier disconnects while some
consumers are still attached it has to shutdown() its filter to
clear the extra reference counts.
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.cpp:
* orbsvcs/tests/Event/Basic/MT_Disconnect.cpp:
* orbsvcs/tests/Event/Basic/Negation.cpp:
* orbsvcs/tests/Event/Basic/Observer.cpp:
* orbsvcs/tests/Event/lib/Driver.cpp:
Fixed several memory leaks in the tests, to verify that the
service is ok.
* orbsvcs/tests/Event/Basic/run_test.pl:
Increased allowed execution time for one of the tests.
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set.cpp:
Cosmetic fixes
Tue Jan 25 18:51:42 2000 Jeff Parsons <parsons@cs.wustl.edu>
* tests/IDL_Test/interface.idl:
* tests/IDL_Test/generic_object.idl:
Added tests for new forward declared interface rule
mentioned below and for generic objects used as
parameters in a request.
Tue Jan 25 18:19:49 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/tao_idl.cpp:
* TAO_IDL/ast/ast_interface_fwd.cpp:
* TAO_IDL/be/be_visitor_scope.cpp:
* TAO_IDL/be/be_visitor_factory.cpp:
* TAO_IDL/be/be_visitor_root/root.cpp:
* TAO_IDL/be/be_visitor_interface_fwd/interface_fwd_ch.cpp:
* TAO_IDL/be/be_visitor_interface_fwd.cpp:
* TAO_IDL/be_include/be_visitor_interface_fwd.h:
Modified files
* TAO_IDL/be_include/be_visitor_interface_fwd/cdr_op_ch.h:
* TAO_IDL/be/be_visitor_interface_fwd/cdr_op_ch.cpp:
New files.
Changes to support code generation and C++ compilation
for IDL files containing a forward declared interface
that is not defined in the same file. The spec had
required that the definition be present, but the rule
was recently relaxed in CORBA 2.3.1 (99-10-07). If
the full definition is not in some generated .cpp file
that is included in the build, the C++ compiler will
produce a link error. This closes Bugzilla entry #401.
Tue Jan 25 14:23:04 2000 Ossama Othman <ossama@uci.edu>
* docs/pluggable_protocols/index.html:
Updated/corrected some URLs.
Tue Jan 25 14:05:42 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/Leader_Follower.h (TAO_LF_Client_Thread_Helper and
TAO_LF_Leader_Thread_Helper): Auto-pointer like classes that
help manage the leader follower object.
TAO_LF_Client_Thread_Helper helps with set_client_thread and
reset_client_thread while TAO_LF_Leader_Thread_Helper helps with
set_leader_thread and reset_leader_thread.
* tao/Wait_Strategy.cpp (TAO_Muxed_Wait_On_Leader_Follower::wait
and TAO_Exclusive_Wait_On_Leader_Follower::wait):
We now use guards to make sure that the leader follower object
has consistent information even when errors occur. Previously,
there were many return and exceptional conditions where the
leader follower object was left in an inconsistent state.
Tue Jan 25 12:17:35 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_operation/rettype_vardecl_cs.cpp:
Missed some cases in the previous checkin of this file.
Tue Jan 25 10:38:31 2000 Pradeep Gore <pradeep@flamenco.cs.wustl.edu>
* orbsvcs/orbsvcs/Notify/Notify_SupplierAdmin_i.cpp:
* orbsvcs/orbsvcs/Notify/Notify_ConsumerAdmin_i.cpp:
Fixed "unreachable statement" warnings on kai.
Tue Jan 25 02:07:58 2000 Nanbor Wang <nanbor@cs.wustl.edu>
* examples/Simple/time/README: Added description on getting
JACE.Misc.GetOpt.
* examples/Simple/time/Time_Client.java: Modified the Java client
so that it, like C++ client, can use command line args to
determine ways of acquiring Time objref. This test now uses
JACE.Misc.GetOpt (a complete JACE installation is not required.)
Mon Jan 24 18:47:22 2000 Ossama Othman <ossama@uci.edu>
* docs/Options.html:
Clarified what UID and GID in the documentation for the
-ORBSetUID and -ORBSetGID options.
Mon Jan 24 13:17:31 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_operation/rettype_vardecl_cs.cpp:
Russell L. Carter <rcarter@pinyon.org> reported a bug when
the return type of an operation is CORBA::Object. It turns
out that CORBA::TypeCode, interface, forward declared interface,
valuetype, and forward declared valuetype needed the same fix.
Sun Jan 23 18:15:32 2000 Krishnakumar Elakkara Pathayapura <krish@polka.cs.wustl.edu>
* docs/releasenotes/index.html:
Updated the release notes for the Logging Service
Sun Jan 23 14:09:30 2000 Ossama Othman <ossama@uci.edu>
* docs/Options.html:
Added descriptions of the newly added "-ORBSetUID" and
"-ORBSetGID" ORB options.
* tao/ORB_Core.cpp (init):
Added two new ORB options "-ORBSetUID" and "-ORBSetGID." These
options can be used to the set the effective user and group IDs,
respectively. This is useful when starting the ORB under the
super-user but it is desired to run the ORB under different user
and/or group IDs. A typical use would be to start a TAO daemon,
under the 'nobody' user via System V style initialization
scripts.
Note that these options are only supported if the platform
implements the setuid() and setgid() POSIX system calls.
In general, it is a good idea to only run the ORB with
super-user privileges when necessary, e.g. when performing
benchmarking experiments, and when access to the real-time
scheduling class is needed. If the ORB must be started via the
super-user account, and the real-time scheduling class won't be
used, than the ORB's effective user ID should at least be
changed to an unprivileged one, such as 'nobody'.
Sat Jan 22 21:08:07 2000 Nanbor Wang <nanbor@cs.wustl.edu>
* orbsvcs/orbsvcs/orbsvcs.dsp:
* orbsvcs/orbsvcs/orbsvcs_static.dsp: Renamed
EC_SupplierFiltering.* to EC_Supplier_Filter.*.
Sat Jan 22 18:25:51 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.cpp:
Fixed typo in the inline include file, only a problem in builds
that disable inlining.
Sat Jan 22 17:04:14 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_ConsumerAdmin.h:
* orbsvcs/orbsvcs/Event/EC_ConsumerAdmin.cpp:
* orbsvcs/orbsvcs/Event/EC_Event_Channel.h:
* orbsvcs/orbsvcs/Event/EC_Event_Channel.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.h:
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.h:
* orbsvcs/orbsvcs/Event/EC_SupplierAdmin.h:
* orbsvcs/orbsvcs/Event/EC_SupplierAdmin.cpp:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.cpp:
* orbsvcs/orbsvcs/Event/EC_Trivial_Supplier_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Trivial_Supplier_Filter.cpp:
Added support for atomic reconnections. Before this change the
event channel could loose messages when a consumer reconnected,
because the implementation basically disconnected the consumer
and then connected it again. There was an interval where the
suppliers could push events, while the consumer was temporarily
disconnected.
The new implementation pushes a reconnected() message through
the event channel, the consumer and supplier admins and finally
to each proxy.
There were no changes on the critical path, or regular
connect/disconnect calls, so this only affect the performance of
reconnections.
This fixes bug #271
* orbsvcs/orbsvcs/Event/EC_Reactive_ConsumerControl.cpp:
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.cpp:
Another attempt at fixing this code to compile without CORBA
Messaging support. It happens that ACE_HAS_CORBA_MESSAGING is
always defined, but sometimes to 0 and other times to 1.
* orbsvcs/tests/Event/Event.dsw:
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.cpp:
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.dsp:
* orbsvcs/tests/Event/Basic/Atomic_Reconnect.h:
* orbsvcs/tests/Event/Basic/Basic.dsw:
* orbsvcs/tests/Event/Basic/Makefile:
* orbsvcs/tests/Event/Basic/run_test.pl:
New test to check that Atomic reconnects actually work.
* orbsvcs/tests/Event/Basic/mt_disconnect.conf:
* orbsvcs/tests/Event/Basic/mt.svc.conf:
Renamed because it is also used in the Atomic_Reconnect test.
Sat Jan 22 17:11:30 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_codegen.cpp:
Brian Harris <harris_b@ociweb.com> reported that when using
the -GI option with the IDL compiler to generate implementation
files for an IDL file 'foo', an included IDL file 'bar' caused
the generation of #include "barI.h" in the implementation header
file, when it should be #include "barC.h". However, since barC.h
is indirectly included by the line #include "fooS.h", it does
not need to be there at all. The block that generates this line
in be_codegen.cpp is commented out for now, to see if some use
case arises where it is needed. This closes out bug 410.
Fri Jan 21 21:19:26 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Makefile:
Updated dependencies again, i forget LifeCycle last time.
Fri Jan 21 16:46:36 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event_Utilities.cpp:
* orbsvcs/orbsvcs/Event_Utilities.h:
* orbsvcs/orbsvcs/Event_Utilities.i:
The QoS factories now take an (optional) initializer function
from the user to help with customized Event Channels that
transmit a union as part of the EventPayload.
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.h:
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.cpp:
Fixed some very obscure concurrency problems while disconnecting
suppliers: if a supplier and a consumer where disconnecting at
the same time there was a very small window for a crash. The
race condition has been eliminated.
* orbsvcs/tests/Event/Event.dsw:
* orbsvcs/tests/Event/Basic/Basic.dsw:
* orbsvcs/tests/Event/Basic/MT_Disconnect.cpp:
* orbsvcs/tests/Event/Basic/MT_Disconnect.dsp:
* orbsvcs/tests/Event/Basic/MT_Disconnect.h:
* orbsvcs/tests/Event/Basic/Makefile:
* orbsvcs/tests/Event/Basic/mt_disconnect.conf:
* orbsvcs/tests/Event/Basic/run_test.pl:
New test that performs concurrent supplier and consumer
connect/disconnect calls.
* orbsvcs/orbsvcs/Makefile:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.i:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter.cpp:
* orbsvcs/orbsvcs/Event/EC_SupplierFiltering.h:
* orbsvcs/orbsvcs/Event/EC_SupplierFiltering.i:
* orbsvcs/orbsvcs/Event/EC_SupplierFiltering.cpp:
Renamed the EC_SupplierFiltering class to EC_Supplier_Filter, it
is more consitent that way (next step: rename EC_Filter to
EC_Consumer_Filter)
* orbsvcs/orbsvcs/Event/EC_Basic_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Basic_Factory.cpp:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.cpp:
* orbsvcs/orbsvcs/Event/EC_Null_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Null_Factory.cpp:
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Per_Supplier_Filter.cpp:
* orbsvcs/orbsvcs/Event/EC_Supplier_Filter_Builder.h:
* orbsvcs/orbsvcs/Event/EC_Trivial_Supplier_Filter.h:
* orbsvcs/orbsvcs/Event/EC_Trivial_Supplier_Filter.cpp:
Propagate the EC_SupplierFiltering name change.
* orbsvcs/tests/Event_Latency/Event_Latency.cpp:
The ORB initialized was not the default ORB used by the resource
factory for the event service.
Fri Jan 21 08:39:10 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.cpp:
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.cpp:
More #ifdefs to compile when CORBA messaging is disabled.
Thu Jan 20 09:07:47 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.cpp:
Minimum CORBA does not provide the _non_existent() method, we
disable the test in that case.
Thu Jan 20 07:11:16 2000 David L. Levine <levine@cs.wustl.edu>
* orbsvcs/orbsvcs/Makefile: updated dependencies, in
a workspace that had Notify (and all of the other
orbsvcs) built.
Wed Jan 19 17:51:39 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/ast/ast_check.cpp:
Added update of AST_Decl's filename pointer before
outputting a 'forward declaration not defined' error.
In certain cases, #xxxx statements in the IDL file
cause the preprocessor to add #<line number> "<filename>"
statements where <filename> is a repeat of the main
file name. This would invalidate
the filename pointer held deep down in a forward
declared interface node. If a forward declared interface
is found not to have a full definition, the error
function that is called tries to use this filename
pointer, causing a crash. Thanks to
Hessel Idzenga <idzenga@lucent.com> for sending in
the example file that uncovered this bug.
Wed Jan 19 15:44:23 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_union/union_ci.cpp:
* TAO_IDL/be/be_visitor_union/union_cs.cpp:
Uninlined union destructor. In certain cases
where the union in nested inside other structs
and/or unions, Dec Unix picks up on the fact that
the inner union's destructor is called before
it is defined inline. Although he didn't report
this bug, thanks to Hugh Arnold <harnold@itginc.com>
for sending in the example IDL file that rooted it
out.
Wed Jan 19 12:07:20 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_union_branch/cdr_op_ci.cpp:
Fixed demarshaling for unions having members with
multiple case labels. Thanks to
Christopher Kohlhoff <chris@kohlhoff.com> for reporting
the bug.
Tue Jan 18 20:45:36 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tests/Timed_Buffered_Oneways/run_test.pl:
TAO/tests/Timed_Buffered_Oneways is timing out on our HPUX_aCC
auto compile. Added extra timing information to narrow down
this problem.
Tue Jan 18 20:26:11 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/Transport_Mux_Strategy.cpp
(TAO_Exclusive_TMS::dispatch_reply): The following change to
this method was incorrect:
Thu Jan 06 22:26:42 2000 Irfan Pyarali <irfan@cs.wustl.edu>
This method is not the correct place mark the transport as being
available for reuse. If we do so here, another thread can start
using the transport and reset the message state before we have
had a chance of reading the reply. However, we still need to
find a place to safely mark the transport as idle otherwise we
will not end up reusing network connections. Note that this was
not a problem for Oneway, Twoway, and Locate Request invocations
since their destructors were setting the transport to idle
anyway. The problem was for DII and Asynch invocations since
their destructors were (correctly) not setting the transport to
idle.
Solution: First we revived
<Transport_Mux_Strategy::idle_after_reply> which was there but
commented out. <TAO_Exclusive_TMS::idle_after_reply> idles the
transport while <TAO_Muxed_TMS::idle_after_reply> is a no-op.
Note that this is the opposite of how <idle_after_send> works.
Second we make <~TAO_Asynch_Reply_Dispatcher> and
<~TAO_DII_Deferred_Reply_Dispatcher> call
<Transport_Mux_Strategy::idle_after_reply>. These reply
dispatchers are told about their transports when
<Invocation::invoke_i> has succeeded.
Third we change Oneway, Twoway, and Locate Request invocations
to call <idle_after_reply> instead of <idle> in their
destructors. This will prevent double idling of the transports
when used with the TAO_Muxed_TMS strategy.
* examples/Buffered_AMI/client.cpp (main): Checking
<message_count == -1> was not a effective trick since the
<(iterations % message_count) != 0> failed. Therefore,
introduced a new variable <setup_buffering> to simplify this
madness.
Tue Jan 18 18:45:13 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_union.cpp:
Increased the signed long default discriminator
value by 1. It was being set to ACE_INT32_MIN,
which is -2147483648. The literal constant is
generated in the _default() method, and when
MSVC sees the line
this->disc_ = -2147483648;
it outputs the warning "unary minus operator applied to
unsigned type, result still unsigned". If the line is
this->disc_ = -2147483647;
we get no warning.
Tue Jan 18 17:26:18 2000 Jeff Parsons <parsons@cs.wustl.edu>
* tao/deep_free.cpp:
Fixed a memory leak in recursive unions. The fix also
removes leaks in unions that contain anonymous
sequences. Thanks to
Lothar Werzinger <werzinger.lothar@krones.de> for
reporting this leak and sending in an example IDL
file.
* tao/DynStruct_i.cpp:
Fixed some mismatched forms of ACE_CHECK that I added
earlier today.
Tue Jan 18 13:04:24 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.cpp:
Fixed warnings on platforms without native C++ exception
support.
Tue Jan 18 12:08:00 2000 Jeff Parsons <parsons@cs.wustl.edu>
* tao/DynStruct_i.cpp:
Added patches to from_any() and to_any() sent in by
Philippe Merle <Philippe.Merle@lifl.fr>.These
patches (de)marshal the repository ID if the Any
contains an exception.
Tue Jan 18 11:37:31 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tests/MT_Client/client.cpp (parse_args): Only shutdown server
when specified by the user.
* tests/MT_Client/run_test.pl: Added explicit shutdown of the
server.
Tue Jan 18 08:56:20 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.h:
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.i:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.h:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.i:
Added a protected modified for the supplier (or consumer in the
ProxySupplier case), this is useful for users that inherit from
the ProxySupplier or ProxyConsumer and change the policies used
to communicate with the peer. Notice that this changes should
only be made once the connection is complete, i.e. after
connect_push_XXX() has completed. Thanks to Dave Meyer
<dmeyer@std.saic.com> for suggesting this.
Mon Jan 17 22:13:05 2000 Darrell Brunsch <brunsch@uci.edu>
* orbsvcs/ImplRepo_Service/tao_imr_i.cpp:
* orbsvcs/ImplRepo_Service/tao_imr_i.h:
Added a new command, ior to the tao_imr program. This
command creates a simplified IOR by attaching a given
name with the IMR's endpoint. This will only work with
servers that register the simplified IORs with the
IOR lookup table (Interoperable Naming Service stuff).
* orbsvcs/tests/ImplRepo/run_test.pl:
Change the nestea_ir test to use this scheme.
* docs/implrepo/index.html:
* docs/implrepo/tao_imr.html
* docs/implrepo/usersguide.html
Updated the documentation.
Mon Jan 17 22:53:53 2000 Carlos O'Ryan <coryan@cs.wustl.edu>
* orbsvcs/orbsvcs/orbsvcs.dsp:
* orbsvcs/orbsvcs/orbsvcs_static.dsp:
Updated the NT project files.
Tue Jan 18 20:44:02 2000 Carlos O'Ryan <coryan@uci.edu>
* tao/Object.cpp:
The _non_existent() method must catch the
CORBA::OBJECT_NOT_EXIST exception and return the appropriate
value.
Tue Jan 18 20:39:50 2000 Carlos O'Ryan <coryan@uci.edu>
* orbsvcs/orbsvcs/Makefile:
* orbsvcs/orbsvcs/Event/EC_SupplierControl.h:
* orbsvcs/orbsvcs/Event/EC_SupplierControl.i:
* orbsvcs/orbsvcs/Event/EC_SupplierControl.cpp:
* orbsvcs/orbsvcs/Event/EC_ConsumerControl.h:
* orbsvcs/orbsvcs/Event/EC_ConsumerControl.i:
* orbsvcs/orbsvcs/Event/EC_ConsumerControl.cpp:
* orbsvcs/orbsvcs/Event/EC_Reactive_ConsumerControl.h:
* orbsvcs/orbsvcs/Event/EC_Reactive_ConsumerControl.i:
* orbsvcs/orbsvcs/Event/EC_Reactive_ConsumerControl.cpp:
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.h:
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.i:
* orbsvcs/orbsvcs/Event/EC_Reactive_SupplierControl.cpp:
New components to the event channel that deal with stale
consumer and/or supplier proxies.
The first implementations simply poll the consumers and
suppliers periodically, using the _non_existent() method to
check if the object is there, to avoid dead-locks they use the
timeout policies in the ORB to bound the time the spend
polling.
* orbsvcs/orbsvcs/Event/EC_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Null_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Null_Factory.cpp:
* orbsvcs/orbsvcs/Event/EC_Basic_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Basic_Factory.cpp:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.h:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.i:
* orbsvcs/orbsvcs/Event/EC_Default_Factory.cpp:
The factory must create the new components now.
Use CORBA::ORB_init() instead of TAO_ORB_Core_instance() to get
to the ORB. This is standard and it can be easily configured
using the ORBid.
* docs/ec_options.html:
Documented the new options in the TAO_EC_Default_Factory.
* orbsvcs/orbsvcs/Event/EC_Defaults.h:
New defaults for the consumer and supplier control policies.
* orbsvcs/orbsvcs/Event/EC_Event_Channel.h:
* orbsvcs/orbsvcs/Event/EC_Event_Channel.cpp:
The event channel creates the consumer and supplier control
policies (using the factory), and then activates them.
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.h:
* orbsvcs/orbsvcs/Event/EC_ProxyConsumer.cpp:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.h:
* orbsvcs/orbsvcs/Event/EC_ProxySupplier.cpp:
Addded methods to invoke _non_existent() while holding the state
lock, to avoid race conditions (as in one thread disconnecting
a consumer while another thread tries to check if the consumer
still exists).
* orbsvcs/orbsvcs/Event/EC_SupplierAdmin.h:
* orbsvcs/orbsvcs/Event/EC_SupplierAdmin.i:
New accessor to use the collection lock.
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set.h:
Fixed nasty bug in the definition of the Disconnect command
object, the typedef was wrong.
* orbsvcs/orbsvcs/Event/EC_ProxyPushSupplier_Set_T.cpp:
Removed old debug message.
* orbsvcs/orbsvcs/Event/EC_Reactive_Timeout_Generator.cpp:
Removed references to TAO_ORB_Core_instance().
Mon Jan 17 19:54:34 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/GIOP.cpp (write_request_header_lite): Made
write_request_header_lite similar to write_request_header_std -
both for checking the <response_flags> correctly and for dealing
with the SYNC_WITH_SERVER option. Thanks to Bala for pointing
this out.
Mon Jan 17 15:49:25 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_field/field_ch.cpp:
Fix to scoping of field name generation in header
file when structs and/or unions are nested. Thanks to
Hugh Arnold <harnold@itginc.com> for reporting this
bug and sending in the example file.
* TAO_IDL/be/be_visitor_scope.cpp:
Cosmetic changes.
* tests/IDL_Test/union.idl:
Added example IDL file mentioned above to test suite.
* tests/IDL_Test/idl_test.dsp:
Updated project settings to make custom build settings
identical for all the IDL files in the test suite.
Tue Jan 18 10:54:49 2000 Carlos O'Ryan <coryan@uci.edu>
* docs/Options.html:
Documented the -ORBReactorRegistry option in the resource
factory.
Sun Jan 16 15:54:39 2000 Ossama Othman <ossama@uci.edu>
* tao/default_resource.cpp:
Moved template instantiations related to ACE_LOCAL_MEMORY_POOL
to ace/ACE.cpp. The idea is to place this set of template
instantiations in a "common" area, since ace/Configuration.cpp
also needs these templates instantations.
Fri Jan 14 08:00:00 2000 Michael Kircher <Michael.Kircher@mchp.siemens.de>
* tests/AMI/README,
docs/releasenotes/index.html,
examples/AMI/FL_Callback/README:
Realized that IDL_HAS_VALUETYPE is defined by default in the TAO
IDL compiler. Updated the documentation about AMI accordingly.
Fri Jan 14 06:01:00 2000 Michael Kircher <Michael.Kircher@mchp.siemens.de>
* TAO_IDL/be/be_interface.cpp,
TAO_IDL/be_include/be_interface.h:
Added a method named 'replacement' giving access to an
interface node, which can serve as an replacement for the original.
* TAO_IDL/be/be_interface_strategy.cpp,
TAO_IDL/be/be_visitor_ami_pre_proc.cpp,
TAO_IDL/be/be_visitor_interface/ami_interface_ch.cpp,
TAO_IDL/be_include/be_interface_strategy.h:
Enhanced the AMI preprocessor and the associated interface
strategies to handle replacement nodes.
This was necessary to pass information around for AMI code generation.
The bug triggering this was a multiply defined typedef. Thanks
to Bala for figuring this out.
Thu Jan 13 20:11:55 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/ORB_Core.cpp (run): Generalized to work with <perform_work>.
In <perform_work>, we only run the loop once.
* tao/ORB.cpp (perform_work): The method now calls
orb_core->run(), specifying that it is <perform_work> which
prevents looping of the event loop.
* tao/ORB.h (CORBA_ORB): There are now three versions of
perform_work() and are similar to the run() methods.
Tue Jan 11 21:27:51 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao/Messaging.pidl (Messaging): Added two TAO specific SYNC
options that control how buffering takes place in the ORB.
(a) SYNC_EAGER_BUFFERING is the same as SYNC_NONE. These
options will first queue the message in the buffering queue and
then check the buffering constraints. If the constraints have
been reached, the buffered messages will be delivered to the
server. If the delivery of the messages to the server times
out, the queued messages will remain queued until the buffering
constraints are reached again.
(b) SYNC_DELAYED_BUFFERING will first check the buffering queue.
If the queue already has buffered messages, then this option
will behave the same as the above option, i.e., buffer the
message for later delivery. However, if the queue is empty, it
will try to deliver the message immediately. If the message
delivery to the server times out, the message will be queued for
later delivery when the buffering constraints are reached.
* tao: Added code to handle the new options and strategies related
to the above change. The following files were effected:
- Invocation.cpp
- GIOP.cpp
- MessagingC.cpp
- MessagingC.h
- ORB_Core.cpp
- ORB_Core.h
- ORB_Core.i
- Stub.cpp
- Sync_Strategies.cpp
- Sync_Strategies.h
- TAOC.cpp
- TAOC.h
* tao/GIOP.cpp (write_request_header_std): Comparison with majic
numbers is evil ;-) Changed it to compare against the constants.
* tao/Pluggable.h (TAO_Transport): Added friendship between the
transport class and the sync classes.
* tests/Timed_Buffered_Oneways/client.cpp
(setup_buffering_constraints): Added the ability to either
choose eager or delayed buffering.
Tue Jan 11 17:25:16 2000 bala <bala@cs.wustl.edu>
* TAO version 1.0.12 released.
Tue Jan 11 16:46:40 2000 Balachandran Natarajan <bala@cs.wustl.edu>
* tao/PollableS.cpp:
* tao/PollableC.cpp:
* tao/MessagingS.h: Fixed compile errors in AMI code.
Tue Jan 11 11:20:27 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* performance-tests/RTCorba/Oneways/Reliable/server.cpp (set_rt_mode):
* performance-tests/RTCorba/Oneways/Reliable/client.cpp (set_rt_mode):
If errors occur while setting the thread priority, print a
message but don't abandon the test.
Tue Jan 11 00:27:51 2000 Nanbor Wang <nanbor@cs.wustl.edu>
* orbsvcs/Notify_Service/Notify_Service.dsp
* orbsvcs/Notify_Service/Notify_Service.dsw
* orbsvcs/orbsvcs/orbsvcs.dsp:
* orbsvcs/orbsvcs/orbsvcs_static.dsp:
Added Notify Service.
* orbsvcs/orbsvcs/Notify/NotifyPublish_i.h:
* orbsvcs/orbsvcs/Notify/NotifySubscribe_i.h:
* orbsvcs/orbsvcs/Notify/Notify_Constraint_Interpreter.h:
* orbsvcs/orbsvcs/Notify/Notify_Constraint_Visitors.h:
* orbsvcs/orbsvcs/Notify/Notify_ConsumerAdmin_i.h:
* orbsvcs/orbsvcs/Notify/Notify_Dispatcher.h:
* orbsvcs/orbsvcs/Notify/Notify_EventChannelFactory_i.h:
* orbsvcs/orbsvcs/Notify/Notify_EventChannel_i.h:
* orbsvcs/orbsvcs/Notify/Notify_FilterAdmin_i.h:
* orbsvcs/orbsvcs/Notify/Notify_FilterFactory_i.h:
* orbsvcs/orbsvcs/Notify/Notify_Filter_i.h:
* orbsvcs/orbsvcs/Notify/Notify_ProxyConsumer_i.h:
* orbsvcs/orbsvcs/Notify/Notify_ProxyPushConsumer_i.h:
* orbsvcs/orbsvcs/Notify/Notify_ProxyPushSupplier_i.h:
* orbsvcs/orbsvcs/Notify/Notify_ProxySupplier_i.h:
* orbsvcs/orbsvcs/Notify/Notify_QoSAdmin_i.h:
* orbsvcs/orbsvcs/Notify/Notify_StructuredProxyPushConsumer_i.h:
* orbsvcs/orbsvcs/Notify/Notify_StructuredProxyPushSupplier_i.h:
* orbsvcs/orbsvcs/Notify/Notify_StructuredPushConsumer.h:
* orbsvcs/orbsvcs/Notify/Notify_StructuredPushSupplier.h:
* orbsvcs/orbsvcs/Notify/Notify_SupplierAdmin_i.h: Added
TAO_ORBSVCS_Export. Thanks to Stephane Chatre
<schatre@oresis.com> for noticing these.
Mon Jan 10 18:12:55 2000 Balachandran Natarajan <bala@cs.wustl.edu>
* tao/PollableC.cpp: Fixed two instances of wrong usage of
ACE_NEW_RETURN.
Mon Jan 10 16:43:09 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_union_branch.cpp:
* TAO_IDL/util/utl_scope.cpp:
Lookup and code generation of case labels for unions
in IDL did not correctly handle cases where the
label names are scoped (for example, if the enum
discriminator is defined in a different module than
the union, the case label names may or may not be
scoped). Thanks to Hessel Idzenga <idzenga@lucent.com>
for reporting this bug.
Mon Jan 10 12:53:08 2000 Jeff Parsons <parsons@cs.wustl.edu>
* tao/GIOP.cpp:
* tao/GIOP_Server_Request.cpp:
Since Reliable Oneways depend on a feature of GIOP 1.2,
and GIOP 1.2 has not yet been fully implemented in TAO,
a temporary workaround was added so that a TAO server
will not confuse a twoway request from a GIOP 1.1
client with a oneway SYNC_WITH_SERVER request. When
GIOP 1.2 is fully implemented in TAO, the TAO server
can then check the version in the message header and
behave accordingly. [bug 397]
Mon Jan 10 11:51:33 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_exception/any_op_cs.cpp:
* TAO_IDL/be/be_visitor_structure/any_op_cs.cpp:
* TAO_IDL/be/be_visitor_sequence/any_op_cs.cpp:
Turned out to be the same change as below, but these
were warnings from SunCC.
Mon Jan 10 11:40:20 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_union/any_op_cs.cpp:
Fix to generated code for >>=. We were getting a
compile error from HPUX.
Sat Jan 8 12:13:43 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_union_branch/cdr_op_ci.cpp:
Fixed some uninitialized declarations. Reported by
g++ David Levine.
Sat Jan 8 03:30:00 2000 Michael Kircher <Michael.Kircher@mchp.siemens.de>
* TAO/docs/releasenotes/index.html
TAO/examples/AMI/FL_Callback/README,
TAO/examples/Buffered_AMI/README,
TAO/tests/AMI/README:
Made the documentation of 'defines' necessary to compile
AMI consistent across various documentations.
Currently, you need to set TAO_HAS_CORBA_MESSAGING and
TAO_HAS_AMI_CALLBACK to activate AMI code in the TAO library.
(TAO_HAS_VALUETYPE is needed as well, but is defined by default)
IDL_HAS_VALUETYPE needs to be defined for the TAO IDL compiler,
this is due to the fact that AMI uses valuetypes for exception
handling, though this might change in the future, because the
spec might change in this area. Use -GC on the IDL compiler
to generate AMI stubs.
Thanks to "Russell L. Carter" <rcarter@consys.com> for pointing
this out.
Fri Jan 7 15:22:38 2000 Ossama Othman <othman@cs.wustl.edu>
* tao/ORB.cpp (destroy):
Corrected behavior where ORB::destroy() would throw a
BAD_INV_ORDER exception if the ORB was already shutdown, which
does not agree with the CORBA specification. It is now possible
to call ORB::destroy() after ORB::shutdown(). Thanks to Mogens
Hansen <mogens_h@dk-online.dk> for reporting this problem.
Fri Jan 7 11:52:17 2000 Jeff Parsons <parsons@cs.wustl.edu>
* TAO_IDL/be/be_visitor_union_branch/cdr_op_ci.h:
* TAO_IDL/be/be_visitor_union_branch/cdr_op_ci.cpp:
Fixed bug where discriminator in a union with an
explicit default case was not getting set after
transmission. Thanks to Hugh Arnold <harnold@itginc.com>
for reporting this bug.
Thu Jan 06 22:26:42 2000 Irfan Pyarali <irfan@cs.wustl.edu>
* tao: Minimum CORBA in TAO meant a minimum POA. This
relationship was changed such that minimum CORBA does not have
to mean a minimum POA. This will allow the user to use a
minimum CORBA build of TAO with the full functionality of the
POA. The default, however, remains the same, i.e., minimum
CORBA enables a minimum POA.
Added a new variable TAS_HAS_MINIMUM_POA to orbconf.h.
TAS_HAS_MINIMUM_POA support is disabled by default if TAO is not
configured for minimum CORBA. If TAO is configured for minimum
CORBA, then TAS_HAS_MINIMUM_POA will be enabled by default.
Also, TAO_HAS_MINIMUM_POA_MAPS is now influenced by the value of
TAO_HAS_MINIMUM_POA rather than that of TAO_HAS_MINIMUM_CORBA if
not explicitly set by the user.
The following files were involved in this change:
- POA.cpp
- POA.h
- POA.i
- POA.pidl
- POAC.cpp
- POAC.h
- POAC.i
- POAManager.cpp
- POAManager.h
- POAManager.i
- POAS.cpp
- POAS.h
- POAS.i
- Object_Adapter.cpp
- Object_Adapter.h
- Object_Adapter.i
Thanks to Erik Johannes <ejohannes@oresis.com> for motivating
this change.
* tests/Timed_Buffered_Oneways: New test for oneways with
buffering and timing constraints. The client sends is setup to
send large requests to the server. The server is setup to take
a long time to process these requests. The combination will
cause flow control for the client. The timing constraints on
the client ORB will prevent the client from blocking because of
flow control. The request is queued up for later delivery once
the flow control subsides.
* tao/IIOP_Connect.cpp (handle_timeout):
* tao/UIOP_Connect.cpp (handle_timeout):
Added code to access the thread or ORB roundtrip timeout
policies. This allows us to perform a timed send operation
instead of a blocking one. We do not consider the object
roundtrip timeout policy since we do not have access to it.
* tao/ORB_Core.h (stubless_relative_roundtrip_timeout): Added a
new method to access the RoundtripTimeoutPolicy policy set on
the thread or on the ORB. In this method, we do not consider
the stub since we do not have access to it.
* tao/Sync_Strategies.cpp: Moved the TAO-specific SYNCH_FLUSH
option from the Messaging::SyncScope policy and changed it to
BUFFER_FLUSH as one of the buffering constraint modes in the
TAO-specific BufferingConstraint policy. Since we now have
BUFFER_FLUSH, we don't need BUFFER_NONE anymore since they both
do the same thing. Also, we didn't need the
TAO_Flush_Sync_Strategy anymore and the Buffered_Oneway example
was changed to accommodate this change.
* tao/Pluggable.cpp
(send_buffered_messages): The return value of 0 is considered
EOF. Therefore, changed the return values in the case of
timeouts and empty queues to be 1. Also, fixed how timeouts are
handled.
(reset_message): The resetting of the queued messages was not
correct. It was deleting excessively. This was fixed. Also,
the resetting was decoupled from the queue so that it can be
used with independent message blocks.
* tao/Wait_Strategy.cpp:
TAO_Exclusive_Wait_On_Leader_Follower::handle_input: The check
for <expecting_response_> was conflicting with the ability to
buffer asynchronous calls. If we mark the asynchronous call as
a twoway call, then buffering cannot take place. If we mark it
as a oneway call, then the check for <expecting_response_>
fails. For now I have selected to disable the check. The long
term fix is to separate out the two concerns (a) can the call be
buffered and (b) are we expecting a response.
* tao/Transport_Mux_Strategy.cpp
(TAO_Exclusive_TMS::dispatch_reply): Once we receive our reply,
we need to mark the transport as being available for reuse
again.
* tao: Renamed TAO_RelativeRoundtripTimeoutPolicy_i to
TAO_RelativeRoundtripTimeoutPolicy.
* tao/Messaging_Policy_i.cpp (set_time_value): Factored out common
code in Invocation.cpp and IIOP_Connect.cpp.
* tao/MessagingS.cpp: Added an overloaded relative_expiry(), which
is a hacky TAO extension to reduce a call to
CORBA::Environment::default_environment () since this method
will never raise exceptions.
Thu Jan 06 21:26:17 2000 Nanbor Wang <nanbor@cs.wustl.edu>
* tests/Param_Test/param_test_i.cpp
(test_unbounded_struct_sequence): Changed the explicit cast to
[(unsigned int) 0] to unsigned literal [0u] to avoid MSVC fro
whining.
Wed Jan 5 15:54:38 2000 Jeff Parsons <parsons@cs.wustl.edu>
* tao/Any.cpp:
Fix to new Any extractor to pointer to const Any,
prompted by a SunCC 4.2 compile error.
* TAO_IDL/be/be_visitor_union/any_op_cs.cpp:
Removed a couple of unnecessary 'ACE_reinterpret_cast's
in generated code for the new Any extractor to pointer
to const Any. They were causing compile errors with
Kai, SunCC and other compilers.
* performance-tests/Latency/deferred_synch_client.dsp:
Added missing paths for library linking.
* performance-tests/RTCorba/Oneways/Reliable/client.dsp:
* performance-tests/RTCorba/Oneways/Reliable/server.dsp:
Put in the complete relative path to tao_idl.exe, as
it is in all other projects with IDL files.
Tue Jan 4 11:46:42 2000 Jeff Parsons <parsons@cs.wustl.edu>
* tao/Any.h:
* tao/Any.cpp:
Added >>= operators to extract to const CORBA::Any *&,
as well as to const char *& and const wchar *&, as outlined in
the CORBA 2.3.1 C++ mapping.
* TAO_IDL/be/be_visitor_exception/any_op_ch.cpp:
* TAO_IDL/be/be_visitor_exception/any_op_cs.cpp:
* TAO_IDL/be/be_visitor_sequence/any_op_ch.cpp:
* TAO_IDL/be/be_visitor_sequence/any_op_cs.cpp:
* TAO_IDL/be/be_visitor_structure/any_op_ch.cpp:
* TAO_IDL/be/be_visitor_structure/any_op_cs.cpp:
* TAO_IDL/be/be_visitor_union/any_op_ch.cpp:
* TAO_IDL/be/be_visitor_union/any_op_cs.cpp:
* TAO_IDL/be/be_visitor_union/any_op_cs.h:
Added code generation for Any extraction operator to
pointer to const whenever extraction to a pointer is called
for. Left the former generated operators (same but without
the 'const') for backward compatibility, but marked them
as 'deprecated' in the generated header file. Also modified
code for both operators to return a null pointer if the
extraction fails. All this is in accordance with CORBA 2.3.1.
* TAO_IDL/be/be_visitor_interface/any_op_cs.cpp:
Modified Any extraction operator to return a null pointer
if the extraction fails, for the reason given above.
Tue Jan 04 07:30:04 2000 David L. Levine <levine@cs.wustl.edu>
* TAO version 1.0.11 released.
Tue Jan 4 00:34:57 2000 Nanbor Wang <nanbor@cs.wustl.edu>
* orbsvcs/tests/Trading/Service_Type_Exporter.cpp:
* orbsvcs/tests/Simple_Naming/client.cpp:
* orbsvcs/LifeCycle_Service/Factory_Trader.cpp: Fixes for KAI that
missed the last round and fixes that were needed after the last
round.
Mon Jan 3 20:33:11 2000 Balachandran Natarajan <bala@cs.wustl.edu>
* orbsvcs/Trading_Service/Trading_Service.cpp:
* orbsvcs/tests/Simple_Naming/client.cpp:
* orbsvcs/tests/Trading/Service_Type_Exporter.cpp:
* orbsvcs/tests/Trading/Offer_Exporter.cpp: Fixes for KAI that
missed the last round and fixes that were needed after the last
round.
Mon Jan 3 15:08:47 2000 Balachandran Natarajan <bala@cs.wustl.edu>
* orbsvcs/Trading_Service/Trading_Service.cpp:
* orbsvcs/tests/Trading/Offer_Exporter.cpp:
* orbsvcs/tests/Trading/Offer_Importer.cpp:
* orbsvcs/tests/Trading/Service_Type_Exporter.cpp:
* orbsvcs/tests/Simple_Naming/client.cpp:
* orbsvcs//LifeCycle_Service/Factory_Trader.cpp:
* orbsvcs/tests/Property/client.cpp: KAI needed fixes in these
files to compile clean. There was a need to explicitly use an
unsigned integer to avoid overloading amibiguity. Essentially
the same fix as 'Mon Jan 03 10:55:36 2000 David L. Levine'.
Mon Jan 3 14:45:22 2000 Nanbor Wang <nanbor@cs.wustl.edu>
* tao/ORB.h:
* tao/corbafwd.h: Removed compiler-specific alignment adjusting
preprocessor directives for MSVC and BCB completely. Thanks to
Christopher Kohlhoff <chris@kohlhoff.com> for reporting this.
Mon Jan 03 10:55:36 2000 David L. Levine <levine@cs.wustl.edu>
* orbsvcs/Time_Service/Clerk_i.cpp (get_first_IOR):
explicitly use unsigned 0 index for bindings_list
array element to avoid overloading ambiguity on Irix
and HP/UX.
Mon Jan 03 10:46:31 2000 David L. Levine <levine@cs.wustl.edu>
* orbsvcs/orbsvcs/Trader/Offer_Iterators.cpp (next_n):
modified look index so that its comparison is > 0, not
>= 0. Now that it's unsigned, it was causing a compiler
warning.
Sun Jan 02 10:48:59 2000 Michael Kircher <Michael.Kircher@mchp.siemens.de>
* orbsvcs/orbsvcs/Trader/Offer_Iterators.cpp:
Fixed a MSVC warning, complaining about an array
operator.
Sat Jan 01 09:18:59 2000 David L. Levine <levine@cs.wustl.edu>
* orbsvcs/orbsvcs/Makefile: updated dependencies, including
Notification Service files.
Sat Jan 01 09:17:40 2000 David L. Levine <levine@cs.wustl.edu>
* ChangeLog: moved to ChangeLog-99c.
|