Source

Coin / docs / BUGS.txt

Full commit
   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
There are additional buglists which may contain important bugs and
unsupported features for Coin. For the Coin satellite libraries,
extensions and example programs:

        simage/BUGS.txt
        SoQt/BUGS.txt
        So[Qt|Xt|Win]/src/Inventor/Qt/common/BUGS.txt
        SoWin/BUGS.txt
        SoXt/BUGS.txt
        SIMVoleon/BUGS.txt

..and buglists less likely to contain important bugs:

        SoGuiExamples/BUGS.txt
        SmallChange/BUGS.txt
        SoGtk/BUGS.txt
        Cute/BUGS.txt


Also, a number of bugs deemed to be obsolete due to age or obscurity
of environment they can be reproduced in, has been moved out of this
file. So also for bugs of little consequence, or for other reasons of
being of very low priority. They have been moved into catalogs under:

        Coin/docs/bugs/


Now, for the list of (possibly) significant bugs in the core Coin
library:

=====================================================================

005 Free memory read on connected SoSFEnum fields?

    This is an old problem which might have been fixed now, but we
    should check it with Valgrind: Purify used to report "Free memory
    read" when deleting SoSFEnum fields that were connected to other
    fields.

    <mortene@sim.no> 20020521.

=====================================================================

031 Setting up the alpha channel for the offscreen renderer doesn't
    consistently (between hosts) produce a transparent background when
    using setComponents(SoOffscreenRenderer::RGB_TRANSPARENCY).  It
    works at home on my Debian Linux w/ GeForce4MX440 with the 3123
    NVIDIA GLX sources, but not at work with whatever gfx system I
    have there.

    Index: src/misc/SoOffscreenGLXData.icc
    ===================================================================
    RCS file: /export/cvsroot/Coin/src/misc/SoOffscreenGLXData.icc,v
    retrieving revision 1.17
    diff -u -r1.17 SoOffscreenGLXData.icc
    --- src/misc/SoOffscreenGLXData.icc     19 Sep 2002 10:08:04 -0000      1.17
    +++ src/misc/SoOffscreenGLXData.icc     27 Sep 2002 08:50:02 -0000
    @@ -199,6 +199,8 @@
         attrs[pos++] = 4;
         // FIXME: won't get an alpha channel in the context unless we also
         // request a particular ALPHA bitsize. 20020605 mortene.
    +    attrs[pos++] = GLX_ALPHA_SIZE;
    +    attrs[pos++] = 4;
       }
       attrs[pos++] = None;
       return pos;

    20020927 larsa.

    UPDATE 20020927 mortene: this could simply be caused by the GLX /
    X11 drivers lacking support for any alpha-enabled visuals
    (according to spec, it is not guaranteed that any such visual is
    available). Anyway, the right way to solve this problem is to 1)
    write our own weight-function for X11 visuals (like it's been done
    for MSWin visuals in SoWin's SoWinGLWidget.cpp), for better
    control of the selection process. Then 2) warn the user if we
    can't get hold of any visual capable of all requested features.

    UPDATE 20021022 larsa: this was the case - old XFree3 with old
    Mesa does not support alpha for offscreen rendering.

    UPDATE 20021121 larsa: The offscreen renderer GL context creation
    is nevertheless bogus (with GLX at least), and should be fixed so
    the best possible context is created.

    UPDATE 20060203 mortene: a *lot* has been improved about offscreen
    rendering contexts since the last update on this bug, but the
    selection process among the available contexts has still not been
    fixed, nor exposed to the user in the form of any public API.

=====================================================================

045 SoCenterballManip rotation off-center.

    When dragging the 'stripes' on a CenterballManip node, rotation
    happen around an off-center axis. (Off CenterballManip center)
    This happen for rotations around all the axis.

    NOTE: The bug is not visible if there is no connected geometry, or
    if the geometry is 'simple', like one cube or one cone.

    A simple scene-graph to reproduce:

    ----8<---- [snip] ---------8<---- [snip] ---------8<---- [snip] --
    #Inventor V2.1 ascii

    Separator{
        CenterballManip {}
       Array{
              numElements1 2
              numElements2 2
              numElements3 2
                separation1  2.0 0.0 0.0
                separation2  0.0 2.0 0.0
                separation3  0.0 0.0 2.0
                      Cone {}
       }
    }

    ----8<---- [snip] ---------8<---- [snip] ---------8<---- [snip] --

    20021209 rolvs

    UPDATE 20030408 mortene: there is also a second bug here; the
    scene bounding box is not calculated correctly, as can be seen by
    loading this into e.g. the ExaminerViewer. The near clip plane
    will clip part of the manip's geometry, something which should
    never happen.

=====================================================================

047 SoAntiSquish seems to trigger a bug in our matrix factoring code.

    From a coin-discuss posting by Philippe Debann�:

    ----8<---- [snip] ---------8<---- [snip] ---------8<---- [snip] --

    I am planning to design my own Dragger class and have been looking
    at the Toolmaker Ch. 8 examples: TranslateRadialDragger and
    RotTransDragger.  I found an unexpected behaviour of
    RotTransDragger (top-left of scene): When dragging one of the
    SoRotateCylindricalDraggers, the dragger will switch from rotation
    to scaling up as the mouse position approaches and goes beyond the
    bounding cylinder of the (spherical) dragger.  In fact, the
    dragger should not do any scaling up/down!! This happens
    systematically when dragging the horizontal band left or right
    (Y-axis rotation).

    When I comment out the SoAntiSquish part of RotTransDragger, this
    unexpected behaviour dissapears (But obviously you want the
    AntiSquish's anti-squishing behaviour!).  Alternately, if I set
    its 'recalcAlways' field to FALSE, I get a different strange
    behaviour: manipulation of the RotTransDragger affects the
    rendering of the RotTransManip to the right of it: its
    antisquishing scaling factor and fluctuations of its wire-mesh
    sphere.  All this seems to be independant of the AntiSquish
    'sizing' field, e.g.  BIGGEST_DIMENSION or LONGEST_DIAGONAL.

    ----8<---- [snip] ---------8<---- [snip] ---------8<---- [snip] --

    pederb has investigated, and found that this seems likely to be a
    problem with our SbMatrix code that factors the translation,
    scaling and rotation out of a matrix.

    20021218 mortene.

=====================================================================

051 ExaminerViewer's automatic initial "view all" doesn't work on
    a scene graph with SoImage.

    Reproduce with the following source code snippet. The
    "aass_logo.rgb" is an image of dimensions 80x68 (I haven't
    investigated if this matters or not).

    ----8<---- [snip] ---------8<---- [snip] ---------8<---- [snip] --
    #include <Inventor/Qt/SoQt.h>
    #include <Inventor/Qt/viewers/SoQtExaminerViewer.h>
    #include <Inventor/nodes/SoSeparator.h>
    #include <Inventor/nodes/SoImage.h>
    #include <Inventor/nodes/SoCone.h>

    int
    main(int argc, char* argv[])
    {
      QWidget * mainwin = SoQt::init(argv[0]);

      SoSeparator * root = new SoSeparator;
      root->ref();

      root->addChild(new SoCone);

      SoImage * img = new SoImage;
      root->addChild(img);
      img->filename = "aass_logo.rgb";

      SoQtExaminerViewer * viewer = new SoQtExaminerViewer(mainwin);
      viewer->setSceneGraph(root);

      SoQt::show(mainwin);
      SoQt::mainLoop();

      delete viewer;
      root->unref();

      return 0;
    }
    ----8<---- [snip] ---------8<---- [snip] ---------8<---- [snip] --

    Notice how the cone is a tiny dot in the lower left corner of the
    image at first. Then click view all (the eye icon) and see how the
    camera is then positioned where it should be. Correct behavior
    would be for the camera to be positioned there from the start.

    20030103 mortene.

    WONTFIX? 20030225 rolvs: In this example, setSceneGraph calls
    viewAll() before the window is 'show'-ed. The viewport region
    passed to (*SoCamera)->viewAll( graph, vpreg ) is therefore wrong,
    causing a malformed boundingbox for the image and thus a malformed
    view.

    UPDATE 20030225 mortene: is there any simple and clean way to
    detect and handle this with a) a warning message for the app
    programmer, b) using a decent "dummy" / default region?

=====================================================================

054 SoComputeBoundingBox engine probably fails with viewport-dependent
    geometry.

    The SoGuiExamples/engines/computexfbbox example fails with a small
    scene with an SoText2 in it:

    ----8<---- [snip] ---------8<---- [snip] ---------8<---- [snip] --
    #Inventor V2.1 ascii

    Cube { }
    Translation { translation 10 0 0 }
    Text2 { string [ "hepp", "tjo-bing", "hupp" ] }
    ----8<---- [snip] ---------8<---- [snip] ---------8<---- [snip] --

    ..and pederb tracked this down to a failure of the engine's
    evaluate() to be called again when the camera position changes.

    Note that this bug is probably rather hard to fix properly.

    20030128 mortene.

=====================================================================

060 SoToVRML2Action failing with VRML97 nodes.

    Reported by kintel:

      "SoToVRML2Action fails [with this file]:

      ----8<---- [snip] ---------8<---- [snip] ---------8<----
      #VRML V2.0 utf8

      Transform {
        children Shape {
          geometry IndexedFaceSet {
            coord Coordinate { point [ 0 0 0, 1 0 0, 1 1 0 ] }
            coordIndex [ 0, 1, 2, -1 ]
          }
        }
      }
      ----8<---- [snip] ---------8<---- [snip] ---------8<----

      ...and outputs an empty file:

      ----8<---- [snip] ---------8<---- [snip] ---------8<----
      #VRML V2.0 utf8

      Group {
        children
              Group {
              }
      }
      ----8<---- [snip] ---------8<---- [snip] ---------8<----

      This is a special case as it is already a VRML2 file. However, I
      guess the action will also fail given a mixed scene graph as
      input. This is a bigger issue since we often have situations
      with mixed scene graphs and want to convert them to VRML2."

    We've also had at least one external report from a user who found
    this to be a real problem. From the problem report:

      After some tests I've found that the problem seems to be that
      VRML shapes instead of being copied through by the ToVRML2
      action (they are obviously already VRML2), they are deleted
      (ignored?). It looks like that src/actions/SoToVRML2Action.cpp
      lacks one more callback to handle nodes already in VRML2 format.

    This is a medium priority item.

    20030307 mortene.

=====================================================================

066 Viewer "view all" functionality not 100% correct.

    See the VRML1 file below containing 5 line segments as separate
    linesets.  Using this file, the viewAll button in the examiner
    viewer only _almost_ zooms out the camera enough: parts of the
    model is outside the view.

    To reproduce:

    o Open model in an examiner viewer.
    o Resize the window so height ~= 2* width. (at least.)
    o Rotate the model so you get minimal total height in the view.
    o Press viewall. If my problems doesn't show, rotate around the up
      axis and press viewall repeatedly until error occurs.

    ----8<--- [snip] ---------8<--- [snip] ---------8<--- [snip] ---
    #VRML V1.0 ascii

    Separator {
      Separator {
        Coordinate3 {
          point [ 372031 7.20114e+06    32.4351, 372031 7.20114e+06   -465.843 ]
        }
        IndexedLineSet {
          coordIndex [ 0, 1, -1 ]
        }
      }
      Separator {
        Coordinate3 {
          point [ 386603 7.19641e+06    32.1387, 386603 7.19641e+06   -348.644 ]
        }
        IndexedLineSet {
          coordIndex [ 0, 1, -1 ]
        }
      }
      Separator {
        Coordinate3 {
          point [ 384767 7.19186e+06    25.2236, 384767 7.19186e+06   -285.863 ]
        }
        IndexedLineSet {
          coordIndex [ 0, 1, -1 ]
        }
      }
      Separator {
        Coordinate3 {
          point [ 377824 7.20872e+06    32.4382, 377824 7.20872e+06   -329.9802 ]
        }
        IndexedLineSet {
          coordIndex [ 0, 1, -1 ]
        }
      }
      Separator {
        Coordinate3 {
          point [ 383236 7.18857e+06    40.1396, 383236 7.18857e+06   -371.016 ]
        }
        IndexedLineSet {
          coordIndex [ 0, 1, -1 ]
        }
      }
    }
    ----8<--- [snip] ---------8<--- [snip] ---------8<--- [snip] ---

    I tried the model in the SoGuiExamples/engines/computexfbbox
    example, and the bounding box gets calculated correctly. I guess
    that means there's a bug in either SoCamera::viewAll() or in
    SoGuiViewer::viewAll() -- most likely the latter.

    20030507 mortene, reported by kintel.

=====================================================================

072 Assert failure when writing extension nodekit in binary format.

    Running the example program below causes the following assert to
    hit:

      test: /home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:248: void set_current_writeref(const SoBase *, int): Assertion `rc >= 0 && "buggy writerefcounter"' failed.
      Aborted

    This points at some very nasty error, where some part of the scene
    graph is not tagged for writing on the first SoWriteAction pass,
    but is still attempted written on the second pass.

    It's really lame that we have this kind of bug at all, as having
    the two passes traverse the same structures shouldn't be rocket
    science.

    When writing as ASCII, the assert does *not* hit -- which means
    traversal is different somewhere for binary and ascii export.

    --------------8<----------- [snip] ------------------8<--------
    #include <Inventor/actions/SoWriteAction.h>
    #include <Inventor/nodekits/SoSubKit.h>
    #include <Inventor/nodekits/SoNodeKit.h>
    #include <Inventor/nodekits/SoBaseKit.h>
    #include <Inventor/nodes/SoGroup.h>

    class SoTestKit : public SoBaseKit
    {
      SO_KIT_HEADER(SoTestKit);
      SO_KIT_CATALOG_ENTRY_HEADER(A);

     public:
      static void initClass();
      SoTestKit();
    };



    SO_KIT_SOURCE(SoTestKit);

    void
    SoTestKit::initClass()
    {
       SO_KIT_INIT_CLASS(SoTestKit, SoBaseKit, "BaseKit");
    };

    SoTestKit::SoTestKit()
    {
       SO_KIT_CONSTRUCTOR(SoTestKit);

       SO_KIT_ADD_CATALOG_ENTRY(A, SoGroup, FALSE, this, \x0, TRUE );

       SO_KIT_INIT_INSTANCE();
    };

    int
    main(int argc, char ** argv)
    {
      SoDB::init();
      SoNodeKit::init();
      SoTestKit::initClass();

      SoTestKit * testkit = new SoTestKit();
      testkit->ref();

      char * buf = (char*)malloc(1024);
      SoOutput out;
      out.setBuffer(buf, 1024, realloc);
      out.setBinary(TRUE);

      SoWriteAction wa(&out);
      wa.apply(testkit);

      testkit->unref();

      return 0;
    }
    --------------8<----------- [snip] ------------------8<--------

    Problem reported by Tamer Fahmy.

    20030602 mortene.

    UPDATE 20050610 mortene: it seems someone (pederb?) rewrote the
    assert, and made the code "kinder", as there are now warnings
    instead:

        Coin warning in SoWriterefCounter::setWriteref(): writeref < 0 for <noname> <0x806d6a8>
        Coin warning in SoWriterefCounter::<cleanup>(): Not removed from writedict: 0x806d6a8, Group:<noname>

=====================================================================

073 Multiple complexity settings for same SoTexture2 does not work.

    Using the same SoTexture2 nodes with different
    SoComplexity::textureQuality settings is something which works for
    SGI Inventor, so it should be considered a Coin bug.

    Reproducible problem, example file:

    --------------8<----------- [snip] ------------------8<--------
    #Inventor V2.1 ascii

    Separator {
       Separator {
          Complexity { textureQuality 0.1 }

          DEF mytex Texture2 {
             image 2 2 3
             0xffffff 0x00ff00   # white   green
             0xff0000 0xffff00   #  red    yellow
          }

          Cube { }
       }

       Separator {
          Complexity { textureQuality 1.0 }
          USE mytex
          Translation { translation 5 0 0 }
          Cube { }
       }
    }
    --------------8<----------- [snip] ------------------8<--------

    pederb's comments on the bug:

        We handled this earlier, but when adding proper support for
        auto-caching, it was easier to store everything for a texture
        object inside the texture object, and then identify it with a
        single uint32. Then it's quite easy to figure out the current
        texture object after a display list has been executed (the
        render-cache must store the OpenGL state changes it does so
        that SoGLLazyElement is updated after the display list has
        been executed). I could add additional variables for storing
        the filter settings, but I just couldn't see any practical
        reason for doing this.


    20030916 mortene.

    UPDATE 20040714 mortene: an external Coin user (Nigel Murry) had
    an ingenious way to use a texture multiple times with varying
    complexity settings, which now fails -- so there are indeed
    practical uses for this. So we should really consider this a bug,
    and a fairly high priority item aswell, as this is a regression
    versus older Coin versions.

    From Murry's bug report on coin-discuss:

        I have been using textures to show how some parameter (like
        say�temperature) varies over the surface of an
        object. Sometimes I would like to see distinct color bands and
        other times I would like to see a continuous variation of
        color. So I have a SoTexture2 defined using 8 or 16 color
        levels. Then I change bewteen a "banded" and a continuous
        display using a SoComplexity node with textureQuality set to
        0.00001 for a banded display.

        That normally works fine, but when the texture is used more
        than once there appears to be some confusion as to how the
        textureQuality is used. In the attached little iv file, if the
        switch at line 7 is off (whichChild=-1) then the second
        (legend) use is correct. If the switch at line 7 is on
        (whichChild=-3) then the second use shows a continuous
        texture.

=====================================================================

074 SoComplexity::textureQuality changes must be made before
    SoTexture2 traversal.

    In the "Inventor spirit" of appearance settings not having any
    effect until hitting a shape node, it should be possible to do
    this:

    --------------8<----------- [snip] ------------------8<--------
    #Inventor V2.1 ascii

    Separator {
      Texture2 {
         image 2 2 3
         0xffffff 0x00ff00   # white   green
         0xff0000 0xffff00   #  red    yellow
      }

      Complexity { textureQuality 0 }

      Cube { }
    }
    --------------8<----------- [snip] ------------------8<--------

    ..but the SoComplexity node will have no effect in this case. I
    feel it should.

    Note that SGI Inventor doesn't handle this either, but I still
    think it is a weakness in Coin which we should fix.

    20030916 mortene.

=====================================================================

075 SoHandleBoxManip does not behave correctly when locking a
    translation axis.

    To reproduce, load this simple iv-file into an examinerviewer:

    --------------8<----------- [snip] ------------------8<--------
    #Inventor V2.1 ascii

    HandleBoxManip { }
    --------------8<----------- [snip] ------------------8<--------

    ..then hold down SHIFT while LMB-clicking on one of the sides (for
    translation). The translation will correctly be locked to one
    axis, but when releasing the SHIFT button, the manip will jump to
    the current cursor location. This last behavior is incorrect.

    20030922 mortene.

=====================================================================

076 SoHandleBoxManip does not behave well upon "deep Z translation".

    To reproduce, load this simple iv-file into an examinerviewer:

    --------------8<----------- [snip] ------------------8<--------
    #Inventor V2.1 ascii

    HandleBoxManip { }
    --------------8<----------- [snip] ------------------8<--------

    ..then move the box out to the left side of the view
    (i.e. translate in -X), before LMB-clicking the right side of the
    box to translate in the Z direction. The manip will correctly move
    in the Z direction when moving the mouse cursor to the right, but
    upon continuing moving the mouse cursor to the right it suddenly
    misbehaves by first jumping back close to the camera, and then
    starting to move in the *negative* Z direction.

    This is clearly bad handling of some kind of border case.

    20030922 mortene.

    UPDATE 20060208 mortene: this may be a bit hard to reproduce. To
    make sure to trigger it, one can first move it way back as
    described above, then do "view all", and then go ahead with the
    above description again.

=====================================================================

090 Scene graph copy behavior deviates from SGI Inventor.

    For the code example below, we get different behavior on
    connection copying between Coin and SGI Inventor. For Coin, the
    node with the slave field is not copied (the connection is just
    shortcut), while it is copied for SGI Inventor.

-----8<------ [snip] ------------8<------ [snip] -------
#include <assert.h>

#include <Inventor/SoDB.h>
#include <Inventor/nodes/SoRotor.h>
#include <Inventor/SoLists.h>

int
main(void)
{
  SoDB::init();

  SoRotor * masterrotor = new SoRotor;
  SoRotor * slaverotor = new SoRotor;

  SoRotor * rotor = new SoRotor;

  rotor->rotation.connectFrom(&masterrotor->rotation);
  slaverotor->speed.connectFrom(&rotor->speed);

  printf("rotor==%p, masterrotor==%p, slaverotor==%p\n",
         rotor, masterrotor, slaverotor);


  SoRotor * copiedrotor = (SoRotor *)rotor->copy(TRUE);

  SoField * f;
  SbBool ok = copiedrotor->rotation.getConnectedField(f);
  assert(ok);
  SoRotor * copiedmaster = (SoRotor *)f->getContainer();

  SoFieldList slavelist;
  int nr = copiedrotor->speed.getForwardConnections(slavelist);
  assert(nr==0 || nr==1);
  SoRotor * copiedslave = nr ? (SoRotor *)slavelist[0]->getContainer() : NULL;

  printf("copiedrotor==%p, copiedmaster==%p, copiedslave==%p\n",
         copiedrotor, copiedmaster, copiedslave);

  return 0;
}
-----8<------ [snip] ------------8<------ [snip] -------

    A test run:

    Coin:
        rotor==0x807d3c0, masterrotor==0x807c740, slaverotor==0x807ce18
        copiedrotor==0x807e248, copiedmaster==0x807c740, copiedslave==(nil)

    SGI Inventor:
        rotor==0x8058710, masterrotor==0x8051690, slaverotor==0x8057ae0
        copiedrotor==0x8059cb8, copiedmaster==0x8051690, copiedslave==0x8059df0

    (Notice the differences for 'copiedslave'.)

    This is a quite nasty bug, which can cause hard to debug errors
    when moving code from SGI Inventor to Coin.

    20031121 mortene.

=====================================================================

092 SoToVRMLAction outputs empty scene graph when given a Coin/Inventor
    scene graph.

    SoToVRMLAction only supports VRML2 scene graphs, not Coin/Inventor
    scene graphs.

    20031126 kintel.

=====================================================================

095 SoIndexedTriangleStripSet renders much slower with PER_FACE normals.

    According to pederb, this is not expected, so something fishy is
    going on.

    This is seen on a Linux NVidia box with v41.91 driver and
    GeForce2.

    Here's a model which reproduces the problem. Just load it into the
    examinerviewer:

-----8<------ [snip] ------------8<------ [snip] -------
begin 644 makrell.iv.gz
M'XL(",G[S3\"`VUA:W)E;&PN:78`Q%U=CYVW<;[7KSA`;EK`..!P2`ZG=ZX@
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MY.<W;^6O;@[RAXW(_<?3Z=WA/Z\_O7_$^?;5[[[_ZLN7KPXWINA>N=\+A``?
MWIU^N/K\X=/A'^Y.]_)?11"XJ\/[SW\^'3Z>[G[07Q$ECV`?;N_O__%X^/+M
MV]L[X_OI5N1$^9^^$$!A>?_^]O.'=Z+KT^%/)]/Q]NK^='P`>`1Z<SK=&,_#
MZS]<O[N^.OSV^N;S7P[O[JZ%WJ'`D>&+PZUR^9?35[=W;T]YCY%30@#@PX^W
MTALWCUI^NOKP^?38[D=#__]N=<&7M]:*JT\G?/R9_//Q5OKZ\!^'7N&8,B8Z
M8#NFCE#ID(^Y%LB=O_@_>1-$ZKV:8)8_5A&LK1/`S^72D7OOZ'+04E,Y;*7G
M_A0PK+E`@>Z"A1NH(%4&S+\X12SRVP>LQTZML<B1_(.I[@`K<76YBLD$L1?J
M95YSJ"G+&0J@$&LJV)B@JB`GKKWT'2)B*@^"U$6PMY8;E7G5H;:LIRB(J;KF
MBKUU0Z0D/;4?.="Z"Y8$J$.1<^7.;5YUK#'K.<(Q]V;](DR3_L8Q)4J)G^)U
M[J6X''70,2N3D'KJTXJ#35G-4/`:9U;!5*0W2`&!A,P>L<,F6#/W;`QE=7JZ
MY(RHCC1E/4/0196%83D*/P!40:ZRT.[&C0S039!3)FN*0,OJ,Z\ZUICU'`6Q
M5!6K.:/W"HJU\VYMZKJJFB`6+#9-F1JF/*\XUI35#!4/P!"QRSK23%!&6JM[
MQ-R-(4H34/?=7I'P`L6!AJSF)WAH,ZK(*,/<BHJ5FNE<0U+.)DBRT>IP[9"P
MEFG%L8:L9ZB(K7<5!)1]0P>7^`:-ZKXIXE-8YT%ALN$J>V_I9Y:1N.I08]9S
M5,2*HAK5D\"J'&6OS-SVC2E-%V,1I);0!!LGXGZ!ZE!CUG-\-#C*<IME<U3!
M+`OJF6$K"PX;HJRNK(*520;P):I#C5G/42=7!S)$"07,;ZH@NT>A_<"5_YF@
M[+>LD[5TH5OF-<?:LIRB;I.0T0$S%)U:*'(R"W>`9"N/`F)JJEF6IR0!Q+3F
M6%/64P0-X3*8(`,GG5FY]\9\QM$I9!P9<FOHT1\E+O.J8XU9SU$]7U#$HAX1
M%?5\`:DD.N/$)Y=KK623$V\CI3*M.=B6Y10?(QQ90SNZMR\>5-XW17X9##!+
MS(O5Y,2;ZM-Z8PU935#PL&?#DVBWR.X*1RX->;\-Z8_=D\B4A*$*%@FBYC6'
M6K*>H2#6[)J+S#T4.0GG$/+.C^:6W%TLA;+LP2J8.]&\XEA35C/4#(W$MBK7
M"A"I''66]72?"A`]YE:VVG0.BB"#+-#SFD,M6<]0$)E-L9"2T:=RB8KXR+O4
M7FJ535!VW6*`LC:?V9WCFF-M64X1CK)'5-(@5EQ`^;\("E,9G&?RF38215!"
MX6YR)&%OF=8<;<IRA@59O=YZ+`DELE?!D@KNW6/U%2RK(2$1;(*%:J=IS=&F
MK*=82F.C6#5R5SEJLMCNNT]\8TMJU"Y_<$'Y^7[(QC5'V[*<8BEL>3BB4HHJ
M)I)(ELX`ML0N*-Y%4T&)B]*9?HYKCC5E.<,B`]`2=BRN+B@BYRX[S'[@U$1]
M$R3;@("KL+E`<ZPIRQD6F4%LM836P%(T4*&?\7<3%O+J1"N,Q3PFW4SKM.9H
M4]93++F`"]*6OI)]0X+\??]AKJY:0ESS$BLC(UZ@.=B6Y11+3IM<AV*5%HEO
MZIE\:\H,7GJ3D+AJ,"#])S^L\YJ#35E/4;W"K8Q8*UI@4\49VL?M*=>&6[VQ
MNVH)#V7'F5<=;<QZCN&::+C*.J`ZV)@8Q](E@).`5%:)7#-:YD7&6NWGQ$K?
MY(KU2^:>VA.&+MC`!<7;<,!&]*3-+@@/B,S*4`));F<%@PSCBK,+%JB6@A`F
M[<FF,<APO>HP(F^"8A/K9W'-Q17X&ZH;9],LP$]"M4&*RS5'`2W9I(*ZH9D@
M[O+F/]?<N553C96PSU,<T.Q#I\F4-D`M/5Y`<+7B,!Y@V?!2;R8H\3'_C<G2
M,'E+2N5+"*[6&P84E]D%9<VS19NAR%;\5]>E1F!R6EIGRA<P7*TX#`A$+BA#
MP.0D.F[MW";@@`38FFM&[)<P7*TX"IB3KS4RWXNM-9V3^!D7:`XCYNQ3C]#S
MB;+40*?ZU_<`JH5\^ZFR[U_$<;GJ,&+V;S4R4?,%IQ,__0C#51.[H,S/LDT7
M>))4'^2X7'48<;..++*P62>W<TUA'[42DY)W(#;*[1*&BQ4'\1[$L*9M#F#%
M,W#B!;M@:9MG!1*>X@*"?U=S\Z[KI(DA\R:S!),74(QJSFEK2M<,LVH6'Q[P
M`HK+-8<!M_'?6RX>#(`PQ$LTAQ&!-^,P;8X+=[I(=1CQ0;!#VOQ3<13K"M5_
M%Q'P09`=L5"KEZD>1RSN-U4M@/T\XA1!*F3EA"I+DF:HCN)+-W&R]H(ULPO6
M3-D$BVP?=$8PU0=$_9,(-BX)SJC>Y"KJ,#I6$M?G:1EJ2'.X+06["[;63%"\
M]]TWQ6.J@VV):^Y;6XC($!M(N^`2B@^JN>GR+*K%J6[Y`M7AQH15-R]`5TCR
M]RI8>?>=V"#%[(#`:(`BCPSSFL-->1`$Y.::N["\2'.P*7'-==-LJ\512VJ4
MZ7+;0$/-_(IBIO3T(Z@1Q>&6Q#7WS8A%_MI4-_DA74!Q^Y931J+A==D!S\V^
ML.)P4Z*:M4)F@&3Y2M%<),2[Q#:E.<7\L.3(;H1/G*PQU=&VQ#4W_Q0W0^FV
M0%3Q?!$NH1A5W5/9!*GZT)&N[/T"CH5]B<^:%-XT,U^@.=R6L&;_**&*HP_9
M-5?F2XSSL*UI@@M=->*Y12>L.MJ6N.8"FV#V42NM*CE?0+$VEY-YYQ03I-XN
MT!QNR[#F9M]DZ'3I.>$%%*T8K((,R5TGPG,+;5ASM"F-7+!`P>H^6T_GO)>P
MYFA3].,ZTYRS`[:F7TM?HCG:EKAJ_\I=VF*_(JJ[N->76">LFOK6EEZM`QO6
M6NL%%..:=<<UP>2[?BOP-(8?HAC7W&`3A-I<,[>4+Z$85EWKHZ`'$)7.>O$A
MC@.55D[Z-9Y^[,#)X]X&/7''V;.=8<`X16S;\=.:T(Y8H@8?E6<I=L'SLYWB
MYX$"=CTG@7WV;&<<<+41PPQ+K@;86O%"G3CS5'*?/MD91UQNQ##%7+ME8"0X
M1@M193GFMC\'$S[:&4=<;\8PQT0/E04@5H=#O_UI^Z_9@H<[PWCKK1ADJ!4$
MRX$FT+A!Y4C"I]RFCW;&$1?;,,Y05D#;_V2SK4VW#/V\+>T_+0X?[8PCKK=B
MF*-^?FL?D7<N5J;DKH?6Z^3QSCC>>BM&&69/@J#$IVP,"VBP.GW`<P!QL0WC
M#!.8WY3U@R63DQ6%I@]W1O'66S#,L';/0R3FEKOO?0DJ3A_M'$!<;,4X0\LX
MX9$US:2*97'`GFGZ8.<`XGHK1CDV2'[>KY!_L*@A`9TY^A8]V3F`N-Z.48[B
M"/B1S:1G*-3O+>H*SY_MC".NMV.88R]^8E-]`CO/FC4ATJ>/=H8!UULQ3)$?
M#O%*9)KMNX.<,I\9.]&CG7'$Y6:,4NQ9MP%=)B"1C3'9-2KW^=.=8<3U9HQS
M+'6[CH#0(I>DR=6]=Q<]WQG&6V[$,$/Q='P+)+%RT9,AE"0RGS[>&09<;<(P
M05DAW-5A`=0C3RPQ6]_/J>CQSCC@<AL&&4K0XD9L28]JJB#(]@\X?;HSCKC<
MB&&*>:ML=]GI6563;!S[33)ZNC.,M]Z&489^5EN"4ZZ=[2Q?UH]7I@]W1@%7
MFS!.L!5+26F1P,YK-BA\YOA;]&!G%&^U`>,$R7+D>K4;0-$E1#9&V-]R%CW6
M&<9;;<$P/T3+)1+:^482]QX`9H]TAN$66R]*KVF^5'/D4,&/:589./N\0_@\
MYP#B8@M&"5H2JAT3BD.C%1/HK$>ZI@]SA@'7FS!.D5)VCOIUJ1U7U%3Y_,'4
M`<359HQ3Q.(42ZN\G7^LN-\6PX<Y!Q#7FS'*$;GVK=#7[02=GC3*^V@T?)@S
MCKC>CG&.P?+F^H+I>CL&$PK"B)-?ZR9K(AC'0C+;\ERL/H`7)N@^M@'*"-=T
M4$IG+I*,$A0/H'ERHOIL04)5/1FG#P"NMF`L3S9@P>5=$F;8"-PV2)Y)4!\%
M&"<SHB.(S\D1*VZWKR5;X8ESZU#GAR(F<D3Q:/U8%W"G?7`6S,@,(*Z?+K&4
M_X@9UW=,G&,I5IN35=M3HGKU7MK?SA7OF&I)5BV=)%\EQ(4'W!<18JFH$<#U
M8R=2T1HPXOIN"3.L_C%<!G1'56)ID9NM7`XA/AM#\LP19J[V*`&SGNBLTP.Q
M4;>!F.7O&UF.59:4.IEP'`%</56")?D!(Z[OECA'@N))4;&GEO+$]99-_\QG
M1=%^Z1F<HO@Y:#&JS%9IWV2:=01Q]=`)?FXR8,7U_1+GV)/+93_(*\-;XVV:
M[I>:[(O[HJFI3&KNWF3XPFS.?P!Q_=@)?DT5-^/R?HDSE"7!I@'W;)^NZVVX
MI>T=G7B_R'0R0?&EH5BRO+=R)L<<2JH/X"T?.&&&=;M7*E49,"J&6;94F,NJ
M#^`MMF#PT\B!0;-^&(8Y%@#TNP\DMH!FB*!;YG2OM(Y>(LA5O5JU=LYU[RF&
M2@DC>(N'3?";WP$3KN^4.$?]:LH*/*T66SA[PYSK;*=0]1MPA+Y?K2NK2:(V
M5T$9P%L\9H+?LL?MM[Q#X@QE&3`3MISLG$W*19P<GN\2(69X7:]#T.5?IE7M
M?:ID-(2W=LP$3VG$+;B\2^(,'TQ#7-FC<-FX4Y\]C3.`^(P<M21@B*P)7_7O
MQ7T7<[?)*L((XC-R1&[DJO5B#DO_"9U49N=?X^UY0'7JBP5H(&%NGBR(CB"N
M7B'"#.MVRQ&3.+E>^@)]YV"N'#H"N-Z(48JR(?J0%42T2WR$*^8R60T=0EQM
MQ3C%YO?>"]-BF58$67O/U;U#Q=`AQ/5F#'/,[LG*$D%V42]6S7'DR6+H".)Z
M.X8YAA?&]4OM8CM&"X1Z7BTG+QJ)/D*K.$ILT'BN)CH"&&*H2=/.;0.L]@6_
MN4<TSU!<Z>SO3,E6:F5@B>.!>:Y.-@"XWH:QBNB($9?W2IABY_SPY)D=R2&]
M&!/J9&EP!/#Y&,IP,D'4)=L?K,2R7^S"([&"W_F1Q`TBOY$/V[Y3@G7!`<#E
M4R56%QRPX?I>&>#8/&V<9(+8YTKZ&-;N:OZ1?J'$V>M?'?S@1Z^IE=E"V0CB
MZK$3+(@.F'%]QPQP)/#W6'-*NJ<?Y3_US$X>K`X.(3XC1TZ6X4;UT`VQYD;[
MJ"0^'+M>F&MEO\)V+A52U?NC)NN#(XC+ITRP/CA@Q_4]$^?(V3]S*/J8M)7`
M$LDTF^X8_7VO#S9JE@$2+PDRS=8'1Q"7#YY8@3!NQ>7=$B.H2X)=$RZ"6J[3
MTA$TNZ!]OE<R;E?QMI+84Q>YMSY9'XP#KA\WP7+9@!67=TN<(J#?X<4225KQ
M"+1B?Z;X$>X7_5S<;PB155/MW0FI],E2]0C@ZJ$39J@%(ZO]R9[:]"A;U\NQ
M.\T5"4<`5]LP6",<&#CKAV*<8V:R="FPG]LY:IJ=RGR_=(3MQ8C2J[WQ*HI3
MFRT5CB"N'CJQ4N&`$==W2YAB(7\$J0!FNT\5)/J$EB_H%LZ6N].RE%NQ@NRG
M=:Y<.`*X?.3$ZH4#5ES>+6&&LG%[&5P"-[O\.TEH5WF^5S@AN6)Q[>QH?:=S
MJ=Q8#6D$</&P"=9'!VRXO%/B%#OY]Q@DOHV]\"6F;LPT61X=0'Q&C@#=Z_]=
M1IB5[$C/H.T_(0_6%480GY&C[/5>M`,29UM3@K6+36%^"O;M$6/2%[K)`_A:
MVVR-=`1Q^3(1IM@LE26"W/4;)K$WVJ'&R2+I".)Z,\8Y(OFX3:W:BWS89(Q1
MFRR2CB"NMV.<8P+G"%CM=2?4+Q]ZGBR2#B$NMV.8H\1JN"T49&][%O%V$O;)
M(ND0XGH[ACE&E\?U"^YR.P:KAGBT](M?Y06ZH&B9;;9,&D(;X588O5[(]C:S
M;H$DOD*?YT?VXH96G*"P'_NU*T0FRV<CB&MM&*J0ADVXOE,&*")W\*.JW9^&
M[!)7[UZ!"9<+AQ"?CV)U<\LZE[I==,NYG<E"AL<A=T]?:\["WO;4BX!@GWJ-
M%0M'`!=/E7"Q,&K%Y=TRQM!?+)._M%,A6=\@+;.=HH&I5\QDY['3OJB9]$J3
M%;,1Q+7C)EP<#=IP=9\,\6OX4-O*5CB2@)73;,UQ!/#Y&&[/D9B?HMDFE)FS
MSZG$AZ&X>>2'5'NRR$:WH31;'AP`7#Q1HO7!J`V7=\H`P9+`;%U('R[VW+$L
MXGF^4V1G\/LUB7/RZQ2HI=8FZX,#B.L'3JP^&#;C+]$O<8KDCP\V]B><C[ED
MO9UJOE]D1?7Z8*WZ=7762ZYD"I6Y^N`(X/JA$RN6#9AQ?<?$.4H$ZH*@-R8J
M8M/'A_MTQW1_1HKTX6(SMRQ2Y<R4CI5%!P"7CYPH0T[%KRO@)C;VSQPD!N]U
MLD`X@KC:BL$"X<#`63\4XQPI>WY4=C[U!?5]"6*>[Q9@SQT#Z].W-L3*[O*[
M<'UP`'#YP`D6",,V7-TG<7[=2V"H3Z_;BHABE[TW%.^37)IIQB)31:]3D(U8
M=EB>+!`.(2X>-\$"8=R*R[LES%"<+T_PBBV[O3*N#S\BXR7=XAEUO4_6KK]+
M^M$NY;F2T0C@\I$3K(D.F'%]Q\0YYLI6K6J0`3S"J;)8X&1-=`#Q.3GJ$\A^
MIC9WL&O,M*B99\L((XC/R-$=$[W'FLDN)JPDODR?GX1`VZN'LF6"W5`K!+!.
MED0'`->O$E&&R<*^:J4"^W)"C(V8<;(@.H*XW(I1BO+WX)HE%K*9*EXOG"D&
M!>NA0XC+S1CGJ/4&FRY5HA9[54G?VLF3]=`AQ/5V#',$H&V1Z%XK$W>G(D_6
M0T<0U]LQS#&\-*Y?;)?;,5@F+!H?U.Q'R$I/8*<%FQYXFZB+CJ"%^;7MD)*X
M"G9@G%-.?;]5QNB)\2U\%YG.?@)6(N.<)VME(XAKS1>KB`Z8;WE_A!G6[)?+
MR<BN6Q4E\>[Q@6AQ<`CPV2B6GBT!"?JNKCG[K<.YFD)T'$I4T/V64<ALUTG4
M7K1".%D?'$%</U<B%<(1*R[OEBC!K68EH5U'"]5JJF<<EWBG@#])C'K=H4BQ
MA@4X62N+POT"0R94$!TQX.H.B1/,VXF[BIJ%R^*LI7VU*E8;',%[-G[-;PJN
M1?./$E6P7@HZ/P(1O!0C2X,[/KGMW]B.E@4'\!9/D5A-,&[`U1TR0%`]52W7
M%`TQ%9'UQ2N<[I&2L3I#KI;#D+%8,\Q6IT<0%P^:8#EPP(CKNR7.L69_;).T
MLJ%#1]^*QM3F^Z5R]HJ@;!,-?4Z)X&QU>@1Q^=@)5L<&[+B^9^(<FY=JN\32
MND/K*>R&^Q18O&,>WO"478?L:A2]7.A,E!XJ@P[@K1\Y48:<>+,UZ1U`,L)8
MW-?]5Q_!DN``XFHCADJ"\4&S?!1&^1'X\='4**4-+]7]LAWN$0EE/2^?K'9N
M/5(3["^3B]4#1P"7#YI807#`BJM[)4RP;]DV62L9?364@<0PWRF5K;B?0;S_
M;/,$M/HV60T<05P];H+%P`$K+N^6.$6UB1\'1R,A2S:)%]POZ)="#X*$X`^%
MR5BDR1K1&.+BL1.L@`[8<7W/A#EV*#UO%?GN%RAP[VE_L#Y8`1U!?$:.J%D1
M'68RN>PV@2RK.,Q6(@8`GX]A8[__NS:QNYWQ:9SV7UJ&9V#GUJL/6CNKJ'=;
M9<R4YRJ@(X#K%XDHQ9Z\/%1+\XL[6U.\V0+H".)R*X8ITG8[B(1GG55SDSB]
M0IZL?PX@_F][W[(KV7);.9:^XGZ`40A&,%[?H88GA@>-AF`(<`L-M3PR].^]
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M5TKFQJ./<\>0Z:Q-RD'-CNPC91PGLED`+3]N7,*HG\!TC[@-U#L21=!H2S7A
M;Q@3L5:=9TIA#/!M%MJMYRPOZ2K@DD7)/@S".I?=/-PZTCF)0AUK^4PFC`"F
M=A*G1!C@+]TA$1.-PC[DOD4T9%F7/%B5>WW2NIT_!LE=][",Q:,]B!F?0!A!
MS`X;ISX88#'?+WX;D:KT2Q7=\B>YF)#'?L$QO$SUAW^0.8EC:E_E@6-\&F$$
M,3UX?)I9@,9\Q_A-7+;.MUO*\=/E7?EZ[A7D#6RGN!OKA2.R8W4>:M4!O/RP
M\5HXK_!:57<%V4[+NWMTW"IA!#&91*=,Z(Z:Y"#TFR>WV2K3+`?7%+$+"><N
M6=<%"NAT4R]0$)>4W@Y%P@!B?M@X5<(`C_F>\=NX.AGA?4G][)L<Q7VPRS3@
MF#JUXEJX;+VZJ\^!Q*4?BH0AQ.S@<8J$?AK3_>*W<+.])D\-Z:;DX2P[GN=\
MP2]$%^*2G1YR_@-+N]H.M:,08G;L.(71`(_YGG';N,M:=@>`'#TCC3)\QP.%
MRR6-1@#?9Z*]3@'`)4_$R7*?ZFKC3$\(X+W+OD[VO;*GA?5\`&:U<5\Y]O>^
MJ?NU9&Q"/]2ZNKRZ=RQ7!Q#S1PB_C5CE;O.+5>&V-+L_QN:410.`Z2SZ3>QK
MVU,?&Q%CC\VAXW8^5$4CB.DT^DT<7,W$Z[CB'A,T\J$H&D),I]%OHSWJ@E&"
M-"??6'9TYD-5-`*8SZ+;1.?0F#[6)K/HU@>'#"*7I(;A<.CU;[UB+&V'HF@$
MT6VC7(YHB&BH4E`MC2J?VPB^#5'N35<5:B/+J^U0*8L@YO/H4T4C/.9[QF\C
M#3N\@J5L540YS_)@]>O3!T.([[.1UU46;W8'8\,2CQ:=A^-B.U_3,'9*RH,H
M&V6M0XW0#_@[=!B72!C@,-TI;@.GG5-%EJ8O:7UKJ_9U#^AWRF[;5)+29S6R
MY<!</Y/-(H#I<>,31@,LIKLE8.+HABB;J:4AR];!!_LP?%IA#/&--EH=M".+
M)97#&*G\/@Y&X'PWL1+K)>K4!\U]*!9&$+/[BU,K#+"8[Y>(C3;_29?6W0X\
MI^BY+SAFVJR!'U.5QR5[C!_DM#ZM,(*8'CQ.L3#`8[YG`C9>!QJQ.!I:9$32
MS6N,<\_0LJO%M:*J-\VT\5!8\4F&$<3TZ'%):`$6\_WBM;`WXW".0K+_X5MO
MF"$>5-+<7JE+9Z*!/&9?NVVI/5B"NE32`%Y^V'@M;-.$5^`-?=IE8I&[[O<V
M.B7#"&(RB4[!,!`WV7$8L'!?9&-1I]D:6!UUCW.O(+71KT;&*,.Q>*5B+IB'
M>F$$,3UR?()A@,9\Q[A-Q.1G=//2+!3AWGB6_8)?NNU6WJU1J]9?UMTCEVZ]
M,(28'3M.O3#`8[IC_";.2P^79TELH;,Q2-`KCL$PJ`VIR-TD50^*ED>;`%Q2
M4@@Q/7A\*FF`Q]_!,VX;T=JB;#:NI*M?V4JXSZ32".#;3$2R435VJ$K=5;Z9
M8$&G0XDA@O@^$VNQ!U>Q&IO:D.3*_[W.NV!;=ME]D3<K5"/""+1.)>P`8/X@
MX381:8KY1;:VZ,Z".=9=?=.KE$8`TUET6WAMZR%\W=(B(W)SWJ<Z:00QG42_
MB6V-*\1VUU=/T!.YCD.A-(283J/?1IKM&B/:KG898!T/SN&YE-(08#Z-7A.]
MXV+^2)O+HELQG-_*:/527ZKF">!]]%+Y4">-(+IM[-,V@K?&>I'*-_SWI/F*
MC8L-D?NPR;</$3`.Y;,(XN_`HTLGC?"8[QF_C5CM&H_(S'0'[D:?V/-,-HP`
MOM%$FM=-D@4-;%F[YGWW]T>CW"QAY[^FWD,MN^KQAT/=,`*8WV%\PF&$QG2_
M^$W4@&$1%G2:TX<_N="Q7^121@/LFCE1[UCKG4K8`<#TP/$II1$.\[T2L+%>
M[[J6-51?6!CL'^K#'M4PAOA.&R\YI_=B8DEI#[8SN8-QETK7V4;589`$RSRS
M#T7#`&)^A_%IA@$6T]T2L-"&;M$,9"3])LH!]?&"6WA/,Q&&Z54&\N9\'8>*
M80@Q/71\FF&`QWS/!&R\7J@>1/J<MCP66._*W@'/$):P9F.O8VB*/C#LT#P4
M#B.(Z='CU-$"/.9[)F!C[RIQ#$R2>C-\EQ<2[[?L!3RS;34($G6K)Y;6?95Y
MII<&\/)#QVLA#4N=X)*F'!;"<%%/5>P(8C*)3N$P$#?YD>BW$=^X+KJG'O?K
M6,C71>=^P71AMQH,>./:83Y7J8?*800Q/7:<TF&`QWS/!&SLUG#)(UM51\:V
M5Z<7/%/MLA)1`>?W/G,WGKBEPQ!B=O0XI<,`C_F>\=N(U%Q_M1RFJ+KB87FF
MM;_@&;)KM>0)^ZX5C+G[6'2H*840LZ/'J9@&>,SW3,#&W8V>6ILF,KSE;?)Y
MJ)D&$-]IXZ8R[)T&#.>:T$LJPOU0;8@@OL]&+-8L>O;D,K0\N,L"E>>]D*X=
MR+L5NMZU'_/NKF&O;!H!S!\HO"9BJ6?1B!%ZZPL:7-!Q3W73"&(ZBP$3B\U&
M8[,&HT1O&^M0-@TAIM/HMW%:)H/4MDQ[$J=CR3P/==,08CZ/;AO[OH8)Y`=;
M11NY=[@=*J<AQ'P>W39Z!\?\X3:91[>&.)$6%,L>F:3&^FME>QXJIQ%$MXUC
MVK%CIFH+7`;>>,7$9CO.68XOLXJ-L(#GH9X60LRGT:6<!FA,]XO?0K[N*:V%
MV4JRG4NOAR)B"/&--C8+,FY(VYHM_G=;KT3C=8R$16N7)4\;R&X?E`E\0F($
M,;_'^(3$"(_YGO';6*]B^6!]#.);IR$7I)U[9ET52AY#"_'HME@;/2C_N32U
M"&!^\/C4TPB-^8X)V#@L>-:TX\R]U7:H3(;@WFI>+Q8_2-"XW5U`$0G#0?6R
M3T\^(DU&!C?'H80804SO*3X%,<)BNEL")BX[HMCJU,?F.I+I,E]PR[J>@I6K
M`MGN.MAWN:I;/PPAIH>.3S\,T)CN%[^%UV:-CG'>`&=C>N1IKU]VL^U-G9NM
ML:J\9M/VH8`80/P=8L>GID5XS/>,W\;K5NB!*6`IXD+>W>L+GIFV(Q>#I/T8
M)`;(%_A,.`T!9@>/U\11:%PL`EOK3UV^>!\JB!'$;!:=`F(@=/*#,6#CM3<5
M@4":2P/Q[JKID%_X\HO<OD_F%Z0^]5`_C""FQXY3/_33F.X7OX6SEFI7"8QM
M>R'DL?7[>3K@EVXW,<U2;0&%#M,>J+!.]3"$F!T[3O4PP&.^9P(VZ@%N(+;6
MNTV7M7%_P3%\/8T@3R)H27F3K!`/%:408G;P."53/XWI?O%;N,KUTLEH5*Y,
M=2X^U$L#@&^T<-H]%',-KGH[P6*L2.E08@@@OM/&W8OM9-M8D6EIE7I;XX7^
M5\?W$;21+N8;8:+9AWII!#%_C/#:B-63E>7GVGKWEMRN-EJA0[TT@IC/H]_&
M:5O.Y[Y4XB8WX3T*1Y=>&D/,YM%OX[`[$^=NI&&&;YZMMD.]-(28SZ/;QFX%
M?(06+;-1WNFB8[TT@)C/H]]&W_"8/MZFL^A3#M<WPB+H.H[#^B>Y17`_*M_[
MU-((HMM$OI[<`^#4;'7L4>_W:41,M!(^[*O7FEI.W<U#"2V$F$^C2RZ-\)CN
MF(")UV_NC2V%J=3O]WTX9<,(X!LMO&ID!7/;56I<2/9>B<7%-P569)!PT:F(
M'4),[R\^U3#"8[YG#FS$E&YEFU'*J"]X9IB^(ALS3:7J6`.<"FD!P/S8\<BE
M$0[SO>*WL!J%M)4:.8+;]JET>(+X#@NM(;4]#!$+[_M"8R00^Q6(6`9J'&*Z
M>:"R^X3#"&!V5W'IAA$.\[T2-W$T79OW*1>,[U>\LJY09+TK2[X9?SD/9<,0
M8G;D^%3#"(WI?@F8:.L<_&(3Q1?51_.SWRWS>@L8,X(FLUUV'=^+)4[1,(*8
M'SH^`2U`8[I?`A9>@TG]OA);F!0>I9Y^OUQ7;%=BW5<O.C'?[WQR2J4AQ/38
M\=NXK%@+'N6E`$%$9USK4#4,(/X./+IDPTCPY(>CWT0K'2+AT'JIXMUO0(JX
MI9E;FEPN9FY!DM,.1<,08GKH^$1#/XW9;@D8^#VC;*US,T2LW/HK;F&Z$*N-
MM5U.RIQ*AB'$[-#Q:88!&O,=$S!Q7/OU:%KH%"KW1=43MXS1S2V\:FF'2E((
M,3MT?$JIG\5LK_CM&^4*FTW%XFNV1^F:2R2-`+[/0*OM=R9NMGVKPIAUIBI$
M`-]HHCS7JX"U6D<9`Q%.KW2]6B^W*+8\%<OS6"`-(>8/#VX;KQ4@T[HZ*7IK
M[X?ZJ!\PG\6(B7;E%&*G6%X^F!J/0WDTA)A,8\1$VPX&$TWTA(E[E'JHCH80
M\VGTVV@/7^DP87)!G^OAN./11T.(^3SZ;?0-CNFC;3:-Y=M>2[26\:W(:QER
M*3S&/G[0D)?68-`0`5/DLKO:\/,^OH!,`&H26V@HKYYK0_2R^K&A&S%B8Z]V
M,\:>>F]PE4)ZPS^Y0^S37NY>Z#I%GNF9#3WFXV(B@OA&&^6TKC9L<MY-&A:Y
MV''>(;9FK\$@,<"P+Q>'H`^-^0+B^VQ$J!@BIG1YU[W*1EPLB>]YU#O-Y)5V
M+I*SR6,/F"?V.$=\HXUS;D7LZ"O+KIQ$VD;M8\.U%[.]`3`7Z6,/LB6]K&/$
M=]K8I]T2(?5AO3R7]NX?93$`+JKV)!,6.DNB;$\,/JT>`[[/0EFX:,/!LA05
MMC>^_&YP7$/ORY8IH=2I)@*ZS76.^$8;*UFY'60O;4?(=OH]C5+.LD4KRZX!
MA-A&6ECJ,>#[+"PR,JC"<+T-B(F^W%.(:<,>-][ROH,VNSM!%$![DW67OLW,
M96HSXG%W5;[@%4M*ZE0E51J6UC]6-T*(;[/0-I4W4:N[W;<X[N[6!"#K!5?\
MC3"`:`3*?7CKOC.Y`=]F894GI?2;J39[G%ETG37N$8<,KPT#S"AZ02&Z89E[
MO8#X1AO;UB5WX=5T!3@:QL)U#[AEL0_`0G+=3-4$KK9]#O@^"P==W[Q(;Z+K
M6.FL^J"W++DPHV&<!:*PS0MK1SX'?)N)&"%D1`0@ED':;+2[0^F`FT3=X$"U
M?"]&FX*4X!3O7>9AFC.BF4A7^`TY+8T'JP:6:X30D.K0M6Z=,E6\@/A&&UOI
MUG!-6]F(H'<_RV/RU])"&8.K-,/*"</I,=[[#.RZN&=Y*+3I8Z9$&%4>I&?Z
MHB:&3:3Y^K08)IG.ZQCO7?9A<K#A?Q19&*`=1@=:]UG/]V<Q6IU8VO]"^!M&
MO/$YXOM,1$MK.#'5L33$"K[0W4R_1['=+&!8'C:@;[*]=LYSP+>9R&5;]8]W
MD<.;:#>QI&WW+*YFBNWHHU="P[D)@VX]1WR?B8U-`Y8+!\JTAGNW^UI=&7;.
M<V$>9464XP\/*D(^Q/?:V)?=!+'+PFBH#7OE/N\KE.O[-9:M+K23MS%&?0'P
M?28"QNY*9<GHT`Y+NOVQNP`/:T0MU+'<_=*E(:*RKWD,^"X#D=4L>RUS8PS1
MAEVF@ONXZ:U;&6HUV4:*AA-_?Q^)?L0WVHCI3MN-2K-*N\5M/RB^=Q6_.W*X
M*5(!B?YW?T]R!/!]%O:M);7)LEE>&DZ0^#'$2&^ZLXIZQ])8I@S:G>]K;P'`
M]YE8N]98,1/L4;7A+/.^MV`-9&K#8'G!"<NAABEV]V/`=QF(-%V*Q6@H=WMU
M70$REX^*K2"J,H:&LVY=*78Y'M1>`'R?B7V;0E;ERE5I*,-DNP^;"G8-<*\N
MZ0"\@K_LYXAO-%$O^A"UCZS@0O)\XL?W4`51-_ZJ?K@TFZO(^O@^8P\@OL_&
MM=NE<7;2'+;2&*7,%U13-^(;;70JNZXTEK=6)'55B:%&BM"MU5KG61W`#^<U
M;G6[\`(S8-&$KK)DJ7QNWRAV,`+VE2$5/3D$MF<_J?:$\)(9=)K7KQ.!^,JI
MM?F.I&_>5ZU\Q;P(8C:!`0N+[4G#8@DCL-1B"R.2CFJV(;QT"CVEF5`W2>]W
M3@L7Z?LT#!PY@RSML,HI+_C$]A"W6JC)M#-6PYIJGFD1(<#<H/&5H@(,9GO$
M;V`CHQH\(S73AO7N-<R02U:Y;DQ>6W_);)@\6SM2V2*`Z5'CMG!<CV'6N<E>
MD)L5"="9EAH"3*;05VT,A$UZ'/HMQ#+].AU=Y84_J;#*=N;Y@E>F2:F#]?Y2
M:;CZHTG^^1:!$%QRU/C*JR$*TYWBMW'4:U\GUBG$6E/&V%Y/G3*W[1(755%3
M6Q'=RJ"C;2\AP-RH\563`P2F>\1OX;JTDMF6O,L*0'FVX8%6XO7)'L7>BUNM
MZ6I]X3>-.L[V<X40L\/&5T"/L)CO%[>-2H;=&8@HTSJC$$9T5$0/`;[/Q*9/
MNDF[.=>V.C'OXU!<I!7YNN&,H5M-ZUH/MD5X]F<&X-)[BD\R\!.8[`^_>5V?
M<>NR5X--&)K<:^_''FEL!O*6YP!UHB!@'FX[#B'F!HU+(8E0F.\4MXFC7JH0
M;:F,D)PVGNW!%F&W4SK9[G(@-]U3/[=<0MS.]M2'$-,#QZ<+17A,=TS`Q&E?
MC:Q1J@HR/\JMK_O<,7/:F#CU=AS)1V'LF&=G1@*`^:'C-G&7:0\^KXVI5PN-
M>R)E&&?'@T*(V33Z361-V-$0_V28#+!F?6"B2R\((:;3Z%(^(_TEOP?Z3>1Z
M103IXF^W@G%EGPF?`<#W&=C+U)%[(RSTO`/A#VL?ZIXQQ/?9R'KG"Q!;E_6J
MJ+V8K\8^TSUCB&^TL5\\,FU#Q(JVW5>U`Z-.[=>@8T<[,>#`18?2;`@P?5ST
MF^AE\7?P2R:-;J$(#9N6FFH%L)2:,/CL^U6@0_IT0[D-6]6VDN*K6740(*^/
M4X#?N&ZGOVHK8D"5:X$7?;C%R"U[1@`3N?/;5NQ]XSJT&E7UOEGJZTSQ#"&F
M\^<VD9==C=EH=8TL9#3CH^CCE0!C@-DLND3`4"=)[W9>$[E<HH6\45>U&2$E
MW\=>Z==-J8R%@B[9^QZ[]B,),(27'#4N/2Q`8+9#W/95OBZWGCRFMBMH]JB;
M>#UBHV''Q*X'T">6CZVM,\TS!)@=,FX#K[T!H]6I9X)D2/EX5:%7``P!9G/H
MT_\B89,>AWX3N32[1F?TV;HV++,\V.?C=<L@D_:$9)/S,>&O>20!AO"2P\:E
M`$88S/>)V\3>["+:@0AG:S=WK>/8*9/6]9#<9%WTXV>,]F`\="F`(<3DN/'I
M?Q$2T[WB-W%V>TIUR3[PI@TWDNY^[I8U6$L[!+MT3%R8!=J#/6<N12R$F!XZ
M/M4SPF.^9_PV[GD]V[7+DN"9\I;.?3N/$AB!>Y-]G:SP2P,+.M7$M@CZYW&(
M,;:9>$:D*TGIHSR/E.@`7'X_\:F``0J3/>*WKPVK)($XA(HBEKX+G;JDDIUE
ME-VRI+J0K/#J/!,`0XBY0>.3_R(4YCO%;R-?]3TI,191"V9%Q&T^=PML5,&\
M[P4@F>H7]WD_X_H$P!!B>NCX9*$(C_F>"=BH+R"@H6R7U5D/<]Y^P3&C5&TX
MZR)5/>67[/M%@4OU#`'FQX[71+N#4F2SAG1/M4RL)FB=B9X1Q'06W28B5DTJ
M1`*D.7ME1A#QF402`4PGT25[!CI+>N_S&\C%O#SQ@]=UW4=;AYIG"/%])K92
ME<35Y]*KQTB.`^Q])GF&$-]H8UU-.SY6H<W.$^\Y#^7$"-[;[+OV$"&-*W9C
MT28LN-OY>%-KK=>`,PJKR-"X43N5.P.`V4.BVT(WB?EN22;1K1&5U>T1,]VM
MJ=>9%>)Y*'_ZT=SV;=EBH-7O.8FT89&']<[LFYWMQ=E"JPG1"#8DLN-4^O0#
MYM+G-4^VUBEY>U95>O?]_;$1U=,/F,V?W\)F#P@A^]]:H9`7%(C/Q6@_8C:+
M7AW0WTG2NYW71"(NUZN86)CJ_7*8\HA?\`K94U8+JP95`AOQ'J="H!<N/6C<
MFIB7P=<]\H<__.&I;;^VH3HO00I#5:FF_769K7]M\KD/?FU21C'182!S9\W`
M%CI'[S<PG^F;01B'-9\$0]`6>SYCHJOI*D_NVZ.Z;F`^5?."."G4_$"YB_@[
M*6Q<UK3O#P@B8Z*F;385*0X%2;;;+U>K53?WS=%D`_1O83Y5Y$(H21[_@?H6
M(B>+9)<]O*['X7J1E3UPBKR.LZ,LLXW!>[?6[`KQ(=_U6YC/1;483HK3?R"?
MA<C)(MEECUR\KVVF'MI#&RP%=ME1EN?U;@/U.50&Q0I/WB[[)\KG4E$$)<GC
MGPM_(6JR*/:8,Z\Z69%36'K/Y^AC[-^V^5SM"L%\H36[#`N*N<I0H1:KZC)6
M-'!6N=Y*(7E,6(6+67G<</RICA6"R8GC'RA6(7*2./:84\MW<AB,&`XZ8^M!
MCC>1<L,#ZP6[8Z..<8/RN0P5@LEQ^`\$IQ`W61S[[&E67<-<0DUQ]#')$609
M285^5>]#MU,CFI>>_/PGRN="4@0EQ^.?:PDA9K(8=IE#S;@!*2Q?M=JH6A.,
M48Q$U8Z7CK:59*S$U[R9%W^@?\5P<ESNL8:1X-@Q13LSRUBV\FTR\[G`%8+)
MXL9C#E(R^ZY5JA:0:I<<]';N_*SJ'8))8N<'@E4HDI,ZA,<<VNV2'W?526^/
MU6^GO$_EJ`#(EUEB+^5V769IAD*]R;N/O\7Y3&Z*H7RA-:/9(>2)K]!+D)$E
MEU)O4#[55H(X7V>-/D(@.`-K"NF6I8]Z,_9Y>O<LU@_0NXLF.G+17NE\8\ZG
M.E$,)V>P\9GC8">+Y21Z?E`EO]HLI'1]7Z_H]JV:3MWX4[FI-WZN^\1P//:T
M;DFG/+:G]4^TV;?539<]=+TN@<7-TEP:<<&UWF8[GZH[,9PL?CSVE'WQ@[R=
M22_?!5&W];G/Y)L02A8Y'FN0#6K="`.8'CKH%9'"-['\N6`1Q,D@YZDXX0SD
MI`[A,:=?.\M%69LZ-A&2_QKGN"Y[3!M,ZLGU+GNN-P?T"2]*CL>?EO)]Y"1Q
M[+%&GD56<K!PUY(M^IV\4A^D6+9R"8K<%2GV8@'1]=6"@.CBQ<EQN,L:ZE;.
M-IBQY,*#'I(GO#`IW#Q7)WP>3PH<ESFSFD0K^U)U"Z24AE?C4Y8W4;-3@[2H
MCQ61)[PP*4Y_KD[XV,EBV67/*M9&9@I[AJ8U*HNC+,N35%(8YU&+E3Z6[`(-
M"11>G!2G/Q<H?.QDL>RR9U=;*V)E45701$_M.Z@#"3^+-=O%XDPG/GS/J'4'
MBOD^D"27/]==?-QD<>RQ!ZMSRQTQNK2M91]J6-_&1`HGSE?:0Z-H?HG&78_0
MRDFTC[G(T^"AOH9)$-T>"`)(9PJI%#Z8C$!^KE#XB,DBV&5/E3*UM*F[ZWW:
M4Z)BU2C'5>]>[]]8]G>P==!.-\+R4Y7""9/C\N<JA8^=+)9]]MCUQHS&50O`
MR$3Z3<740W)K9C%6)H-T(3/6OM%@GVD4+I`D?S\O>[N(2>+79\U5@>N5BVU2
M`?!M/OV<8+[*BEAWLFZTE&ON^UPAV<6)D^%NES%C74>?9!6MA0_\Q>`:$5V<
M,%G<.,P1ARO,GI85U8&HN:F//Z\2.W&2V/FQ\N**X:RN\-043,KF)30M6A<I
M;7[X/4]E%R?.EQF#'-M\O623J&HJ%>2,D.[BQ?DZ>[`(&=\9'-L>=:P897=(
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MLW=!+DSZ5<A_=IQB>WM!DBG+0+%`Y9A0$8-YS=]/50HG-5D4N^R12PKT9V&8
M:9:S+7W@(DCRM.P08YH=ZAXR18Z84N'%R7'Z4Z'"24\6S1Y[R(JE&"Y%#R*Y
M[(P^[E%RL-R+/;560:"^6#IVQ8*R1ZKZ7I@<IS^77WSLY)#LLJ9>CW01\=:P
M(*1D)29Y^&"^T)IV76^"A>:J4ZM6#4-AC4:.59,;4.R<M>@`?49T"A]*3A0_
M5RE\U&11[+*'B>@ZR'2=!T7LKRC'Z`WVHN">K2H[4@)O*R15.'%27/Y4J7!1
MD\2PSYAM*%-F!-T`(;)Z"S-L!`.>B6VUN=>8$:G"AY+D[J<%<!\W211[K&G;
M!A.6NQ1T$!VRQWP%*9X7C&QF')+EX?_O63@DOSAQ<CSNL0;&V%6<6(7J,,J\
MVX<M<L_4%Q],%C<.<P8U8Z?+%?=:8D'"]MOCEYXZL1,GAY[G,HPODI,ZA,^<
MZUG87O5H,MH,#"DM),+X8+[0FHH$T31`9DWMD53W=;O;_:D$X\/Y4GOT)5+%
M6;J)OQ(&U=U"2H43Y@NMF6SDR),+JIMPJ=Q7N)]W2V70SYOV3[GOJ]$.ZB\^
MG*1QQV6/AY\LGG/X\92^FT8/B^U33Y!7V=W1:E!^\<`XK.%"MNU]2^'-?KFD
M)Q0V!YW$[-GHXR*ML)ZY"<HO/IP<>CSFK$LLVU3M#?NV>Y\W2Q6/_N+"R:+'
M94]K5AUJO9!.Z@VI_(C*7#Z<)'Z>*Q6N:,[J%0YSJ-@#:[*$(KV8%>O'FZ63
MCV2RDU%[==91CB?2BIORD$>J<.%D.?UY7=_#3@[''EN&W1$G1>]BA\FEMMK"
M%%L93Z8&:2#OO9:@P.7`2/*UPY:YKSI?'U9%Q@H5BX"@3N&#>9T9CT;A<G5.
MQ+C,J96JR09VXK_LW7][XM])\)J&LO5[$"+\48)\IDYX,#(<[5$F'*PDD>NS
M9MA!_\*KZU43\'-O/<ROE83D,FT5D;!RQG]&4)?PX;SN;H<HX2`FB5^/,:V1
M%=X+VZ.%15Z1J3W*[RKVF@_XZG9O,GX"KZB<Y<))\K=#:_'0D\2RRYH]^/I1
M51\&05@4XA64)5PX7V@.]Z+L5*2(^A(4I@>YFRL<.]RN8S6-=<V,")(0B0D3
M/IB<4'9($QYVLEAVF;.F*094JUZ?AF_<:XXHR<N.\L&87JU_/MAY^%2<<.'D
M^-RC3GC822+99\XNUZV$^$XMD191A:(24).M$EI`IKVW+C.1>M+'[:_/]`D?
M3([3/35O#SU9-/OL&7;:3+:9ZRS!4AR>,9;Y>Y&9QY@Z+/?"'X9NC^[BPDER
MNL<<N=Q!V\@DHPDTU46+8[*+"R:+'8<Y&(&M0"KG/[2PLN3AB6!MV`&3Q(U#
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MPG5S^9Z/XF%[KY$\6%F1D6+TWH+2A`\GQ>6>0KZ'GB26?>9(\J_L2,54=ZD4
MHK;"+-OCO%+BT_MT92Y$#XA(+BZ,+(<_M86M9U*?1><\.&#B/T&%PH630(Q#
MGO"X.BMD/.;,:DMHDAL?=7\4TN;%*\XP6;U5G@!4=C!%E)MUZU.!P@>2XVZ'
M1.'A)HEBES5L.]D(,Z[&'_Z`G[=C#-O75"2#TSKNV..VW.A0)WPPKSO;HTUX
M>$FBUV?.NF(/>:CN92)YB*W4,,5L:4.=L-AV\@ZY6#Q8Q7?AY+C<H[EXZ,FB
MV64/@H^NWT5LSUGB5]UF50YUPH7SI?9T6T6"E78-3/*J2`^'S^C+SK"4;<&S
M;E?/'GG"@Y(3RAYUPL--%L<^>R8;.[3GN'`D#L(L;SL9U@8-W<R+A2M3I:`\
MX</)\+I'G/"PD\5RS)Z.D+UZ*)*Y%F6YDSUF)(VW\C/:K+<RK4.?<.$D>=U1
M]O;0DT6SRYQ^279@P,;3-LI:%&9YS*O-8GVF":M_]/:H[.+!R?*ZRQ[\*VN#
MU=C2?%HDF.537OX=K?Z!__L??_O+__SK?_SGG__T][_]Y?_\Z<]__^6__RC_
M^J__];__]<]_^_M?_M>?_^\O__9+^Y>'_VM?]//#GQ_^_/#GAS\__/GASP]_
M?OCSPY\?_OSPYX<_/_SYX<\/?W[X\\/_#S[\XR]W_V,5(/NS_O<__OB//_X_
(XF.\,Q9<`@``
`
end
-----8<------ [snip] ------------8<------ [snip] -------

    pederb has a patch below that "fixes" the problem, but this is not
    a proper fix. Somebody needs to investigate to figure out what is
    really going on. A good start would perhaps be to test on a
    variety of different OpenGL drivers to figure out whether or not
    it is a weakness with NVidia's drivers.

-----8<------ [snip] ------------8<------ [snip] -------
Index: src/misc/SoGLnonindexedTristripSetTemplate.icc
===================================================================
RCS file: /export/cvsroot/Coin/src/misc/SoGLnonindexedTristripSetTemplate.icc,v
retrieving revision 1.6
diff -u -r1.6 SoGLnonindexedTristripSetTemplate.icc
--- src/misc/SoGLnonindexedTristripSetTemplate.icc       14 Feb 2003 17:31:28 -0000       1.6
+++ src/misc/SoGLnonindexedTristripSetTemplate.icc       20 Nov 2003 12:43:03 -0000
@@ -67,8 +67,12 @@

     glBegin(GL_TRIANGLE_STRIP);

+#if NBINDING==PER_VERTEX || NBINDING==PER_STRIP || NBINDING==PER_FACE
 #if NBINDING==PER_VERTEX || NBINDING==PER_STRIP
     currnormal = normals++;
+#else
+    currnormal = normals;
+#endif
     glNormal3fv((const GLfloat *)currnormal);
 #endif
 #if MBINDING==PER_STRIP || MBINDING==PER_VERTEX
@@ -80,8 +84,12 @@
     SEND_VERTEX(idx);
     idx++;

+#if NBINDING==PER_VERTEX || NBINDING==PER_FACE
 #if NBINDING==PER_VERTEX
     currnormal = normals++;
+#else
+    currnormal = normals;
+#endif
     glNormal3fv((const GLfloat *)currnormal);
 #endif
 #if MBINDING==PER_VERTEX
-----8<------ [snip] ------------8<------ [snip] -------

    20031203 mortene.

=====================================================================

099 View frustum culling is not done unless there are cached bounding
    boxes available.

    E.g. if an app programmer calls setAutoClipping(FALSE) on a
    standard viewer, SoGetBoundingBoxAction will not be invoked each
    frame, and there is no other mechanism in Coin which triggers the
    calculation and storage of bounding boxes.

    This is not a bug, but a performance enhancement suggestion. It
    can have a dramatic effect under certain circumstances.

    pederb suggests this can be solved by letting SoGLRenderAction
    invoke an SoGetBoundingBoxAction on the root node before
    rendering.

    20031211 mortene, reported by pederb.

    UPDATE 20040909 mortene: the above suggestion for a fix may not be
    such a good idea:

      On Wed, Sep 08, 2004 at 04:27:55PM +0200, Karin Kosina wrote:
      :
      : [...]
      :
      : And: why don't we implement pederb's suggestion (applying a
      : GetBoundingBoxAction before rendering)?

      [pederb:]

      I feel it should be possible to toggle this feature on/off.
      Other applications might apply an SoGetBoundingBoxAction before
      they call SoSceneManager::render() (or render the scene using an
      SoGLRenderAction directly). If you have a scene with lots of
      animations, SoGetBoundingBoxAction can be pretty slow.

      If you have an application that does multipass rendering, and we
      apply an SoGetBoundingBoxAction every time
      SoGLRenderAction::apply() is called, that would be pretty bad.


    Note: when this bug has been fixed, remove the warning note on
    So@Gui@Viewer::setAutoClipping() which mentions this problem.


    UPDATE 20060208 mortene: pederb added support for automatically
    applying an SoGetBoundingBoxAction, but it is at the moment only
    possible to activate with an envvar -- as it is basically a policy
    decision whether we want this or not.

=====================================================================

102 Lack of control over OffscreenRenderer context setup

    There is currently no way to enable stencil-buffer and other GL
    features when creating the SoOffscreenRenderer context.  There
    should be a way to control offscreen renderer features in some
    way.  Requested by Josh Flowers.

    20040203 larsa.


    UPDATE 20040203 mortene: my suggestion is that the API should
    present the possibility of setting up a callback weighting
    function to run the available GL-capable contexts through, because
    this will give the API client full control if he/she so wishes.

    Something like this has already been implemented internally for
    SoWin, see SoWinGLWidget::weighPixelFormat(). I.e. default
    weighting algorithms like that function should be available inside
    libCoin, with the option of overriding them through the API.

    Designing the API so it manages to abstract away the peculiarities
    of each window-specific (GLX / WGL / AGL) context struct format
    might be challenging.

    UPDATE 20040922 mortene: there have been several more requests for
    this feature. Lacking control over the GL context capabilities can
    be a serious show-stopper for programmers wanting to use e.g. the
    stencil buffer, so I consider this a medium-to-high priority item.

    UPDATE 20050510 mortene: something like this has been implemented
    by TGS, with among others the SoGLGraphicConfigTemplate class.
    Consider using the same design and API.

    UPDATE 20060224 kyrah: mortene has implemented a hack to allow
    requesting a stencil buffer via the environment variable
    COIN_OFFSCREEN_STENCIL_BITS. Just noting this here so we don't
    forget to get rid of this workaround when this issue is finally
    resolved.

=====================================================================

108 SoVRMLBackground (and SoVRMLImageTexture?) texturing doesn't work
    with SoOffscreenRenderer.

    Most likely because SoVRMLBackground uses an SoVRMLImageTexture,
    which loads textures in the background (if Coin is configured with
    platform-independent mt-support, which is the default). So on the
    first frame, which is what the SoOffscreenRenderer grabs, textures
    will not have completed loading yet.

    Should perhaps best be solved by using a new element
    "SoSynchronizedElement", or some such, which tells traversal
    operations whether or not they are allowed to run in parallel
    threads. SoOffscreenRenderer should then add such an element
    before invoking the SoGLRenderAction.

    20040427 mortene.

    UPDATE 20050505 mortene: Colin Dunlop reported another effect of
    this; fractured textures (i.e. SoTextureScalePolicy::policy set to
    FRACTURE) will typically render in low quality. The fix outlined
    above should also be the underpinning for taking care of this
    problem.

    When this problem has been properly taken care of, remove the
    following hack from src/misc/SoOffscreenRenderer (which doesn't
    seem to work):

      const int bigimagechangelimit = SoGLBigImage::setChangeLimit(INT_MAX);

    (And check whether or not there are any similar hacks to also take
    care of.)

    UPDATE 20050512 mortene: correction; fractured textures are not
    scaled or otherwise prepared in separate threads, but gradually in
    multiple render passes. The SoGLBigImage::setChangeLimit() is
    there to make sure that all necessary preparations are made on the
    first pass. This seems to work correctly upon some slight testing
    with the following iv-file:

-------------8<---------------- [snip] --------------------------
#Inventor V2.1 ascii

Separator {
   DirectionalLight {
      direction 1 -1 -10
   }
   PerspectiveCamera {
      position -0.47643667 0.5896858 6
      nearDistance 0.1
      farDistance 100
   }
   Separator {
      TextureScalePolicy {
         policy FRACTURE
      }
      Texture2 {
         filename "/tmp/M245-m040904.png"
      }
      Cube {
      }
   }
}
-------------8<---------------- [snip] --------------------------

    I'm therefore not quite sure what the problem is at Colin's end,
    so I've asked for more information. (My best guess is that it is
    downscaling of tremendously large textures to not-so-large
    offscreen pixmaps which may look bad, due to scaling algorithms
    tuned for speed, not quality, in SoGLBigImage.)

=====================================================================

114 VRML-nodes are erroneously renamed upon import.

    VRML-nodes can legally be named in ways which are not ok for
    ordinary Inventor/Coin nodes. They are therefore sometimes renamed
    in error upon import of VRML97-files, from within SoBase::setName().

    I suggest this is fixed within SoBase::setName(), by making the
    code check whether or not the class is internal and has the
    "SoVRML" prefix, and if so; do the Right Thing wrt the VRML97
    spec.

    (BTW, do we have the same problem wrt VRML 1 nodes?)

    According to kintel, this is probably quite a bit of work to do
    properly, one will for instance have to read node names in UTF-8.

    20040521 mortene, bug reported by kintel.

=====================================================================

121 SoOffscreenRenderer tiled rendering causes problems with
    subsequent viewport changes.

    The SmallChange ViewportRegion node sets up a sub-viewport
    according to the SoGLRenderAction's viewport settings, and does
    not know about whether or not the current rendering is part of a
    tiled rendering. This causes artifacts for the commonly used
    "sub-viewport + orthocam + scene graph" trick for rendering
    overlay graphics.

    This needs to be fixed partly in Coin (there is now no mechanism
    for passing on information about whether or not we're doing tiled
    rendering), and partly in SmallChange (ViewportRegion must set up
    the sub-viewport according to the current tile).

    As for the Coin part, this could perhaps best be solved by using a
    new element "SoViewportTileElement", or some such, which should
    contain information about tilesize (absolute and
    relative). SoOffscreenRenderer should then add such an element
    to the state stack before invoking the SoGLRenderAction.

    This is a tricky issue, of medium priority.

    20040705 mortene.

    UPDATE 20050520 mortene: bug item #191 depends on this bug.

    UPDATE 20050520 mortene: As a side effect of this fix, SoCamera
    could then also do the view volume narrowing itself, removing the
    need for the current ugly hack in SoOffscreenRenderer to do this.

=====================================================================

129 Dragger bugs. This is a collection of bugs mostly related to
    scaling of draggers.

    JackDragger: Scale down as much as possible. Use viewAll to keep
    the geometry in focus. When a certain limit has been reached, it
    is no longer possible to click on the drag tabs. Might be because
    JackDragger uses childdraggers to do it's job, and child draggers
    currently do not have a good enough minimum scale restriction.

    --------8<------------- [snip] --------------8<------------- [snip]
    #Inventor V2.1 ascii
    JackDragger {}
    --------8<------------- [snip] --------------8<------------- [snip]


    Scale2Dragger: Scale down alot, and you will have problems with
    viewAll. Also had problems with grabbing the dragger. The geometry
    of the dragger does not get scaled properly. Some of these
    problems are probably related to the lack of a minimum scale
    restriction when using child draggers. Others are problems
    relating to the way the dragger geometry/tabs are scaled.

    --------8<------------- [snip] --------------8<------------- [snip]
    #Inventor V2.1 ascii
    Scale2Dragger {}
    --------8<------------- [snip] --------------8<------------- [snip]


    SoTabBoxDragger: Lack of minimum scale restriction because it uses
    child draggers (SoTabPlaneDraggers). The dragger geometry does not
    look correct with extremely small scale.

    --------8<------------- [snip] --------------8<------------- [snip]
    #Inventor V2.1 ascii
    TabBoxDragger {}
    --------8<------------- [snip] --------------8<------------- [snip]


    SoTransformerDragger: Geometry does not look correct when having
    non-uniform scale in the scenegraph. Load the following file into
    examinerviewer and click on the rotation tab on one of the sides,
    and you should see the problem.

    --------8<------------- [snip] --------------8<------------- [snip]
    #Inventor V2.1 ascii
    TransformerManip {}
    Text3 {
      string "b"
      parts ALL
    }
    --------8<------------- [snip] --------------8<------------- [snip]

    Also, free rotation (holding shift while rotating) does not behave
    100% as expected when the locater position goes outside of the
    projection sphere. Using the same scenegraph, press shift and
    rotate the center tab, and you will see the problem. This problem
    happens when having non-uniform scale.


    SoTransformBoxDragger: Unable to grab hold of the dragger tabs
    when scale is small. This is probably also realted to the lack of
    minimum scale restriction when using child draggers.

    --------8<------------- [snip] --------------8<------------- [snip]
    #Inventor V2.1 ascii
    TransformBoxDragger {}
    --------8<------------- [snip] --------------8<------------- [snip]


    Scale1Dragger: Impossible to grab hold of dragger tabs again after
    scaling as small as possible. Tabs should have a certain dimension
    even though the dragger has been scaled down.

    --------8<------------- [snip] --------------8<------------- [snip]
    #Inventor V2.1 ascii
    Scale1Dragger {}
    --------8<------------- [snip] --------------8<------------- [snip]


    Scale2Dragger: Difficult to repick dragger tabs when scale gets low.
    viewAll does not work properly.

    --------8<------------- [snip] --------------8<------------- [snip]
    #Inventor V2.1 ascii
    Scale2Dragger {}
    --------8<------------- [snip] --------------8<------------- [snip]


    Scale2UniformDragger: viewAll does not work as expected when scaling
    down alot. Also the geometry of the dragger gets strange when having
    low scale.

    --------8<------------- [snip] --------------8<------------- [snip]
    #Inventor V2.1 ascii
    Scale2UniformDragger {}
    --------8<------------- [snip] --------------8<------------- [snip]


    General observations:
     A general problem is that it is difficult to separate and click on
     a tab of choice when scale is small. This should be fixed for
     easier use. It would be nice if the tabs are spaced so that it is
     possible to see which is which. Dragger geometries should also be
     fixed so that they are intuitive and easy to use even when scale
     is small/big, etc.

     OIV does not have a restriction on the scale when using child
     draggers, but we should since without there will be trouble.

     viewAll does not always work as expected, not zooming in enough on
     some draggers under certain circumstances. This should be checked.


    Priorities:
     1) Fix minimum scale restriction when using child draggers
     2) Fix scaling/positioning of dragger geometries so that
        they are intuitive and easy to use at any scale
     3) See if the viewAll works after these bugs have been fixed.
        If not, find out what's wrong and fix it.
     4) Roam all the draggers, have fun with them and look for
        problems to fix
     5) Fix all other dragger bugs in this list :)
     6) Fix the rest of the bugs in this list (:


     20040810 jornskaa

=====================================================================

133 SoExtSelection selection area rendering bug

    SoGui toolkits needs to set the window id element so the
    SoExtSelection node can know which window id it is being
    manipulated in, so it doesn't render the selection area rectangle
    in all available views.

    You probably also have canvas refresh issues related to this -
    make sure the inactive views does not refresh their displays
    during interaction.

    20040831 larsa.

    UPDATE 20040831 mortene: This is a high priority item, as it
      causes Coin to be unusable when using multiple views and
      SoExtSelection together.

    UPDATE 20061026 mortene: we have recurring support requests for
      this bug to be fixed.

      The above suggestion for how to fix it may be be unnecesary
      roundabout and complicated, BTW. I'm guessing it should be
      possible to just store the GL context id (taken from
      SoGLRenderAction::getCacheContext()) when a selection is first
      activated, and then only render it to that context.

      With regard to picking up events from only the same context, I'm
      unsure as to whether we're currently able to detect which
      context an SoHandleEventAction comes from, though.

=====================================================================

135 Elements used by <node>::doAction() functions not automatically
    set up.

    When writing extension actions, it is a major hassle that one
    needs to explicitly enable all elements used by nodes' doAction()
    functions for the new action. It is a hassle for many reasons, not
    the least that one can not find out which elements are actually
    used without reading the source code. In addition, these are
    really implementation details that may change without notice,
    rendering client code buggy without any warning.

    So this needs to be fixed. Would probably best be done by changing
    the internals somewhat so that extension actions automatically
    inherits a list of enabled elements from SoAction aswell, and not
    just from the other actions, where this works as it should
    already.

    As an example, it should be possible to do something like this for
    e.g. SoLevelOfDetail:

    ------8<----- [snip] --------------------8<----- [snip] --------------
    SoLevelOfDetail::initClass(void)
    {
      SO_NODE_INTERNAL_INIT_CLASS(SoLevelOfDetail, SO_FROM_INVENTOR_1);

     // FIXME: these elements are used from doAction(), so they should be
     // enabled for all actions.
     //   SO_ENABLE(SoAction, SoComplexityTypeElement);
     //   SO_ENABLE(SoAction, SoViewportRegionElement);
     [...]
    ------8<----- [snip] --------------------8<----- [snip] --------------

    Medium priority item, we've had support requests / problem reports
    from at least two external developers about this. It is a
    non-trivial issue to fix.

    20040902 mortene.

=====================================================================

140 Rust in the SoVRMLParent machinery for keeping both an SoMFNode
    and an SoChildList of its children.

    Jochen Stier reported a problem with this on coin-discuss, and
    provided the following code for reproducing the problem:

----8<----- [snip] --------------------8<----- [snip] ----------------
I have a little piece of code that demonstrate the problem. As far as
I understand, all the print statements should output 1. But that is
not the case..... The refcount of the children of lGroup is
incremented during the copy, but never decremented :(
��� SoVRMLGroup* lGroup = new SoVRMLGroup();
��� lGroup->ref();
��� lGroup->addChild(new SoVRMLGroup());
��� ((SoVRMLGroup*)lGroup->getChild(0))->addChild(new SoVRMLGroup());
��� ((SoVRMLGroup*)((SoVRMLGroup*)lGroup->getChild(0))->getChild(0))->addChild(new SoVRMLGroup());
��� LOG<12> () << lGroup->getRefCount();
��� LOG<12> () << ((SoVRMLGroup*)lGroup->getChild(0))->getRefCount();
��� LOG<12> () << ((SoVRMLGroup*)((SoVRMLGroup*)lGroup->getChild(0))->getChild(0))->getRefCount();
��� SoVRMLGroup* lCopy = (SoVRMLGroup*)lGroup->copy();
��� lCopy->ref();
��� LOG<12> () << lGroup->getRefCount();
��� LOG<12> () << ((SoVRMLGroup*)lGroup->getChild(0))->getRefCount();
��� LOG<12> () << ((SoVRMLGroup*)((SoVRMLGroup*)lGroup->getChild(0))->getChild(0))->getRefCount();
��� LOG<12> () << lCopy->getRefCount();
��� LOG<12> () << ((SoVRMLGroup*)lCopy->getChild(0))->getRefCount();
��� LOG<12> () << ((SoVRMLGroup*)((SoVRMLGroup*)lCopy->getChild(0))->getChild(0))->getRefCount();
�----8<----- [snip] --------------------8<----- [snip] ----------------

    At the least, this can lead to memory leaks as some nodes will
    never be destructed.

    The problem is related to the fact that we need to keep two copies
    of references to children nodes for an SoVRMLParent: one from the
    "children" SoMFNode field (because that is in the public API) and
    one that can be fetched from the SoNode::getChildren() method
    (also part of the public API, and used by e.g. traversal methods).

    The issue of keeping these internally in sync at all times has
    been a long-standing source of bugs, and this bug report indicates
    that it is still not working 100%.

    Consider this medium-to-high priority.

    20040908 mortene.

    UPDATE 20060208 mortene: see also items #162 and #196, which are
    closely related.

=====================================================================

141 OpenGL context store/restore "stacking" for
    e.g. SoOffscreenRenderer will fail if different window system GL
    bindings are used in libCoin vs the main windows system binding.

    [We used to run into this on Mac OS X, where Sc21 uses
    NSOpenGLContext, and libCoin attempts to check (and restore) the
    current context with aglGetCurrentContext(), but this is now
    fixed, see update below.]

    We can in theory also get the same problem for other combinations,
    e.g. SoXt for a X11/GLX binding for the application on a Windows
    system where libCoin will use WGL internally.

    20040915 mortene, reported by kyrah.

    UPDATE 20040915 mortene, this is kintel's write-up of assumed
    problem areas and issues to cover:

        Cases in which this has potential of breaking Coin:
        o Offscreen rendering w/Cocoa [fixed]
        o Offscreen rendering using a window system layer on top of
          the native system (as Morten also pointed out), e.g. SoXt
          under Windows.
        o Offscreen rendering of a scene containing an SoSceneTexture2
          node
        o Offscreen rendering done from a callback node
        o Nested SoSceneTexture2 nodes

        What I have here is not a solution but rather requirements for
        a solution that will help implementing this properly. The
        basic mechanism for solving the problem should be the same as
        kyrah suggested initially (i.e. using callbacks into client
        code):

        o Must call back into client code (we cannot e.g. execute
          Objective-C code from Coin)
        o Must have a stack of these callbacks (to take nested
          offscreen renderers into account)
        o Must take into account that multiple threads can
          simultaneously render (by having one stack per thread?)
        o Should be able to share resources between on-screen and
          off-screen contexts:
          - Imagine rendering a huge volume with SIMVoleon and wanting
            to also offscreen render that every frame. Without texture
            sharing, this is not going to be very fast.
        o Should also support sharing between "layered" and native
          contexts:
          - Using e.g. GLX on Windows or NSOpenGL on Mac, we should
            make the offscreen context (WGL and AGL respectively) able
            to share from the layered contexts by e.g. extracting the
            core component.

    larsa's feedback:

        It must behave like a stack, but it must be implemented
        non-stackish.  I'm trying to figure out how this relates to
        how BeOS did things.  There you would have to lock a GL
        context when using it, and unlock it afterwards.  You couldn't
        lock another context while one was locked already, so you
        would basically have to unlock the onscreen rendering context
        while doing any offscreen rendering.  So that's a model that
        should fit below the surface of whatever design we end up with
        on this.  I'm not sure if this is a problem at all yet though,
        just that you wouldn't be able to translate makeCurrent() into
        Lock() calls without pairing every Lock() up with Unlock() of
        the previous context...

    Update 20051108 kyrah:

        Part of this problem (offscreen rendering in Sc21) has
        been fixed by adding a callback interface for providing an
        external offscreen rendering implementation (see gl.c:
        cc_glglue_context_set_offscreen_cb_functions()). Note that
        the other issues mentioned above ("stacked" offscreen
        renderers, GLX-on-Windows) still remain to be resolved.

=====================================================================

143 Warnings typically triggered from faulty input files should be
    easy to hide.

    We should provide a mechanism to hide all the warnings typically
    thrown out when encountering a buggy input file, as a service to
    applications that have little control over their input sources
    (like VRML viewers, for instance).

    A couple of ideas:

        * Provide an envvar, "COIN_DONT_WARN_ON_SCENEGRAPH" or some
          such, that can be set to turn them off.

        * Use a standard "tag" or prefix to those messages, so they
          are easy to filter out for someone installing their own
          error handler on SoError / SoDebugError.

    Both the above suggestions could be implemented. In addition, one
    needs to audit much of the Coin code to find the relevant
    invocations of SoDebugError::post[Warning]().

    Fairly low priority task.

    20040924 mortene.

=====================================================================

157 SoFont does not support styles under Windows.

    When adding a style to a fontname, the fallback font is used
    instead (Arial). This wrl file recreates the problem:

----8<----- [snip] --------------------8<----- [snip] ----------------
#Inventor V2.1 ascii

Font { name "Times New Roman:Bold" }
Text2 { string "test" }
----8<----- [snip] --------------------8<----- [snip] ----------------

    (This works fine when using FreeType on Linux.)

    20041102 handegar

    UPDATE 20041102 mortene: note that it works as expected when
    removing the ":Bold"-suffix, which indicates that the string we
    pass to Win32's CreateFont() is wrong. Most likely it needs to be
    broken up properly, and we should probably pass the information
    about "bold" weighting as one of the other arguments to the
    CreateFont() call, and not as part of the font name.

    When taking care of this bug, please also check how "italics" and
    other styles works (or not works).

=====================================================================

162 Inconsistent traversal behavior upon traversal of VRML97-nodes
    with SoSFNode and/or SoMFNode fields.

    According to pederb, only some VRML97-nodes forwards scene graph
    traversal to the sub-graphs in their SoSFNode- or SoMFNode-fields.

    E.g. SoVRMLText does not traverse its fontStyle field, but
    SoVRMLShape and SoVRMLAppearance traverses their SoSFNode-fields.

    Is there any good reason why we should /not/ let all nodes
    continue traversing down into SoSFNode and SoMFNode fields? If
    not, this should be fixed.

    20041104 mortene, reported by kintel.

    UPDATE 20050711 mortene: see also items #140 and #196, which seems
    to be the same issue.

=====================================================================

168 Memory read/write errors for VRML97 Group / Parent / Switch nodes.

    Valgrind reports

        Invalid read of size 4
        Invalid write of size 4

    for the local WRL file ~pederb/store/vrml97/sony_cam.wrl upon exit
    of the viewer it has been loaded into. The callstack backtraces at
    the point of the errors are:

----8<----- [snip] --------------------8<----- [snip] ----------------
==31646== Invalid read of size 4
==31646==    at 0x408BCA83: SoNodeSensor::dyingReference(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoNodeSensor.cpp:102)
==31646==    by 0x406FD3E7: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:436)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40766885: SoBaseList::truncate(int) (/home/sigma/mortene/code/coin-head/src/lists/SoBaseList.cpp:141)
==31646==    by 0x40701316: SoChildList::truncate(int) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:176)
==31646==    by 0x40701B95: SoChildList::~SoChildList(void) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:84)
==31646==    by 0x407E803D: SoGroup::~SoGroup(void) (/home/sigma/mortene/code/coin-head/src/nodes/SoGroup.cpp:239)
==31646==    by 0x40954FD6: SoVRMLSwitch::~SoVRMLSwitch(void) (/home/sigma/mortene/code/coin-head/src/vrml97/Switch.cpp:179)
==31646==    by 0x406FD43D: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:447)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40766885: SoBaseList::truncate(int) (/home/sigma/mortene/code/coin-head/src/lists/SoBaseList.cpp:141)
==31646==    by 0x40701316: SoChildList::truncate(int) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:176)
==31646==    by 0x40701B95: SoChildList::~SoChildList(void) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:84)
==31646==    by 0x407E803D: SoGroup::~SoGroup(void) (/home/sigma/mortene/code/coin-head/src/nodes/SoGroup.cpp:239)
==31646==    by 0x40949F72: SoVRMLParent::~SoVRMLParent(void) (/home/sigma/mortene/code/coin-head/src/vrml97/Parent.cpp:165)
==31646==    by 0x4093CAD1: SoVRMLGroup::~SoVRMLGroup(void) (/home/sigma/mortene/code/coin-head/src/vrml97/Group.cpp:291)
==31646==    by 0x406FD43D: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:447)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40766885: SoBaseList::truncate(int) (/home/sigma/mortene/code/coin-head/src/lists/SoBaseList.cpp:141)
==31646==    by 0x40701316: SoChildList::truncate(int) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:176)
==31646==    Address 0x45436CEC is 44 bytes inside a block of size 48 free'd
==31646==    at 0x4002CDAD: __builtin_delete (vg_replace_malloc.c:244)
==31646==    by 0x408BD037: SoSensor::~SoSensor(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoSensor.cpp:128)
==31646==    by 0x408BC3A7: SoDelayQueueSensor::~SoDelayQueueSensor(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoDelayQueueSensor.cpp:80)
==31646==    by 0x408BC1B2: SoDataSensor::~SoDataSensor(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoDataSensor.cpp:87)
==31646==    by 0x408BC9B5: SoNodeSensor::~SoNodeSensor(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoNodeSensor.cpp:66)
==31646==    by 0x40747E0F: SoProtoInstance::~SoProtoInstance(void) (/home/sigma/mortene/code/coin-head/src/misc/SoProtoInstance.cpp:130)
==31646==    by 0x406FD43D: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:447)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40748372: SoProtoInstance::sensorCB(void *, SoSensor *) (/home/sigma/mortene/code/coin-head/src/misc/SoProtoInstance.cpp:271)
==31646==    by 0x408BC285: SoDataSensor::invokeDeleteCallback(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoDataSensor.cpp:248)
==31646==    by 0x408BCA7F: SoNodeSensor::dyingReference(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoNodeSensor.cpp:121)
==31646==    by 0x406FD3E7: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:436)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40766885: SoBaseList::truncate(int) (/home/sigma/mortene/code/coin-head/src/lists/SoBaseList.cpp:141)
==31646==    by 0x40701316: SoChildList::truncate(int) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:176)
==31646==    by 0x40701B95: SoChildList::~SoChildList(void) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:84)
==31646==    by 0x407E803D: SoGroup::~SoGroup(void) (/home/sigma/mortene/code/coin-head/src/nodes/SoGroup.cpp:239)
==31646==    by 0x40954FD6: SoVRMLSwitch::~SoVRMLSwitch(void) (/home/sigma/mortene/code/coin-head/src/vrml97/Switch.cpp:179)
==31646==    by 0x406FD43D: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:447)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==

and

==31646== Invalid write of size 4
==31646==    at 0x408BCA96: SoNodeSensor::dyingReference(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoNodeSensor.cpp:103)
==31646==    by 0x406FD3E7: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:436)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40766885: SoBaseList::truncate(int) (/home/sigma/mortene/code/coin-head/src/lists/SoBaseList.cpp:141)
==31646==    by 0x40701316: SoChildList::truncate(int) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:176)
==31646==    by 0x40701B95: SoChildList::~SoChildList(void) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:84)
==31646==    by 0x407E803D: SoGroup::~SoGroup(void) (/home/sigma/mortene/code/coin-head/src/nodes/SoGroup.cpp:239)
==31646==    by 0x40954FD6: SoVRMLSwitch::~SoVRMLSwitch(void) (/home/sigma/mortene/code/coin-head/src/vrml97/Switch.cpp:179)
==31646==    by 0x406FD43D: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:447)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40766885: SoBaseList::truncate(int) (/home/sigma/mortene/code/coin-head/src/lists/SoBaseList.cpp:141)
==31646==    by 0x40701316: SoChildList::truncate(int) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:176)
==31646==    by 0x40701B95: SoChildList::~SoChildList(void) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:84)
==31646==    by 0x407E803D: SoGroup::~SoGroup(void) (/home/sigma/mortene/code/coin-head/src/nodes/SoGroup.cpp:239)
==31646==    by 0x40949F72: SoVRMLParent::~SoVRMLParent(void) (/home/sigma/mortene/code/coin-head/src/vrml97/Parent.cpp:165)
==31646==    by 0x4093CAD1: SoVRMLGroup::~SoVRMLGroup(void) (/home/sigma/mortene/code/coin-head/src/vrml97/Group.cpp:291)
==31646==    by 0x406FD43D: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:447)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40766885: SoBaseList::truncate(int) (/home/sigma/mortene/code/coin-head/src/lists/SoBaseList.cpp:141)
==31646==    by 0x40701316: SoChildList::truncate(int) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:176)
==31646==    Address 0x45436CEC is 44 bytes inside a block of size 48 free'd
==31646==    at 0x4002CDAD: __builtin_delete (vg_replace_malloc.c:244)
==31646==    by 0x408BD037: SoSensor::~SoSensor(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoSensor.cpp:128)
==31646==    by 0x408BC3A7: SoDelayQueueSensor::~SoDelayQueueSensor(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoDelayQueueSensor.cpp:80)
==31646==    by 0x408BC1B2: SoDataSensor::~SoDataSensor(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoDataSensor.cpp:87)
==31646==    by 0x408BC9B5: SoNodeSensor::~SoNodeSensor(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoNodeSensor.cpp:66)
==31646==    by 0x40747E0F: SoProtoInstance::~SoProtoInstance(void) (/home/sigma/mortene/code/coin-head/src/misc/SoProtoInstance.cpp:130)
==31646==    by 0x406FD43D: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:447)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40748372: SoProtoInstance::sensorCB(void *, SoSensor *) (/home/sigma/mortene/code/coin-head/src/misc/SoProtoInstance.cpp:271)
==31646==    by 0x408BC285: SoDataSensor::invokeDeleteCallback(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoDataSensor.cpp:248)
==31646==    by 0x408BCA7F: SoNodeSensor::dyingReference(void) (/home/sigma/mortene/code/coin-head/src/sensors/SoNodeSensor.cpp:121)
==31646==    by 0x406FD3E7: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:436)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==    by 0x40766885: SoBaseList::truncate(int) (/home/sigma/mortene/code/coin-head/src/lists/SoBaseList.cpp:141)
==31646==    by 0x40701316: SoChildList::truncate(int) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:176)
==31646==    by 0x40701B95: SoChildList::~SoChildList(void) (/home/sigma/mortene/code/coin-head/src/misc/SoChildList.cpp:84)
==31646==    by 0x407E803D: SoGroup::~SoGroup(void) (/home/sigma/mortene/code/coin-head/src/nodes/SoGroup.cpp:239)
==31646==    by 0x40954FD6: SoVRMLSwitch::~SoVRMLSwitch(void) (/home/sigma/mortene/code/coin-head/src/vrml97/Switch.cpp:179)
==31646==    by 0x406FD43D: SoBase::destroy(void) (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:447)
==31646==    by 0x406FD99F: SoBase::unref(void) const (/home/sigma/mortene/code/coin-head/src/misc/SoBase.cpp:634)
==31646==
----8<----- [snip] --------------------8<----- [snip] ----------------

    20041105 mortene.

    UPDATE 20050520 mortene: this bug may be fixed, at least I
    couldn't reproduce it now (with Valgrind 2.4.0.rc4).

=====================================================================

191 SoOffscreenRenderer bug with SoOrthographicCamera in multitile-mode.

    The following iv-file results in a faulty image when fed to the
    Ivtools/renderos/renderos example, using a tile size of 512x512:

----8<-------------- [snip] -------8<-------------- [snip] -------------
#Inventor V2.1 ascii

OrthographicCamera {
   viewportMapping CROP_VIEWPORT_LINE_FRAME
   nearDistance 0.1
   farDistance 53000
   height 6393.33333333333
   position -15206.092773 9440.072266 17456.371094
   orientation -0.536894 -0.843216 0.027043   0.798199
}

Cube {
   width 500
   height 100
   depth 50000
}
----8<-------------- [snip] -------8<-------------- [snip] -------------

    To reproduce with a forced tile size of 512x512, set the following
    envvars:

        COIN_OFFSCREENRENDERER_TILEWIDTH=512
        COIN_OFFSCREENRENDERER_TILEHEIGHT=512

    It appears to have something to do with the SoOffscreenRenderer
    camera view volume narrowing to not work correctly with the
    cropped viewport, as commenting out the "viewportMapping" field of
    the camera causes the bug to go away.

    20050519 mortene.

    Yes. SoCamera will render a frame in the viewport when the
    viewport mapping is set to one of the CROP modes. To fix this we
    need a new element called SoViewportTile or something to make it
    possible for SoCamera to detect that the SoGLRenderAction is
    actually rendering a tile in the viewport, and not the entire
    viewport. This is also needed to make other viewport-based
    rendering nodes, such as legend, work.

    I don't know if SoViewportTileElement is a good idea though, since
    it will complicate the CROP rendering code in SoCamera quite a
    bit.

    Another idea: Is it possible to use the WindowSize-parameter in
    SbViewportRegion to specify the entire viewport? This must be done
    _after_ the viewport is initialized in the GL context though, maybe
    in a pre-camera callback or something.

    20050519 pederb

    UPDATE 20050520 mortene: after discussion with pederb, we consider
    the Correct Fix for this to first take care of bug item #121. This
    means the SoOffscreenRenderer adding an element to the state stack
    upon traversal start which identifies the current tile to be
    rendered (i.e. containing row and column index, and ditto
    dimensions), and then leaving it to SoCamera to render the
    viewport cropping correctly. (As a side effect, SoCamera could
    then also do the view volume narrowing itself, removing the need
    for the current ugly hack in SoOffscreenRenderer to do this.)

    UPDATE 20050714 mortene: there's something else odd going on
    here... if the tilewidth and -height is set to 640 and 480,
    respectively, which is the same as the default size of renderos'
    buffer, it also comes out wrong -- but only if tiled rendering is
    *not* used. If we on the other hand force tiled rendering on (by
    using the envvar COIN_FORCE_TILED_OFFSCREENRENDERING=1), and the
    tile size is at least of dimensions 640x480, everything works..?
    Strange.

=====================================================================

194 SoOffscreenRenderer in tile-mode asserts on VRML97 scene graph.

    The following scene graph attempted drawn with an
    SoOffscreenRenderer (use e.g. the 'osgrab' debugging backdoor in
    an examiner viewer to reproduce) causes an assert to hit in
    SoOffscreenRenderer, due to the tail of the
    SoGLRenderAction::getCurPath() being NULL at one point. The tail
    of the path should likely never become NULL.

----8<-------------- [snip] -------8<-------------- [snip] -------------
#VRML V2.0 utf8

Group {
   children [
      Shape {
         geometry Sphere { }
      }
   ]
}
----8<-------------- [snip] -------8<-------------- [snip] -------------

    20050606 mortene.

=====================================================================

196 Bug with VRML97 scene graph searching.

    The below scene graph (with a VRML97 PROTO) demonstrates a problem
    with SoSearchAction in VRML97 graphs: one can not get paths to all
    nodes, which causes e.g. PROTO-connection to "miss out" on certain
    connections it should be able to make.

----8<-------------- [snip] -------8<-------------- [snip] -------------
#VRML V2.0 utf8

PROTO AnnText [
 exposedField MFString sText ""
 field SFFloat sFontSize 0.1
 field MFString sFontFamily "courier"
 field SFString sFontStyle ""
 exposedField SFVec3f slocation 0 0 0
]
{
   Transform {
      translation IS slocation
      children [
         Shape {
            geometry Text{
               string IS sText
               fontStyle FontStyle {
                  size IS sFontSize
                  family IS sFontFamily
                  style IS sFontStyle
                  spacing 2.5
               }
            }
         }
      ]
   }
}

Transform {
   children [
      AnnText { slocation 0 0 0 sText "hepp" }
   ]
}
----8<-------------- [snip] -------8<-------------- [snip] -------------

    pederb suggests that this should be fix be extending the
    SoVRMLParent etc hack by duplicating MFNode contents into "hidden"
    SoChildList instances. pederb has a framework for this in:

        void SoVRMLParent::updateChildList(SoNode * nodewithsfnode, SoChildList & cl)

    Note: is there some other, more fundamental, way we could fix this
    without applying this hack?  It's not very nice, and is giving us
    headaches now and then with bugs which are the results of these
    SoChildList instances not kept in sync.

    20050609 mortene.

    UPDATE 20050711 mortene: see also item #162, which seems to be the
    same issue.

    UPDATE 20060130 mortene:

      This is a /really/ tricky issue, and one which we don't really
      know yet how to fix:

      The source of the problem is that VRML nodes use SoSFNode and
      SoMFNode *fields* to keep track of their child nodes, while the
      "traditional" Inventor nodes use the SoChildList class.

      For certain functionality of the library, we do however need a
      SoChildList -- like for e.g. SoPath-instances being able to
      track their way down through the scene graph.

      So to work around this problem, there is a hack in place at
      various spots in the VRML nodes with MFNode "children" field(s):
      the node internally keeps an SoChildList, which is attempted
      kept synchronized with the MFNode field. This SoChildList
      instance is then available through the virtual
      SoNode::getChildren() method, just as for the "normal" Inventor
      nodes.

      This hack does however have certain ugly weaknesses:

        * it has not been set up everywhere it is probably needed
          (none or few of the nodes with SoSFNode fields have an
          SoChildList "on the side", for instance)

        * it causes various problems with the SoPath class, which has
          no knowledge of SoSFNode / SoMFNode, and is completely and
          utterly geared towards going through SoChildList

        * there is copy'n'paste-jobs of code for
          SoChildList<->SoMFNode synchronization at various places --
          code which is not straightforward to bug-fix and maintain

        * the SoNode::getChildren() method returns a mutable
          SoChildList (!)

      So, this is one of the really major design flaws of the library,
      and it looks like we need a stroke of genius to resolve this
      knot.

      One venue which may be interesting to investigate would be to
      see if letting SoChildList take care of the synchronization vs
      SoSFNode & SoMFNode fields would be sensible. Or perhaps let the
      SoMFNode and SoSFNode classes have a built-in SoChildList which
      they could auto-sync?

      Another option which may be a possibility would be to let SoPath
      know about SoMFNode and SoSFNode. My gut feeling about that one
      is that it would quickly make the SoPath API a lot uglier, tho'.

      I've been asked for an estimate on this task, but given that we
      don't yet know /how/ to fix it, makes this impossible as of
      now. I'm quite certain however that we are talking about > 1
      week of efficient coding and testing.

=====================================================================

203 openal.dll instances may have different calling convention.

    Different than the custom under Windows, and different than from
    what we assume it is. This causes a crash on those platforms when
    attempts are made at using sound. (The first attempt at calling an
    openal.dll function will crash the application, because of
    unexpected stack layout on return.)

    My suggestion for a "best fix" would be to include code in Coin to
    detect the calling convention used for a DLL. To do this, one
    needs to do some trickery with checking the stack pointer after
    calling a function of the DLL -- something which probably requires
    some x86 assembly programming.

    This is still not a perfect fix, as functions in a DLL can have
    different calling conventions. Perhaps it would be possible to
    query the DLL directly about this? Request for assistance at a
    Usenet Windows programming group did not result in any useful
    hints for how to do this.

    20050701 mortene.

=====================================================================

205 Hardcopy export does not work correctly with viewport region
    changes.

    As reported on coin-discuss by user Ole Jacob Hagen:

        > Generating a postscript file when my scenegraph consists of
        > one instance of Smallchange's [ViewportRegion] works, but
        > when I have arranged my scenegraphs into four viewports
        > (2x2-array) it doesn't work.

    Is this a bug or a limitation? Waiting for pederb's input on this.

    20050714 mortene.

=====================================================================

208 Too high tessellation by default for 3D text?

    The 06.3.Complex3DText example updates at a very slow framerate,
    so it seems like our default settings for 3D text complexity is
    too high.

    Investigate, compare with SGI Inventor.

    Not particularly high priority item.

    20051021 mortene.

=====================================================================

209 Bug with SoTranslate1Dragger.

    The following iv-file should work the same in Coin as in SGI
    Inventor, but it doesn't:

----8<-------------- [snip] -------8<-------------- [snip] -------------
#Inventor V2.1 ascii

Switch { DEF DRAG Translate1Dragger { } }

RotationXYZ {
   angle  = DecomposeVec3f { vector = USE DRAG.translation }.x
   axis Z
}

USE DRAG

Cube { width 0.5 height 0.5 depth 0.5 }
Separator {
   Translation { translation 0 -5 0 }
   Cylinder { radius 0.2 height 10 }
}
Translation { translation 0 -10 0 }
Sphere { radius 0.4 }
----8<-------------- [snip] -------8<-------------- [snip] -------------

    In Coin, the dragger geometry is translated, for some odd
    reason. In SGI Inventor, it only rotates along with the rest of
    the geometry, as expected.

    20051208 mortene.

    UPDATE 20060130 mortene:

      As for an estimate on this bug, I expect it to be on the order
      of just a few hours to fix. Need to investigate and figure out
      why it happens, and the fix is likely to just be a little
      modification of the internal default geometry for the dragger,
      i.e. to the Coin/data/draggerDefaults/translate1Dragger.iv file.

=====================================================================

210 SoText2 font size policy in Coin incompatible with SGI / TGS
    Inventor?

    According to Frank Hibbeln:

> The doc's for SoFont indicate the size for Text2 nodes is
> interpreted as pixels whereas the TGS Inventor doc's indicate it's
> "printer points" (ie i point = 1/72 inch ).

    Should investigate this with SGI Inventor, and see if we need to
    fix anything for Coin.

    20051212 mortene.

=====================================================================

211 SoTextureCoordinateSphere et al does not work with some
    primitives.

    SoSphere, SoCube, SoCone and SoCylinder has not been modified to
    work properly with SoTextureCoordinateSphere,
    SoTextureCoordinateCylinder etc, but instead just ignoring this
    texture function nodes.

    20060213 mortene.

=====================================================================

213 SbTesselator fails on polygon which it is supposed to handle.

    The following Inventor file demonstrates a problem with
    SbTesselator (according to kintel, it is supposed to be able to
    handle this case):

----8<-------------- [snip] -------8<-------------- [snip] -------------
#Inventor V2.1 ascii

Separator {

   ShapeHints {
      faceType UNKNOWN_FACE_TYPE

   }

   FaceSet {
      vertexProperty
      VertexProperty {
         vertex [5616.34650  -94900.63137  0,
          19459.61213 -83978.42044  0,
          18951.60431 -69500.14309  0,
          34699.90509 -70643.16262  0,
          35080.91681 -85883.45559  0,
          21491.65119 -91598.56887  0,
          19459.61213 -83978.42044  0,
          5616.34650  -94900.63137  0,
          9172.41681  -108997.90091 0,
          52607.25275 -107600.87747 0,
          55909.31525 -90328.54153  0,
          46257.12775 -61371.98684  0,
          20475.63556 -53243.82669  0,
          282.24494   -65690.06887  0,
          5616.34650  -94900.63137  0]
         orderedRGBA 0xac00ff
         materialBinding OVERALL

      }
      numVertices 15

   }
}
----8<-------------- [snip] -------8<-------------- [snip] -------------

    Example provided by Frank Hibbeln.

    To see how this polygon is supposed to be tessellated, use the GLU
    tessellator by setting the following envvar:

        COIN_PREFER_GLU_TESSELLATOR=1

    20060216 mortene.

=====================================================================

214 VMLScript JS support does not work at all with latest Spidermonkey?

    We've had an external report about the VRMLScript node no working
    at all with JS, not even the models/vrml97/script/simplest-possible.wrl
    test case.

    The errors output from the code is

        Coin info in SoJavaScriptEngine::field2jsval(): no handler found for MFString
        Coin info in SoJavaScriptEngine::field2jsval(): no handler found for SFBool
        [and more of the same]

    There may be bugs in the Script.cpp, JS_VRMLClasses.cpp or
    SoJavaScriptEngine.cpp code which manifests itself with the latest
    Spidermonkey release, and not with previous versions (I'm not able
    to reproduce the errors on my system with the earlier version
    1.5rc6a).

    20060217 mortene.

=====================================================================

215 SoText2 behaviour under the influence of an SoClipPlane is
    peculiar.

    The following scene graph demonstrates how SoText2 nodes doesn't
    work well with SoClipPlane:

----8<-------------- [snip] -------8<-------------- [snip] -------------
#Inventor V2.1 ascii

ClipPlane {
   plane 1 0 0 -1
   on TRUE
}

Text2 {
   string "hepp"
}
----8<-------------- [snip] -------8<-------------- [snip] -------------

    Spin the camera around, and see how the SoText2 string slips in
    and out of view, sometimes partially, sometimes fully.

    Should perhaps try to implement a more consistent behaviour, for
    instance always showing or not showing the string depending on how
    the its 3D origin point is placed vs the clip plane(s).

    We need to make a policy decision on this, and document that, at
    least.

    20060307 mortene.

    This happens since we do culling against the view volume and the
    clipping plane based on the bounding box of the node and each
    character. I'm not sure if we should do culling against clipping
    planes for 2D texts or not. Should investigate how SGI Inventor
    handles this.

    20060327 pederb

=====================================================================

217 SoDrawStyle::POINTS not handled for linesets in SoToVRML2Action

    If a scene contains a line set where the draw style is set to
    POINTS, the lines will still be converted to lines and not points.

    Needs a bit of code to fix so I'll probably delay this until the
    next Coin bugfixing session.

    20060327 pederb

=====================================================================

219 View All doesn't work properly with SoText2 nodes

Pressing "View All" in an examinerviewer with the attached scenegraph
moves the "ZZZ" string outside the frustum and clips the "YYY" string
in half.
The initial view when setting the scenegraph looks correct though.

----8<-------------- [snip] -------8<-------------- [snip] -------------
#Inventor V2.1 ascii

SoText2 {
  string "XXX"
}
Translation {
  translation 0 100 0
}
SoText2 {
  string "YYY"
}
Translation {
  translation 0 100 0
}
SoText2 {
  string "ZZZ"
}
----8<-------------- [snip] -------8<-------------- [snip] -------------

    20061129 kintel

=====================================================================

221 Global field values are not evaluated when reading from file

When creating and connecting from a global field from an .iv file, the
value of the field is not evaluated until the field is touched. If
nobody touches the file from code, the initial value will never be
set.

The offending code is probably related to the "do not notify" flag
being set when making the field connection (SoField.cpp:2283).

----8<-------------- [snip] -------8<-------------- [snip] -------------
#Inventor V2.1 ascii

SoText2 {
  string "XXX" = GlobalField {
    type SFString
    vfctime "Test"
  }.vfctime
}
----8<-------------- [snip] -------8<-------------- [snip] -------------

    20061129 kintel

=====================================================================