Re: Adding skip scan (including MDAM style range skip scan) to nbtree

Peter,

moving the conversation here from "pgsql: Improve nbtree skip scan primitive scan scheduling" thread on -committers.  Attached is another regression test for your consideration, which gives rise to the following assertion:

TRAP: failed Assert("numSkipArrayKeys == 0"), File: "nbtpreprocesskeys.c", Line: 1859, PID: 7210
0   postgres                            0x00000001032f30e0 ExceptionalCondition + 108
1   postgres                            0x0000000102e83100 _bt_preprocess_keys + 6036
2   postgres                            0x0000000102e87338 _bt_first + 168
3   postgres                            0x0000000102e84680 btgettuple + 196
4   postgres                            0x0000000102e75cdc index_getnext_tid + 68
5   postgres                            0x00000001030017a0 IndexOnlyNext + 228
6   postgres                            0x0000000102fe5b2c ExecScan + 228
7   postgres                            0x0000000103011088 ExecSort + 536
8   postgres                            0x0000000102fdbc68 standard_ExecutorRun + 312
9   postgres                            0x00000001031bdfb8 PortalRunSelect + 236
10  postgres                            0x00000001031bdbd4 PortalRun + 492
11  postgres                            0x00000001031bcb18 exec_simple_query + 1292
12  postgres                            0x00000001031b9d1c PostgresMain + 1388
13  postgres                            0x00000001031b59a8 BackendInitialize + 0
14  postgres                            0x000000010310edd8 postmaster_child_launch + 372
15  postgres                            0x000000010311303c ServerLoop + 4948
16  postgres                            0x0000000103111360 InitProcessGlobals + 0
17  postgres                            0x0000000103030c00 help + 0
18  dyld                                0x000000018eb17154 start + 2476

This looks sufficiently different from the assertion mentioned on the other thread to be worth posting here.