Hi Dharin,
Thanks for addressing the FIXME in the alter_table regression test.
The change LGTM. Verifying both planner visibility via EXPLAIN and runtime
correctness with a SELECT after the index rename is a solid improvement,
and using the standard enable_seqscan/bitmapscan/indexonlyscan = off
pattern keeps the test deterministic.
I tested the patch locally and it passes:
make check TESTS=alter_table
full make check
O/P: ok 206 + alter_table 870 ms
I don’t see any issues with the approach.
Regards,
Vasuki M,
C-DAC,Chennai
>