Richard Guo <guofenglinux@gmail.com> writes: > So what I'm thinking is that maybe we can add a new type of path, named > SampleScanPath, to have the TableSampleClause per path. Then we can > safely reparameterize the TableSampleClause as needed for each > SampleScanPath. That's what the attached patch does.
Alternatively, could we postpone the reparameterization until createplan.c? Having to build reparameterized expressions we might not use seems expensive, and it would also contribute to the memory bloat being complained of in nearby threads.
I did think about this option but figured out that it seems beyond the scope of just fixing SampleScan. But if we want to optimize the reparameterization mechanism besides fixing this crash, I think this option is much better. I drafted a patch as attached.
