Amit Langote <amitlangote09@gmail.com> writes:
> Whatever mechanism we will use would still need to involve setting a
> global Snapshot variable though, right?
In v14 we'll certainly still be passing the snapshot(s) to SPI, which will
eventually make the snapshot active. With your patch, since we're just
handing the snapshot to the scan mechanism, it seems at least
theoretically possible that we'd not have to do PushActiveSnapshot on it.
Not doing so might be a bad idea however; if there is any user-defined
code getting called, it might have expectations about ActiveSnapshot being
relevant. On the whole I'd be inclined to say that we still want the
RI test_snapshot to be the ActiveSnapshot while performing the test.
regards, tom lane