>This is because DDL will go through ProcessUtility hook. I don't think >pg_stat_statements should be calling GetCurrentStatementStartTimestamp() >as you have it in v5. We should be setting this the start timestamp in >PlannedStmt during the ProcessUtility hook ( same pattern as ExecutorStart ). >see [1].
Oh, yeah! Indeed! Thanks! > > >> FWIW, I remain opposed to adding *any* additional thing under the spinlock. >> We >> first need to fix the design of pgss, then we can start discussing expanding >> it even further. > >I do think having the execution start time is valuable, but I will do >agree that it should >wait until after the pg_stat_statements redesign is complete which is >being discussed >here [2]. No problem! I completely agree with this. I'm fine to wait and then rebase it on Sami's patch. > > >-- >Sami Imseih >Amazon Web Services (AWS) > > >[1] >https://www.postgresql.org/message-id/CAA5RZ0uLbF_zcu64-K50fepq20s7GEYCsnizVBhm6eMmsfSa7Q%40mail.gmail.com >[2] >https://www.postgresql.org/message-id/flat/CAA5RZ0sQ%2BgDn-J85j1FzOdL1YjVYRegpmQpDiah1%3DREWZSZj%2BQ%40mail.gmail.com#52395851f3df4b02c9ecb3430a7bfc37
