Re: BUG #18292: Unexpected error: "relation "hobbies_r" does not exist" caused by user-defined functions
От | Tom Lane |
---|---|
Тема | Re: BUG #18292: Unexpected error: "relation "hobbies_r" does not exist" caused by user-defined functions |
Дата | |
Msg-id | 1367517.1705287423@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | BUG #18292: Unexpected error: "relation "hobbies_r" does not exist" caused by user-defined functions (PG Bug reporting form <noreply@postgresql.org>) |
Ответы |
Re: BUG #18292: Unexpected error: "relation "hobbies_r" does not exist" caused by user-defined functions
|
Список | pgsql-bugs |
PG Bug reporting form <noreply@postgresql.org> writes: > My fuzzer finds a suspicious alarm in Postgres 17devel: "ERROR: relation > "hobbies_r" does not exist". I report it as I think it might be an > unexpected error. I don't see anything unexpected here: you made a transaction-local change in search_path that renders that table invisible. It's not instantly obvious how the flow of control gets to a lookup of that table after the SET LOCAL; but evidently that's happening, and I don't feel any urge to work out the details. (Note that if you were expecting SET LOCAL to mean "local to this function call", you're mistaken. You can get that effect with a SET clause attached to the function definition; but this is not that.) If you want people to take this sort of report seriously, you need to analyze the behavior yourself, not expect us to look for a bug that probably doesn't exist. regards, tom lane
В списке pgsql-bugs по дате отправления: