Re: pretty_bool in pg_get_constraintdef has no effect since pg >= 9
От | eli.mach@mailbox.org |
---|---|
Тема | Re: pretty_bool in pg_get_constraintdef has no effect since pg >= 9 |
Дата | |
Msg-id | 248997790.14326.1582639167881@office.mailbox.org обсуждение исходный текст |
Ответ на | Re: pretty_bool in pg_get_constraintdef has no effect since pg >= 9 (Daniel Gustafsson <daniel@yesql.se>) |
Список | pgsql-bugs |
The SA-Version I use is the newest version 1.3.13. Seems that the SA regex never had to deal with newlines. I think thatcan be easily fixed by adding a "re.DOTALL" in the SA regex. I open an SA-Issue. Thanks for your help. Greetings, elim. > On 25 February 2020 14:30 Daniel Gustafsson <daniel@yesql.se> wrote: > > > > On 25 Feb 2020, at 13:56, eli.mach@mailbox.org wrote: > > > since postgres 9, `pg_get_constraintdef(cons.oid, pretty_bool)` does not work as expected. The result is always in "pretty"-format(with newlines), regardless of whether `pretty_bool` is true or false. Calling `pg_get_constraintdef(constraint_oid)`without `pretty_bool`, also returns "pretty"-format. > > This is not a bug, but a deliberate change which was made in 62e666400d back in > 2013, the argument being that changes in whitespace should not affect forward > compatibility. > > > I'm migrating from postgres 8 to 11 and a sqlalchemy script throws a warning "SAWarning: Could not parse CHECK constrainttext" because there a no newlines expected in `re.match(r"^CHECK *\((.+)\)( NOT VALID)?$", src)`. > > Surely SA has been updated to work with more recent version of postgres? I've > not used SA myself, but are you sure you are using the right version of the > tool? > > cheers ./daniel
В списке pgsql-bugs по дате отправления: