Re: tripping an assert in 8.1.6 (more info)
| От | Brian Hurt |
|---|---|
| Тема | Re: tripping an assert in 8.1.6 (more info) |
| Дата | |
| Msg-id | 45B659EF.6000703@janestcapital.com обсуждение исходный текст |
| Ответ на | tripping an assert in 8.1.6 (Brian Hurt <bhurt@janestcapital.com>) |
| Ответы |
Re: tripping an assert in 8.1.6 (more info)
Re: tripping an assert in 8.1.6 (more info) |
| Список | pgsql-hackers |
More info on that assert I've hit.
Compile 8.1.6 with configuration options:
./configure --with-perl --enable-debug --enable-cassert
(not sure if --perl is relevent or not, I think not).
This is on Fedora Core 5 on x86-32.
Execute, on a fresh database, the following sql, to recreate the bug:
CREATE TABLE foo
( some_data VARCHAR(32)
, row_date DATE
);
CREATE TABLE bar
( some_data VARCHAR(32)
, row_date DATE
);
CREATE TABLE quux
( some_data VARCHAR(32)
, more_data VARCHAR(32)
);
CREATE OR REPLACE VIEW bazz AS
SELECT ('bar: ' || bar.row_date) :: TEXT AS action, quux.more_data AS more_data, bar.row_date AS
row_date
FROM bar JOIN quux ON bar.some_data = quux.some_data
UNION ALL
SELECT ('foo: ' || foo.row_date) :: TEXT AS action, quux.more_data AS more_data, foo.row_date AS
row_date
FROM foo JOIN quux ON foo.some_data = quux.some_data
;
SELECT action, more_data FROM bazz;
В списке pgsql-hackers по дате отправления: