| From: | Dominique Devienne <ddevienne(at)gmail(dot)com> |
|---|---|
| To: | pgsql-general(at)lists(dot)postgresql(dot)org |
| Subject: | Two test failures on v16 (compared to v14) |
| Date: | 2024-02-14 10:40:47 |
| Message-ID: | CAFCRh-8NDFEhyEibO=4fpcLzJXh+s4c575+abnVOTC2C=B3giw@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-general |
Just an FYI. Running the same test suite against V16 triggered two failures.
First, a test of getting the direct members of a role failed.
Returned one more role. The role that created the introspected role.
That's the new CREATEROLE semantic, which adds the role creator
as a direct member of the created role, with admin option.
Second, a LO test failed to open a given OID.
The test starts a read-only transaction, then tries a loOpen(oid,
INV_WRITE),
but does only a loRead(). Now in v16, the loOpen fails, returning -1, while
it used to be OK.
So this must be a bug fix of some kind I guess. Using INV_READ fixed the
test.
I was not expecting failures, but both have good reasons and are easily
fixed. FWIW. --DD
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Simon Connah | 2024-02-14 11:18:33 | Re: PostgreSQL DB in prod, test, debug |
| Previous Message | Erik Wienhold | 2024-02-14 10:38:20 | Re: PostgreSQL DB in prod, test, debug |