Postgres Pro
Downloads
Home   >   mailing lists

Re: hashing bpchar for nondeterministic collations is broken - Mailing list pgsql-bugs

From Richard Guo
Subject Re: hashing bpchar for nondeterministic collations is broken
Date
Msg-id CAMbWs48dYCWaMaPxFCC4hDTf5ZOFX1xcDdAs4kuPq1b7CCTDAA@mail.gmail.com
Whole thread Raw
In response to hashing bpchar for nondeterministic collations is broken  (Jeff Davis <pgsql@j-davis.com>)
Responses Re: hashing bpchar for nondeterministic collations is broken
List pgsql-bugs
Tree view

On Fri, Dec 2, 2022 at 4:11 AM Jeff Davis <pgsql@j-davis.com> wrote:
CREATE COLLATION ctest_nondet(
  provider = icu, locale = '', deterministic = false);

=# select hashbpcharextended('a' collate ctest_nondet, 3);
 hashbpcharextended 
---------------------
 2792701603979490504
(1 row)

=# select hashbpcharextended('a  ' collate ctest_nondet, 3);
  hashbpcharextended 
----------------------
 -4885217598372536483
(1 row)
 
Good catch.  This really shouldn't happen for bpchar as the trailing
blanks are supposed to be ignored.  I can see hashbpchar is doing it
with the correct "true" length, so maybe this is a copy-and-pasteo from
hashtextextended?
 
Also, does someone have an opinion on backporting this? I'm inclined
to.
 
I'm not familiar with the backporting policy, but I also think we need
to as this fixes an obvious oversight.

Thanks
Richard

pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: INFORMATION_SCHEMA.routines column routine_definition does not show the source
Next
From: Tom Lane
Date:
Subject: Re: INFORMATION_SCHEMA.routines column routine_definition does not show the source