From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | Josef Šimánek <josef(dot)simanek(at)gmail(dot)com> |
Cc: | Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>, Peter Eisentraut <peter(dot)eisentraut(at)enterprisedb(dot)com>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: Add Boolean node |
Date: | 2021-12-28 15:51:36 |
Message-ID: | 1759529.1640706696@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
=?UTF-8?B?Sm9zZWYgxaBpbcOhbmVr?= <josef(dot)simanek(at)gmail(dot)com> writes:
> po 27. 12. 2021 v 16:10 odesílatel Alvaro Herrera
> <alvherre(at)alvh(dot)no-ip(dot)org> napsal:
>> Hmm, interesting side-effect: we no longer assign a column name in this
>> case so it remains "?column?", just like it happens for other datatypes.
>> This seems okay to me. (This is also what causes the changes in the
>> isolationtester expected output.)
> This seems to be caused by a change of makeBoolAConst function. I was
> thinking for a while about the potential backward compatibility
> problems, but I wasn't able to find any.
In theory this could break some application that's expecting
"SELECT ..., true, ..." to return a column name of "bool"
rather than "?column?". The risk of that being a problem in
practice seems rather low, though. It certainly seems like a
wart that you get a type name for that but not for other sorts
of literals such as 1 or 2.4, so I'm okay with the change.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Peter Eisentraut | 2021-12-28 16:22:04 | parse/analyze API refactoring |
Previous Message | Jelte Fennema | 2021-12-28 15:49:00 | Re: [EXTERNAL] Re: PQcancel does not use tcp_user_timeout, connect_timeout and keepalive settings |