From: Tom Lane <tgl@sss.pgh.pa.us>
To: Karina Litskevich <litskevichkarina@gmail.com>
cc: Pavel Borisov <pashkin.elfe@gmail.com>,
        Postgres hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: Invalid "trailing junk" error message when non-English letters are used
In-reply-to: <CACiT8iaMW47nEzXSAiBNVaoapw7v0FSQ2j_TWi8ecWqynF2ngg@mail.gmail.com>
References: <CACiT8iZ_diop=0zJ7zuY3BXegJpkKK1Av-PU7xh0EDYHsa5+=g@mail.gmail.com> <CALT9ZEFG8u=+pBMkON1Ske+We6wtjf=A2SYGvhsZJn5TaHLwLA@mail.gmail.com> <CACiT8iZb+o_=MGJ72qM8HOuam0-hv09bj8-9x-XsODH1PrwP_A@mail.gmail.com> <1709057.1725490341@sss.pgh.pa.us> <CACiT8iaf-ONO6sF0nC-9ZMpO7jq0eWUkEQXuVznegi3agGzWxw@mail.gmail.com> <CACiT8iaMW47nEzXSAiBNVaoapw7v0FSQ2j_TWi8ecWqynF2ngg@mail.gmail.com>
Comments: In-reply-to Karina Litskevich <litskevichkarina@gmail.com>
	message dated "Thu, 05 Sep 2024 18:11:20 +0300"
MIME-Version: 1.0
Content-Type: text/plain; charset="UTF-8"
Content-ID: <1998875.1725554982.1@sss.pgh.pa.us>
Content-Transfer-Encoding: 8bit
Date: Thu, 05 Sep 2024 12:49:42 -0400
Message-ID: <1998877.1725554982@sss.pgh.pa.us>
Archived-At: <https://www.postgresql.org/message-id/1998877.1725554982%40sss.pgh.pa.us>
Precedence: bulk

Karina Litskevich <litskevichkarina@gmail.com> writes:
> On Thu, Sep 5, 2024 at 6:07 PM Karina Litskevich
> <litskevichkarina@gmail.com> wrote:
>> In v3 of the patch I grouped all the *_junk rules together and included
>> the suggested comment with a little added something.

> Oops, I forgot to attach the patch, here it is.

Pushed with a bit of further wordsmithing on the comment.

I left out the proposed new test case "SELECT 1ä;".  The trouble
with that is it'd introduce an encoding dependency into the test.
For example, it'd likely fail with some other error message in
a server encoding that lacks an equivalent to UTF8 "ä".  While
we have methods for coping with such cases, it requires some
pushups, and I didn't see the value.  The changes in existing
test case results are sufficient to show the patch does what
we want.

Also, while the bug exists in v15, the patch didn't apply at all.
I got lazy and just did the minimal s/ident_start/identifier/ change
in that branch, instead of back-patching the cosmetic aspects.

			regards, tom lane