From: Tom Lane <tgl@sss.pgh.pa.us>
To: "Higuchi, Daisuke" <higuchi.daisuke@jp.fujitsu.com>
cc: Michael Paquier <michael.paquier@gmail.com>,
        "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Subject: Re: [Bug fix]If recovery.conf has target_session_attrs=read-write,
 the standby fails to start.
In-reply-to: <1803D792815FC24D871C00D17AE95905B1A4CA@g01jpexmbkw24>
References: <1803D792815FC24D871C00D17AE95905B1A34A@g01jpexmbkw24>
 <CAB7nPqQafwnsKAZ9sv9D8wE_3HU6gge6L0QuPXp9yK=4-2oC1A@mail.gmail.com>
 <1803D792815FC24D871C00D17AE95905B1A4CA@g01jpexmbkw24>
Comments: In-reply-to "Higuchi, Daisuke" <higuchi.daisuke@jp.fujitsu.com>
	message dated "Fri, 19 May 2017 07:05:30 -0000"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <24402.1495225931.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 19 May 2017 16:32:11 -0400
Message-ID: <24403.1495225931@sss.pgh.pa.us>
Precedence: bulk
Sender: pgsql-hackers-owner@postgresql.org

"Higuchi, Daisuke" <higuchi.daisuke@jp.fujitsu.com> writes:
> By adding flex option '-i', replication command parser could be more flex=
ible.=20
> This option is already used for syncrep_scanner.c, so it is not strange t=
o add for repl_scanner.c too.=20

Really?  That wasn't an especially bright idea IMO, because it means that
the scanner's behavior will be dependent on the locale flex was run in.
An example here is that Turkish locale is likely to have a different
idea of what "FIRST" matches than other locales do.  Indeed, if I run
the flex build in a non-C locale, I get warnings like

$ LANG=3Den_US /usr/bin/flex -b -CF -p -i -o'syncrep_scanner.c' syncrep_sca=
nner.l
syncrep_scanner.l:91: warning, the character range [\200-\377] is ambiguous=
 in a case-insensitive scanner
syncrep_scanner.l:91: warning, the character range [\200-\377] is ambiguous=
 in a case-insensitive scanner

We'd probably be better off to implement case-insensitivity the hard way.
There is a reason why none of our other flex scanners use this switch.

While I'm whining ... it looks like the other flex options selected here
were cargo-culted in rather than being thought about.  Surely we don't
run syncrep_scanner often enough, nor over so much data, that it's a
good tradeoff to use the options for larger tables.

			regards, tom lane


--=20
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers