From: Tom Lane <tgl@sss.pgh.pa.us>
To: Erik Wienhold <ewie@ewie.name>
cc: Roman =?utf-8?Q?Fro=C5=82ow?= <rofrol@gmail.com>,
        pgsql-docs@lists.postgresql.org
Subject: Re: T is a mandatory date time separator in RFC3339 but documentation
 states differently
In-reply-to: <7njdkzeco47heg6tglzsm5wmmbb56vd66f7pta6gazb2k52viu@av3wruqnylpp>
References: <169974270449.398198.1438250797414790599@wrigleys.postgresql.org>
 <fojeaxd2hqux2abmft22nnsll435q2qfmecjlsig2g2th5u2bw@cr3m7rzgaop3>
 <CAFZ1f-9rnK6uZwsUUKDwCkNB9h9zP70oaSfduu+TxrQAxEmwjQ@mail.gmail.com>
 <CAFZ1f-9D3iCcGF8+mLgRffUp6i7m-2mh1GJ5UZbfcZiyjaXQzQ@mail.gmail.com>
 <piavtdd7mhmkpzpgvxaek3hz3e2kan3c2fitn5iqta6nyrpgyl@txongxshxkxw>
 <7njdkzeco47heg6tglzsm5wmmbb56vd66f7pta6gazb2k52viu@av3wruqnylpp>
Comments: In-reply-to Erik Wienhold <ewie@ewie.name>
	message dated "Mon, 13 Nov 2023 15:29:35 +0100"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <2037452.1699892596.1@sss.pgh.pa.us>
Date: Mon, 13 Nov 2023 11:23:16 -0500
Message-ID: <2037453.1699892596@sss.pgh.pa.us>
Archived-At: 
 <https://www.postgresql.org/message-id/2037453.1699892596%40sss.pgh.pa.us>
Precedence: bulk

Erik Wienhold <ewie@ewie.name> writes:
> On 2023-11-13 15:24 +0100, Erik Wienhold wrote:
>> I also noticed that when people say "ISO 8601" they usually mean RFC
>> 3389 or some subset of ISO 8601.

> Forgot this fine visualization of the differences:
> https://ijmacd.github.io/rfc3339-iso8601/

I'm inclined not to change anything here, for a couple of reasons:

1. PG accepts a fairly large number of ISO 8601 variants (not all);
not only the RFC 3339 format.  So s/ISO 8601/RFC 3339/g would be
incorrect.  Besides, I think more people know what ISO 8601 is than
know what RFC 3339 is, so that change would also be confusing.

2. If ijmacd's pretty graphic is correct, then what we say about
'T' versus space is correct, even if it isn't the whole truth.
I'm not quite sure that ijmacd is correct, though, because of
this bit in 3339:

      NOTE: Per [ABNF] and ISO8601, the "T" and "Z" characters in this
      syntax may alternatively be lower case "t" or "z" respectively.

which suggests that 8601 is also case-insensitive.  I don't plan
to go buy a copy of that spec to find out, though.  In any case,
we accept 'T', 't', '_', and most other punctuation there, so
we should be able to read nearly any plausible variant.

			regards, tom lane