Subject: Re: PostgreSQL vs SQL/XML Standards
To: Pavel Stehule <pavel.stehule@gmail.com>
References: 
 <CAFj8pRDFQJMoKf-X+3p1ASwU02nnXfAXrg83Z3fHUBd-Gwuuxw@mail.gmail.com>
 <20181025194720.ylmw64xqlnjr5wja@alvherre.pgsql>
 <CAFj8pRB10LxRcYYKWb5omjtAQq_bEVWiQ9B6V=kRZYMvM0B-_w@mail.gmail.com>
 <CAFj8pRBr5tY5b=8gKptwOyuPf8f3fvyHbXNbSCU03guNh-johw@mail.gmail.com>
 <CAFj8pRA6J25CtAZ2TuRvxK3gat7-bBUYh0rfE2yM7Hj9GD14Dg@mail.gmail.com>
 <FB592C2F-E2F3-42FD-A342-EE41BECD3771@winand.at>
 <CAFj8pRC_bi4uzE3Z-XXuGnLFPP-7cZ-zr2JhGauWm3E8K2MqEg@mail.gmail.com>
 <45C74339-3531-46BD-A595-37183C6730A1@winand.at>
 <CAFj8pRB4NHismecE-SnezKz9OektP2H8t_684TbqQ4Ax3swZwQ@mail.gmail.com>
 <5C283168.8060001@anastigmatix.net>
 <CAFj8pRDRrxyr8=2PEFxP_eZCRqJMU4Jpj0whJM9pX0V4Q4EQvw@mail.gmail.com>
 <5C291741.9060506@anastigmatix.net>
 <CAFj8pRAmSUO1g1qnVgXmE=7bPULYjwDfyO5KDohsHMFesDenmA@mail.gmail.com>
Cc: Markus Winand <markus.winand@winand.at>,
 Alvaro Herrera <alvherre@2ndquadrant.com>, Tom Lane <tgl@sss.pgh.pa.us>,
 PostgreSQL Hackers <pgsql-hackers@postgresql.org>
From: Chapman Flack <chap@anastigmatix.net>
Message-ID: <5C297BB7.9070509@anastigmatix.net>
Date: Sun, 30 Dec 2018 21:15:19 -0500
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101
 Thunderbird/38.7.0
MIME-Version: 1.0
In-Reply-To: 
 <CAFj8pRAmSUO1g1qnVgXmE=7bPULYjwDfyO5KDohsHMFesDenmA@mail.gmail.com>
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 12/30/18 15:18, Pavel Stehule wrote:
> ne 30. 12. 2018 v 20:06 odesílatel Chapman Flack <chap@anastigmatix.net>
> napsal:
>> How difficult would it be to make the grammar constructs that currently
>> accept "BY REF" (only as noise and ignore it) accept and ignore both BY REF
>> and BY VALUE?
> 
> This is difficult question - a implementation is probably very easy, but it
> is hard to accept to possible break compatibility due syntactic sugar.
> 
> This is not probably related to just XPath/XQuery question - but it is
> related to different design of XML datatype (based on PostgreSQL TOAST)
> against ANSI/SQL (Oracle - clob).
> 
> So this is complicated topic and my opinion is better to don't touch it
> because we can't to fix it, change it - and I am not sure so ANSI/SQL is
> significantly better than PostgreSQL implementation.

I am not sure I understand your point. It appears that Oracle (18c),
just like PostgreSQL, really only supports BY VALUE semantics. Here is
an Oracle fiddle that shows it:

https://dbfiddle.uk/?rdbms=oracle_18&fiddle=0cb353da0d94c6d5c2659222a1e419fd

When the same element is passed via two parameters, an 'is' test (node
identity equality) of the two parameters returns false, indicating that
Oracle has used BY VALUE semantics, not BY REF.

Oracle uses BY VALUE when BY VALUE is explicitly requested, and also when
no passing method is specified (i.e., BY VALUE is the default). Oracle also
uses BY VALUE when BY REF is explicitly requested, which seems rather rude,
but that must be the behavior PostgreSQL is imitating with the choice to
accept and ignore BY REF.

But the PostgreSQL situation is a little more strange. PG uses BY VALUE
semantics as the default when no passing method is specified. PG also uses
BY VALUE semantics when BY REF is explicitly requested, which is rude,
just like Oracle. But why should an explicit specification of BY VALUE
(which is, after all, the semantics we're going to use anyway!) produce
this?

ERROR:  syntax error at or near "value"

To me, that doesn't seem like least astonishment.

I am not seeing what would be complicated about removing that astonishment
by simply allowing the grammar productions to also consume BY VALUE and
ignore it.

-Chap