Horton

Re: PostgreSQL vs SQL/XML Standards

Chapman Flack <chap@anastigmatix.net>

View thread View on postgr.es
From: Chapman Flack <chap@anastigmatix.net>
To: Pavel Stehule <pavel.stehule@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>
Date: 2019-01-14T01:41:34Z
Lists: pgsql-hackers 
There is a bug that remains, in the
else if (xpathobj->type == XPATH_STRING)
case.

As it is now, it simply passes the string value of the result
into the output column's type-input function, regardless of the
output column type.

If the output column type is xml, this will attempt to parse the
string as xml. The result should simply be xml content consisting of
a text node representing the string (as by XMLTEXT()). If it contains
XML metacharacters, they should be escaped.

For a non-xml output column, the string should be used directly,
as it is now.

# select * from xmltable('.' passing xmlelement(name a)
columns a text path '"<foo/>"', b xml path '"<foo/>"');
   a    |   b
--------+--------
 <foo/> | <foo/>

Oracle fiddle for comparison:

https://dbfiddle.uk/?rdbms=oracle_18&fiddle=26f91a9e55a6908c2bcf848b464ca381

A 	B
<foo/> 	&lt;foo/&gt;

-Chap

Commits

  1. Improve documentation about our XML functionality

  2. Improve documentation about our XML functionality.

  3. Add volatile qualifier missed in commit 2e616dee9.

  4. Fix crash with old libxml2

  5. Fix minor deficiencies in XMLTABLE, xpath(), xmlexists()

  6. Fix the BY {REF,VALUE} clause of XMLEXISTS/XMLTABLE

  7. doc: Update README.links

  8. XPath fixes: