Horton

Re: Fix XML handling with DOCTYPE

Chapman Flack <chap@anastigmatix.net>

View thread View on postgr.es
From: Chapman Flack <chap@anastigmatix.net>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: Ryan Lambert <ryan@rustprooflabs.com>, pgsql-hackers@lists.postgresql.org, Pavel Stehule <pavel.stehule@gmail.com>, Tom Lane <tgl@sss.pgh.pa.us>
Date: 2019-03-28T23:45:24Z
Lists: pgsql-hackers

Attachments

On 03/27/19 19:27, Chapman Flack wrote:
>   A column marked FOR ORDINALITY will be populated with row numbers
>   matching the order in which the output rows appeared in the original
>   input XML document.
> 
> I've been skimming right over it all this time, and that right there is
> a glaring built-in reliance on the observable-but-disclaimed iteration
> order of a libxml2 node-set.

So, xml-functions-type-docfix-6.patch.

I changed that language to say "populated with row numbers, starting
with 1, in the order of nodes retrieved from the row_expression's
result node-set."

That's not such a terrible thing to have to say; in fact, it's the
*correct* description for the standard, XQuery-based, XMLTABLE (where
the language gives you control of the result sequence's order).

I followed that with a short note saying since XPath 1.0 doesn't
specify that order, relying on it is implementation-dependent, and
linked to the existing Appendix D discussion.

I would have like to link directly to the <listitem>, but of course
<xref> doesn't know what to call that, so I linked to the <sect3>
instead.

Regards,
-Chap

Commits

  1. Improve documentation about our XML functionality

  2. Improve documentation about our XML functionality.

  3. Doc: clarify that REASSIGN OWNED doesn't handle default privileges.

  4. Suppress Append and MergeAppend plan nodes that have a single child.

  5. Accept XML documents when xmloption = content, as required by SQL:2006+.

  6. Ensure xmloption = content while restoring pg_dump output.