Re: BUG #16091: xpath fails to compute "name()", regression - Mailing list pgsql-bugs

From Cherio
Subject Re: BUG #16091: xpath fails to compute "name()", regression
Date
Msg-id CAKHqFk+3EWzphe7DTPY_JYb+fHmentAfLCX-rrmAMAj_AZu+sA@mail.gmail.com
Whole thread Raw
In response to Re: BUG #16091: xpath fails to compute "name()", regression  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-bugs
Thanks Tom,

This is what we essentially did to make the old code work.
I guess you are saying this is not considered a regression.
Than this is not a bug.

On Wed, Oct 30, 2019 at 2:10 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
PG Bug reporting form <noreply@postgresql.org> writes:
> After upgrading postgresql from 9 to 12 the following statement no longer
> produces the same result: SELECT xpath('name()', '<abc>xyz</abc>'::XML)

> PostgreSQL 9 returns '{abc}'
> PostgreSQL 12 returns '{""}'

> This behavior changed in version 11 and perpetuated into 12.

This looks to me to be an intentional change in xpath's behavior.
The v11 release notes call out the incompatibility:

    Correctly handle relative path expressions in xmltable(), xpath(), and
    other XML-handling functions (Markus Winand)

    Per the SQL standard, relative paths start from the document node of
    the XML input document, not the root node as these functions
    previously did.

Perhaps 'name(/*)' would do what you want now.

                        regards, tom lane

pgsql-bugs by date:

Previous
From: Tomas Vondra
Date:
Subject: Re: memory problems and crash of db when deleting data from tablewith thousands of partitions
Next
From: Alvaro Herrera
Date:
Subject: Re: BUG #16082: TOAST's pglz_decompress access to uninitializeddata, if the database is corrupted.