Re: BUG #4622: xpath only work in utf-8 server encoding

The following bug has been logged online:

Bug reference: 4622
Logged by: Sergey Burladyan
Email address: eshkinkot(at)gmail(dot)com
PostgreSQL version: 8.3.5
Operating system: Debian testing
Description: xpath only work in utf-8 server encoding
Details:

hello, all !

i am trying for test parse xml string in other than utf-8 encoding, it
correctly loaded but xpath(text, xml) can't handle it:

seb(at)seb:~/tmp/pg$ echo $LANG
ru_RU.CP1251
seb(at)seb:~/tmp/pg$ /usr/lib/postgresql/8.3/bin/postgres -p 5433 -k s -s -D .
LOG: система была отключена: 2009-01-22 16:30:07 MSK
LOG: autovacuum launcher started
LOG: database system is ready to accept connections

seb(at)seb:~$ echo $LANG
ru_RU.CP1251
seb(at)seb:~$ psql -h localhost -p 5433
Welcome to psql 8.3.5, the PostgreSQL interactive terminal.

Type: \copyright for distribution terms
\h for help with SQL commands
\? for help with psql commands
\g or terminate with semicolon to execute query
\q to quit

seb=# select * from (select
xml('<русский>язык</русский>')) as x(v);
v
-------------------------
<русский>язык</русский>
(1 запись)

seb=# select xpath('/русский/text()', v::xml) from (select
xml('<русский>язык</русский>')) as x(v);
ERROR: could not parse XML data
DETAIL: Entity: line 1: parser error : Input is not proper UTF-8, indicate
encoding !
Bytes: 0xF0 0xF3 0xF1 0xF1
<x><русский>язык</русский></x>
^
seb=# select name, setting from pg_settings where name like 'lc_%' or name
like '%enco%';
name | setting
-----------------+--------------
client_encoding | WIN1251
lc_collate | ru_RU.CP1251
lc_ctype | ru_RU.CP1251
lc_messages | ru_RU.CP1251
lc_monetary | ru_RU.CP1251
lc_numeric | ru_RU.CP1251
lc_time | ru_RU.CP1251
server_encoding | WIN1251
(8 rows)

in utf-8 server encoding it work correctly:

seb=> select xpath('/русский/text()', v::xml) from (select
xml('<русский>язык</русский>')) as x(v);
xpath
--------
{язык}
(1 запись)

seb=> select name, setting from pg_settings where name like 'lc_%' or name
like '%enco%';
name | setting
-----------------+-------------
client_encoding | UTF8
lc_collate | ru_RU.UTF-8
lc_ctype | ru_RU.UTF-8
lc_messages | ru_RU.UTF-8
lc_monetary | ru_RU.UTF-8
lc_numeric | ru_RU.UTF-8
lc_time | ru_RU.UTF-8
server_encoding | UTF8
(8 rows)

i am think something is wrong here, string parsed correctly by xml(text),
but it result can't pass to xpath function...

On Thursday 22 January 2009 15:39:00 Sergey Burladyan wrote:
> seb=# select xpath('/русский/text()', v::xml) from (select
> xml('<русский>язык</русский>')) as x(v);
> ERROR: could not parse XML data
> DETAIL: Entity: line 1: parser error : Input is not proper UTF-8, indicate
> encoding !
> Bytes: 0xF0 0xF3 0xF1 0xF1
> <x><русский>язык</русский></x>
> ^

This raises the question: What are the rules about encoding the characters in
XPath expressions themselves? I haven't found anything about that in the
standard. Anyone know?

23 января 2009 г. 0:58 пользователь Peter Eisentraut <peter_e(at)gmx(dot)net> написал:
> On Thursday 22 January 2009 15:39:00 Sergey Burladyan wrote:
>> seb=# select xpath('/русский/text()', v::xml) from (select
>> xml('<русский>язык</русский>')) as x(v);
>> ERROR: could not parse XML data
>> DETAIL: Entity: line 1: parser error : Input is not proper UTF-8, indicate
>> encoding !
>> Bytes: 0xF0 0xF3 0xF1 0xF1
>> <x><русский>язык</русский></x>
>> ^

> This raises the question: What are the rules about encoding the characters in
> XPath expressions themselves? I haven't found anything about that in the
> standard. Anyone know?

PostgreSQL does not use libxml2 internal encoding support and strip
xml encoding from xml body, so i think there is no choice, by default
for libxml2 it must be in it internal encoding utf-8 anyway.

i am not sure about xml standard but may be documentation of libxml2
can help to solve this issue ? see http://xmlsoft.org/encoding.html

"What does this mean in practice for the libxml2 user:
* xmlChar, the libxml2 data type is a byte, those bytes must be
assembled as UTF-8 valid strings. The proper way to terminate an
xmlChar * string is simply to append 0 byte, as usual.
* One just need to make sure that when using chars outside the ASCII
set, the values has been properly converted to UTF-8"

I understand this as: all xmlChar strings must be in utf-8 encoding,
no matter what is encoding of xml body

i try to fix this issue for xpath function, see patch in attachment

by the way, contrib/xml2 also have this issue...