BUG #4987: registered in pg_type by a wrong table name

The following bug has been logged online:

Bug reference: 4987
Logged by: durfenlee
Email address: durfenlee(at)gmail(dot)com
PostgreSQL version: 8.4.0
Operating system: Windows XP Professional(32bit)
Description: registered in pg_type by a wrong table name
Details:

Hi, My name is durfenlee.
I found bug to postgresql8.4.0

When create table by long table name(63byte just)
and last charcter is multi-byte charcter,
It is registered in pg_type by a wrong table name.

At this time, it becomes an encoding error when the encoding of the database
and the client is different.
(There is an influence in processing that uses pg_type. like a pg_dump. )

[My Environment]
OS : Windows XP Professional(32bit)
PostgreSQL : "PostgreSQL 8.4.0, compiled by Visual C++ build 1400, 32-bit"
Database Encording : UTF8
Client Encording : SJIS

/***************************************************************************
******************
* 1. long table name's(63 byte just) last charcter is not muli-byte
charcter
****************************************************************************
******************/

[run psql]
===============================
postgres=# select typname from pg_type where typname like '%abcdefghi%';
typname
---------
(0 row)

postgres=# create table
abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_123(test int);
CREATE TABLE
postgres=# select typname from pg_type where typname like '%abcdefghi%';
typname
-----------------------------------------------------------------
abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_123
_abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_12
(2 row)

postgres=#
===============================

when create table,2 recorde registered in pg_type(table type and enum
type?).
if table names is 63byte, enum typname names is named by the following
rules.

- "_" is applied first.
- The last 1 byte is cut down.

/***************************************************************************
**************
* 2. long table name's (63 byte just) last charcter is muli-byte charcter
****************************************************************************
**************/

[run psql]
===============================
postgres=# create table
abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_あ(test int);

//(last charcter is JAPANESE HIRAGANA. UTF8 code is "0xE38182")

CREATE TABLE
postgres=# \encoding
UTF8
postgres=# select typname from pg_type where typname like '%abcdefghi%';
typname
----------------------------------------------------------------
abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_縺・
_abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_abcdefghi_
(2 row)

postgres=# \encoding SJIS
postgres=# \encoding
SJIS
postgres=# select typname from pg_type where typname like '%abcdefghi%';
ERROR: invalid byte sequence for encoding "UTF8": 0xe381
HINT: This error can also happen if the byte sequence does not match the
encoding expected by the server, which is controlled by "client_encoding".
===============================

For last charcter is multi-byte, It fails in encording if 1 byte is cut down
only.

Thanks!!

"durfenlee" <durfenlee(at)gmail(dot)com> writes:
> When create table by long table name(63byte just)
> and last charcter is multi-byte charcter,
> It is registered in pg_type by a wrong table name.

Hmm, yup, seems to be broken in 8.3 as well. Thanks for the report!

regards, tom lane