UTF8 field displays the wrong length, in the DBgrid.
lazarus / 16-Databases

----------------------------------------------------------------------
PaulWoo · 2011-09-16 07:12

Any DB, with utf8 encoding, like MySQL, Firebird, may have this bug.
My environment is: WindowsXP sp3 Simple Chinese, Postgre 9 + zeos 6 +
Lazarus 0.9.30 + FPC2.44.

Database using utf8 encoding, a field is defined as char (3), such as
his value is '北京市' (Chinese, meaning BeiJing City), but, in the
DBGrid, the display width is only 1 character (the first character,
'北').

I think the cause of the problem, DBGrid calculate a utf8 Chinese
characters to 3 characters.
oh! This is clearly different from the other components of Lazarus.

For example, a Edit1 components, maxlengh = 3, it can enter the 'ABC'
or '北京市', three English character or 3 Chinese characters, this should
be utf8 features, right?

How can I make DBGrid do that, shows 1 Chinese character = 1 English
characters?

There is a similar question:
Re: SQLite field keeps only half of the data
Quote from: atasoft on July 17, 2011, 10:53:22 pm... I am using
ZEOSDBO-6.6.6-stable components. Anyway I have the problem on both
SQLdb and Zeos.

Using database tool Lita or SQLite Sorcerer or Sqliteman I can fill
the full size of the field (e.g 50) using greek or english characters
without problem. Using my application I can read all the 50 english
characters but only the first 25 GREEK characters....

----------------------------------------------------------------------
Arbee · 2011-09-16 09:14

I don't know about all DBMS's but there are  those where the field
length does not give you display length, but rather storage length. 
DB2, I know, is one of them.  A terrible "design" flaw in my eyes, but
there you are.  If that is the reason for your problem too, there is
not much you can do, other than taking it into account when designing
your tables.  Any characters could (when you limit yourself to the
characters in the BMP) be up to 4 bytes, so to keep it safe you should
then design a table with CHAR 12 (or VARCHAR(12)) to be able to cope
with any UTF8 character.

----------------------------------------------------------------------
jixian.yang · 2011-09-16 15:50

There is no such problem in zeoslib now: Just set CodePage to UTF8.

ZConnection1 object: TZConnection
Properties.Strings = （
""Codepage = UTF8""
""Client_encoding = UTF8""）
AutoCommit = False
Connected = True
Protocol = ""postgresql-8""
HostName = ""127 .0.0.1 ""
Port = 5432
Database = ""postgres""
User = ""postgres""
Password = ""mypassword""
Left = 126
Top = 16
end