Hello,
in the FAQ I read "DBF uses only charSet property when the codepage flag in DBF header field is empty".
I have files in cp852 (East European MS-DOS, code page mark 0x64) that have the code page mark filled correctly (I have checked the byte in the file header). I can see the contents correctly in Visual FoxPro.
But I _sometimes_ get garbled accented letters when I try and SELECT with HXTT DBF (versions 3.0.024 and 3.1.035).
When I get my connection with property charSet=cp852, everything is o.k.
When I get my connection with property charSet=cp1250, it is bad. Apparently, the charSet parameter was used even though the header field was not empty.
Is there any explanation for the behaviour I see here?
I think I could provide a test file, if necessary.
Jan Dvorak
MathAn Praha
|
Yeah. If there isn't charSet property, DBF will try to use the codepage flag in DBF header field. If the codepage flag in DBF header field is empty too, DBF will try to use the Java platform's default charset. We will change the information in faq.html soon.
>When I get my connection with property charSet=cp852, everything is o.k.
Yeah. If there's no any charSet property, DBF will uset Cp852 too according to the code page mark in your dbf files.
>When I get my connection with property charSet=cp1250, it is bad.
Yeah. It will try the assigned charSet.
|
Main
webmaster@hxtt.com