Re: bug in PreparedStatement of JDBC in 7.2b4

I had just commited your last patch before seeing this new patch.
However, I noticed the same problem as you did and fixed it myself in
the fix I just committed (although I did the padding in a slightly
different way than you do here). This fix should be in 7.2RC1 (but
won't be in 7.2b5). It is also available in the latest build on
jdbc.postgresql.org.

thanks,
--Barry

Ryouichi Matsuda wrote:

> Barry Lind wrote:
>
>>Thank you for reporting this bug. I will work on getting your patch
>>applied this weekend.
>>
>
> I attach a patch for JDBC1 of this bug.
> Because there was not StringBuffer#replace() in JDK1.1, I changed it.
> I turned the same modification into a patch for JDBC2.
>
>
>
>>Ryouichi Matsuda wrote:
>>
>>
>>>I found a bug in PreparedStatement#setTimestamp() of JDBC included
>>>in PostgreSQL7.2beta4. An attached patch corrects this bug.
>>>
>>>
>>>(1) ".01" sec becomes ".10" sec
>>>
>>> $ create table test(ts timestamp);
>>>
>>> String sql = "insert into test values(?)";
>>> PreparedStatement pst = db.prepareStatement(sql);
>>> pst.setTimestamp(1, Timestamp.valueOf("2002-01-10 19:30:59.01"));
>>> pst.executeUpdate();
>>>
>>> $ select * from test;
>>> $ 2002-01-10 19:30:59.10+09
>>>
>>>Though I inserted ".01" second, ".10" second has been inserted.
>>>There is this bug in 7.2beta4 and 7.1.3.
>>>
>>>
>>>(2) ".876543210" sec becomes ".87" sec
>>>
>>> pst.setTimestamp(1, Timestamp.valueOf("2002-01-10 19:30:59.876543210"));
>>> pst.executeUpdate();
>>>
>>> $ select * from test;
>>> $ 2002-01-10 19:30:59.87+09
>>>
>>>In PostgreSQL7.2, a decimal can insert only two columns, and remainder
>>>is thrown away.
>>>