TomohitoNakayama <[EMAIL PROTECTED]> writes:

> Hello Knut.
>
> By the way, I want to figure it out what is the actual range of
> modification in DERBY-1610.
>
> Knut Anders Hatlen commented on DERBY-1610:
>>Therefore, I think that it is better to fix all of them at once.
>
> What is "all of them" here ?
>
> Just setBlob and setBinaryStream ?
> Or whole case of parameterMapping.java ?

I was thinking about all the setter methods. I think we only need a
method in am/PreparedStatement which checks whether a given type is
compatible with the type of the parameter, and a call to that method
in PreparedStatement.checkSetterPreconditions().

> Too high goal may block other task,
> then reasonable goal should be established.

Good point! If we use a generic approach to begin with, we could add
checks for a subset of the types and increase incrementally. I think
that if we do it for setBinaryStream() we should also do it for
setBlob() and setBytes() in the same run.

Next (in a separate patch), we could do the same for setAsciiStream(),
setCharacterStream() and setString().

Then we fix all the scalar setters (setInt, setFloat & co.), and
finally date, time and timestamp.

> I think it is needed to analyze current difference of
> parameterMapping.out between Embedded and Network Client in order to
> consider the goal...

Sounds like a good idea.

-- 
Knut Anders

Reply via email to