[PyDB2-discuss] timestamp segfault resolved, found another bug

[Jeremy H. <je...@hi...>]

Hi all,

I somehow missed that there was a mailing list for this project and sent
an email directly to Patrick.

Anyway, the problem I had with a segfault on timestamp columns is
apparently already fixed in CVS HEAD.  As a result, that's the version
that I'll be using.

I did find another bug, that I believe I have fixed correctly. I
submitted it as a patch.  The gist of it from my patch submission:

    Select statements of the type for selecting from an insert, update,
    or delete statement are not supported.  i.e.

    SELECT i FROM FINAL TABLE (INSERT into test1(a,b,c) VALUES (?,?,?))

    This type of SQL results in a segmentation fault on Linux when the
    cursor.fetchone/many() Python method is called.

    The cause is that the num_cols from SQLNumResultCols() has a value >
    0 but the row_count from SQLRowCount() results in a number
    indicating how many rows were updated by the DML statement.

    The end result is that the SQL statement was executed successfully,
    but when trying to retrieve the result from the statement PyDB2
    segfaults.

enjoy,

-jeremy

-- 
========================================================================
 Jeremy Hinegardner                              je...@hi...