Phil Steitz wrote:
> Mark Thomas wrote:
>> Phil Steitz wrote:
>>> OK, I am now feeling stupid.
>> You and me both.
>>
>>> GKOP already has LRU-cache like
>>> behavior. (See clearOldest test in allocate()). Given compat
>>> issues, I am inclined to move forward with 0).
>> +1
>>
>> We'll also need pool 1.5.4 as well to fix an edge case where the pool
>> could lock up.
>>
>> Mark
>>
>
> One question. The patch introduces a statement type flag as part of
> the key to distinguish prepared statements from callable statements.
> I am not sure this is necessary. Can anyone imagine a scenario
> where the same exact sql is used in prepareStatement and
> prepareCall? Are there drivers that even allow this ("call" in
> prepareStatement sql, or no "call" in prepareCall) If not, we do not
> need the additional flag.
As CallableStatement extends PreparedStatement there may be some edge
cases where the same sql could be used. It isn't something I would do
but I can how it might work with some drivers (I haven't tested any).
Given that the sql would normally be very different, I don't see the
harm including the statement type in the key apart from a small
inefficiency in key construction but I don't see that as a major issue.
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
