On 09/10/2010 10:45 AM, Alfranio Correia wrote:
> Hi Libing,
> Excellent questions and comments.
> I will try to figure them out and commit a new patch.
> On 09/10/2010 05:00 AM, Libing Song wrote:
>> Hi Alfranio,
>> Nice work!
>> Please find my review comments below.
>> The patch is ok for this bug, but two questions need to be clarify.
>> Not Approved.
>> REQUIRED CHANGES
>> RE1. We have to caches now, statement's cache and transactional cache.
>> Should binlog_cache_use be increased twice if both satement's cache
>> and transactional cache are used.
They are supposed to be increased twice and they are.
>> I suppose you think so. As this, it should be documented. Could you
>> please add comments to mark that it should be documented.
I will add a comment to the patch and bug report and ask Jon to take care of it.
>> RE2. Should binlog_cache_use be increase on the following case?
>> INSERT INTO t1 VALUES(1);
>> I think it should be record too.
It is if the transaction is written to the binary log.
If there is nothing to be written to the binary log, nothing is updated.
If we think we should change this behavior, it is better to file another bug
I will add comments to the code though.