Re: track_commit_timestamp and COMMIT PREPARED

On 2015-09-28 18:59, Robert Haas wrote:
>>
>> The patch looks good to me except the following minor points.
>>
>> * or not. Normal path through RecordTransactionCommit() will be related
>> * to a transaction commit XLog record, and so should pass "false" here.
>>
>> The above source comment of TransactionTreeSetCommitTsData() seems to
>> need to be updated.
>>
>> + * Note: if you change this functions you should also look at
>> + * RecordTransactionCommitPrepared in twophase.c.
>>
>> Typo: "this functions" should be "this function"
>>
>> + if (replorigin_sesssion_origin == InvalidRepOriginId ||
>>
>> This is not the problem of the patch, but I started wondering what
>> "sesssion" in the variable name "replorigin_sesssion_origin" means.
>> Is it just a typo and it should be "session"? Or it's the abbreviation
>> of something?
>
> This should go in before beta. Is someone updating the patch?
>

Sorry, missed your reply.

The "sesssion" is typo and it actually affects several variables around
the replication origin, so I attached separate patch (which should be
applied first) which fixes the typo everywhere.

I reworded the comment, hopefully it's better this way.

--
Petr Jelinek http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services