On Thu, Oct 27, 2016 at 7:38 PM, Michael Paquier
<michael.paquier@gmail.com> wrote:
>> I committed and back-patched this with some additional work on the
>> comments, but I don't understand this remark. That comment seems like
>> it should refer to the checkpointer in modern branches, but isn't that
>> point independent of this patch?
>
> * During recovery, we keep a copy of the latest checkpoint record here.
> - * Used by the background writer when it wants to create a restartpoint.
> + * lastCheckPointRecPtr points to start of checkpoint record and
> + * lastCheckPointEndPtr points to end+1 of checkpoint record. Used by the
> + * background writer when it wants to create a restartpoint.
>
> The patch committed introduces lastCheckPointEndPtr, which is not used
> to decide if a restart point should be created or not.
The comment doesn't say anything about those structure members being
used to decide "if a restart point should be created or not". It just
says that they are used; it says nothing about the purpose for which
they are used.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company