On Mon, Mar 14, 2016 at 6:46 PM, David Steele <david@pgmasters.net> wrote:
> On 2/24/16 12:40 AM, Michael Paquier wrote:
>
>> This has the merit to be clear, thanks for the input. Whatever the
>> approach taken at the end we have two candidates:
>> - Extend XLogInsert() with an extra argument for flags (Andres)
>> - Introduce XLogInsertExtended with this extra argument and let
>> XLogInsert() in peace (Robert and I).
>> Actually, I lied, there was still something I could do for this
>> thread: attached are two patches implementing both approaches as
>> respectively a-1 and a-2. Patch b is the set of logs I used for the
>> tests to show with a low checkpoint_timeout that checkpoints are
>> getting correctly skipped on an idle system.
>
>
> Unfortunately neither A nor B apply anymore.
>
> However, since the patches can still be read through I wonder if Robert or
> Andres would care to opine on whether A or B is better now that they can see
> the full implementation?
>
> For my 2c I'm happy with XLogInsertExtended() since it seems to be a rare
> use case where flags are required. This can always be refactored in the
> future if/when the use of flags spreads.
>
> I think it would be good to make a decision on this before asking Michael to
> rebase.
That's a bit embarrassing, the last versions should be able to apply
cleanly as there have not been changes in this area of the code
lately... But... I did a mistake when generating the patches by
diff'ing them from an incorrect commit number... This explains why
they exploded in size, so attached are the corrected rebased versions.
Too many patches I guess.. And both of them are attached by the way.
--
Michael