Re: delta relations in AFTER triggers

On Sun, Oct 30, 2016 at 10:35 AM, Kevin Grittner <kgrittn@gmail.com> wrote:
> On Sun, Oct 2, 2016 at 11:20 PM, Michael Paquier <michael.paquier@gmail.com> wrote:

>> Not as big as I thought, only 2k when both patches are combined... The
>> patch without noapi in its name needs to be applied first, and after
>> the patch with noapi can be applied.
>>  60 files changed, 2073 insertions(+), 63 deletions(-)
>> Moved to next CF.
>
> In an attempt to make this patch more digestible for reviewers, I
> have split it up as follows:
>
> transition-c-triggers-only-v7.diff

>  17 files changed, 581 insertions(+), 47 deletions(-)
>
> This part is virtually unchanged (just curing bit-rot) since
> August, 2014, when I believe I had addressed all issues raised by
> reviewers.  It does provide a barely usable feature, since the
> syntax for transition tables is added and tuplestores are created
> when needed (and only when needed), with references stored in the
> TriggerData structure.  No new execution nodes are provided, so
> only C triggers can use these relations, and must explicitly and
> directly access the tuplestores from within the C code -- there is
> no support for referencing these tuplestores from within queries.
>
> This is basic infrastructure needed for the more complete feature.
> As far as I know there are no objections to what is implemented
> here.  I have pulled it out to make the review of the more
> controversial portions easier.  Since it had quite a bit of review
> two years ago, I will do some testing to make sure that nothing has
> broken and then push this part in a few days if there are no
> objections.

Hearing none, done.  Hopefully that makes what remains easier to
review.

During final testing I was annoyed by the thin support for CREATE
TRIGGER in the tab completion code, so I improved that a bit and
pushed that, too.

--
Kevin Grittner
EDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company