Re: INSERT ... ON CONFLICT UPDATE and RLS

On 26 February 2015 at 05:43, Stephen Frost <sfrost@snowman.net> wrote:
> Dean,
>
> * Dean Rasheed (dean.a.rasheed@gmail.com) wrote:
>> Attached is a patch to make RLS checks run before attempting to
>> insert/update any data rather than afterwards.
>
> Excellent, this I really like and it's a pretty straight-forward change.
> I wonder if there are some documentation updates which need to be done
> for this also?  I'm planning to look as I vauguely recall mentioning the
> ordering of operations somewhere along the way.
>
> I also addressed the bitrot from the column-priv leak patch.  Would be
> great to have you take a look at the latest and let me know if you have
> any further comments or suggestions.  I'm definitely looking forward to
> getting these changes in soon.
>

Thanks for the update. I don't recall any mention of ordering in the
docs, but if there isn't anything there it makes sense to add
something.

I haven't thought about this patch in a while, but I still like the
look of it. It's definitely neater to do these checks earlier and it's
good to be able to get RLS-specific errors too. I'll take a look at
the latest updates.

Regards,
Dean