Le 14/10/2021 à 14:09, Aleksander Alekseev a écrit :
> Hi again,
>
> > So all in all, -1. [...]
>
> Here is something I would like to add:
>
> 1. As far as I know, "all the rest of DBMS have this" was never a good
> argument in the PostgreSQL community. Generally, using it will turn
> people against you.
I have cited the implementation in the other RDBMS because it helps to
understand the feature, it shows the state of the art on it and
illustrates my needs. If making references to other implementation turns
people against me I think that they have the wrong approach on this
proposal and if we refuse feature because they are implemented in other
RDBMS this is even worst. I'm not agree with this comment.
> 2. I recall there was a proposal of making the SQL syntax itself
> extendable. To my knowledge, this is still a wanted feature [1]. In
> theory, that would allow you to implement the feature you want in an
> extension.
For what I've read in this thread
https://www.postgresql.org/message-id/flat/20210501072458.adqjoaqnmhg4l34l%40nol
there is no real consensus in how implementing this feature should be
done. But I agree that if the implementation through an extension was
possible I would not write a patch to core but an extension, this is my
common behavior.
Best regards,
--
Gilles Darold
http://www.darold.net/