Re: TODO question

> Tom Lane:
> > In fact the patch seemed quite incomplete to me; adding a new parsenode
> > type requires much more than just a struct declaration.
>
> btw, it's not correct, that just a new structure has been declared. I added
> the T_Default to the Type-Enum and it seems to me, my new parsenode type has
> been full-automatically integrated in the parser-workflow. In the gram.y,
> there is a new set of rules describing the DEFAULT value in the INSERT
> stmt - this is the place, where it's being identified and node-ed (using
> it's type), the transformation has got the new T_Default-case leaving this
> node "as is", and it's being transformed (replaced by the default value
> taken from the relation specified by the corresponding parsestate-field)
> later.
>
> > But this isn't
> > the right time of the cycle to be reviewing new-feature patches.
>
> ok, but I hope you've got a 3%-free--ear-capacity at least to answer some of
> my questions (having a very bad timing ,-) ). I don't ask offen and about
> every step, but sometimes it breaks through...

Sure, ask away. We will do our best. In fact, adding a new node is
pretty tricky. There is a developer's FAQ item about it, number 7. I
assume you read that already.

I may be able to take you patch and add the needed node support stuff,
and send the patch back to you so you can continue on it.

--
Bruce Momjian | http://candle.pha.pa.us
pgman(at)candle(dot)pha(dot)pa(dot)us | (610) 853-3000
+ If your life is a hard drive, | 830 Blythe Avenue
+ Christ can be your backup. | Drexel Hill, Pennsylvania 19026