Re: Adding Node support in outfuncs.c and readfuncs.c

Dimitri Fontaine <dimitri(at)2ndQuadrant(dot)fr> writes:
> Peter Eisentraut <peter_e(at)gmx(dot)net> writes:
>> This is a massive amount of code that very few people in our community
>> will use, and very few be able to maintain it, too.

It's not that "massive", at least not as it stands, although I agree it
looks distressingly write-only.

>> If you want to make
>> a lasting contribution on this area, write a program that generates the
>> node handling functionality automatically as part of the build process.

> Can emacs --batch be used there? If not, apart from C and perl, what
> can I use?

You can *not* assume that emacs is available in any random build
environment; and not C either, because it might be a cross-compile.
It'd have to be Perl.

FWIW, even though I use emacs exclusively, I have little or no interest
in this approach myself. I don't think the node functions are as
boilerplate as you think --- for one thing, how will you deal with
typedef'd field types? (Assuming any unknown type name is a node type
is not right.) And even ignoring that, there are always exceptions to
any general rule.

If Peter is seriously suggesting that construction of the backend/nodes
files could be automated entirely, I think he's nuts. The amount of
work to construct a bulletproof tool (if it's possible at all) would
greatly outweigh the benefit. What you've got here could be useful
to people who use emacs and understand they've got to hand-check the
results. I'm not sure how much further it'd be useful to go.

regards, tom lane