Re: [guppi-list] New Guppi Snapshot

[Rob Browning <rlb cs utexas edu>]

Jon Trowbridge <trow@emccta.com> writes:

> It may be beyond the scope of what you want to do with g-wrap, but
> I'd need to have support for generating python bindings.  Ideally,
> it would then later grow to support bindings for other embeddable
> scripting alternatives like perl and gnome-basic.

I'd be happy with that.  It's been something "planned" for g-wrap for
a while, but as yet, no-one's implemented anything.

However, just before I took over g-wrap, Christopher Lee, the original
author, was working on a major re-write that would have made the
g-wrap "backends" much more modular, so before adding other language
backends, it might be worth spending a little bit of time on that kind
of abstraction.  Ideally, I'd like to see the "target language output"
bits be completely modular, with a well specified API so that you
could just "drop in" a backend for a new language.

I've already started moving in that direction with the new command
line args, but nothing more than that yet.  The good thing is that
g-wrap, even as it stands now, is not that much code, so it's not hard
to wrangle it around to whatever's more appropriate.

> I would go so far as to say that g-wrap (or something like it)
> should cease to be an independent entity and should be assimilated
> into the currently-under-discussion gnome-office-lib.  After all,
> multi-language scripting is something that should be encouraged in
> all GNOME apps.

Fine with me.  I'm not stuck on g-wrap either.  I just want the best
solution to the problem.  If g-wrap continues along and becomes that,
great.  If something better surpasses it, then that's fine too, and
I'll be happy to help there.

Thanks

-- 
Rob Browning <rlb@cs.utexas.edu> PGP=E80E0D04F521A094 532B97F5D64E3930