Re: Initial thoughts.

[Andy Wingo]

Hi Rob,

Thanks for the note. It sounds like some really interesting plans are in
the works.

On Wed, 29 Oct 2003, Rob Browning wrote:

>   * "sexp representation for C" (CSE)

Interesting! Sounds like we can add this over time, no?

>   * On another topic, that of automatic wrapper generation, one of the
>     tricky bits is how you get *reliable* API information.  GTK has
>     avoided this problem by (quite nicely) providing an easily
>     parsable spec.

Actually, our spec was produced by h2defs.py and then tuned by hand. The
specs they provide are not what we are using, they're in an old format.

>     one thing people often consider is parsing the headers themselves.

Jeez, lots of changes. This branch of g-wrap is unstable. Can we just
have a release with what Andy (Rottman) has done, then move on to
implementing other things? These are fascinating questions, but we
really need to get on and churn things up a bit. Pieces at a time.

I sure do sound impatient, but to be honest my concern is with making
apps. guile-gobject works now with the old g-wrap, but if the size of
the files and the speed of the bindings are both better with the new, we
go for it. Research into C is not my area. But those are just my
interests. I want releases of all the libraries I depend on so I can
have nice releases, myself.

>   * I also looked at and speculated a bit about using libffi a while
>     ago, but it sounds like Andreas is way ahead of me there :>

Let's move on this one :)

Regards,

wingo.