[Dynamite] another us trip...

[Ian Holmes]

On Mon, 10 Apr 2000, Ewan Birney wrote:

> 
> I am off on another US trip. 
> 
> 
> Ian - if you could give me an implementation task to do on Telegraph 
> I'll give it a go....
> 

Anything in SeqModule! Try EditSeq, then Registry. These are superficially
very simple objects (especially EditSeq) but they will involve testing our
whole IDL-to-C mapping, which makes them less trivial.

I suggest you use the idlstubs.pl I sent you a while ago to generate C
stubs, stripping out everything from the IDL except the particular objects
you want to build. Then hack the C stubs by hand until they compile,
ideally keeping notes of any non-obvious changes you needed to make.

NB I haven't yet tried compiling the C stubs with a C compiler. Ewan I
could understand if you balked at being the first to do this, but I think
you should do it anyway because otherwise we have a situation where
everyone is waiting for me to produce a sparkling final version of the
IDL-to-C compiler before they even *think* of using it, and that is BAD.  
The process of improving idlstubs.pl will go most efficiently if people
are actually using the program; this is the best way of shaking out which
revisions are absolutely necessary and which are merely glitter-paste.

The main problem with using idlstubs right now is that it uses a "has-a"
inheritance model, and it would be nice to switch this over to a GTK-style
"is-a" inheritance model where inherited data members go at the front of
the struct, as we discussed. I think I have a way to do this, but I also
think it won't be too much trouble to change, as long as we don't write
*too* much code using the "has-a" model. So I will take responsibility for
doing the changes. If anyone really can't stand the idea of using an
IDL-to-C mapping that is changing under their feet in this way, then speak
up now. We *could* wait until I change idlstubs, but again, it is an
undesirable bottleneck. Alternatively, we could just stick with the
"has-a" model. (If anyone needs a review of our inheritance-model
discussion then speak up; I'm assuming no-one does, cos I'm lazy and I
hate typing.)

Ian