[GNUccess-design] Some changes

[Bartek W. <ba...@re...>]

Hi,

As you can see in CVS, I've made some changes to GNUccess. 

I've implemented the GcTrigger class more/less according to the idea that I've
learned from Przemek (If you think i did it wrong then tell me please)

I also introduced the class GcException - I think that we would need it anyway
at some point.

One of the things that I'm not so sure concerns GcEnv. I've made it a subclass
of GcObject. It's a little bit tricky - it has a reference to itself in GcEnv. I
thought that it will be more consistent with the idea of keeping the object in
exception - when GcEnv raises the exception there's no problem with the
interface, however it introduces a lot of inapropriate like execute. I'm waiting
for your comments in that matter.

I've cleaned up the mess with GUI a little bit, but it's still a piece of crap.
Anyway some things started to work. You can even use drop-down menu (although it
refers to the object selected in objectsContainer so far ;)

I can see that nobody has written the summary of the second "meeting" so I'll
just say that I'm still thinking about GcScript Object - it should come along in
a few days - Przemek is working on a GcQuery and SQL parser and Kuba is involved
in DataSource stuff (I'm not sure if that's all true - but that's what I remember)

One more thing - maybe the most interesting. I've found a very nice project -
Gadfly -  it is a small RDBMS - written comp[letely in python and it's designed
to reside in memory during execution. I think it will be much easier to just add
the serialization layer to this library (it is currently using marshal, but i
guess it shouldn't be hard to write a nice XMLserializer for it). I'll try to
dig into it (actually Przemek could also be interested - it has some kind of sql
parser in python).

See you guys
   Bartek Wilczyñski