Re: [gtk-list] Re: signals, objects, user_data and stuff.

[Thomas Mailund Jensen <mailund daimi au dk>]

>>>>> "M" == Martin Kestel <mka@mppmu.mpg.de> writes:

 M> Hmmmm My apologies. I was reading too fast and understood you kept
 M> some information inside some file and had to read this in
 M> later. Sorry.

Well, I _do_ store some information in files, but only between
sessions ;)  Actually almost everything I use is stored in either
lists og hash tables.

 M> ....but: what do think of the approach I suggested?
 
It would probably work, but I would need to do a lot of bookkeeping
which I was trying to avoid by using signals.  In you approach the
nodes keep track all connectors linked to them.  In my approach the
don't even know about them, they just emit a signal whenever they are
moved, and anyone interrested in a given node's movement just connects
to the signal.  Propagating the movement and updating all data happens
almost automatically.

This is for the endpoints.  It is not as simple for the bend points,
since they need to know which index they have.  So here I think I'll
go with Tim Janik's suggestion and store the index with the bend
point.  Since each bend point is only associated with one connector
this should work without any problems.

        /mailund