Re: performance of goocanvas

[D M German <dmg uvic ca>]

Ed, Damon, 

I really want to thank you both. I was able to find the reason for the
performance hit. As you indicated, we were "moving" items to the same
location where they were _for the entire document_. I have found some
redundancy in some operations during scrolling (and probably in other
places too) that I'll remove.

thank you very much,

--daniel

 Ed> I think this is a good guess, we had a lot of trouble with foocanvas
 Ed> (the goocanvas predecessor) with this. In the end we were forced to
 Ed> draw certain things ourselves on to the screen in order to get good
 Ed> enough performance.

 Ed> If you are moving the scrolled region as you scroll that would do it....

 Ed> Ed



> > I don't think the problem is drawing. I put a "return;" statement at the
> > top of the GooCanvasPolyline's draw function so it didn't draw anything,
> > but it was just as slow!
> >
> > I think the problem may be that you are sometimes changing items when
> > the canvas is scrolled, and this causes the canvas to have to
> > recalculate the bounds of lots of items, which is slow.
> >
> > What changes are you making when the canvas is scrolled? Is there a way
> > to avoid them?
> >
> > Damon
> >
> >
> > _______________________________________________
> > goocanvas-list mailing list
> > goocanvas-list gnome org
> > https://mail.gnome.org/mailman/listinfo/goocanvas-list

 Ed> -- 
 ------------------------------------------------------------------------
 Ed> | Ed Griffiths, Acedb/ZMap development, Informatics Group,               |
 Ed> |  The Morgan Building, Sanger Institute, Wellcome Trust Genome Campus   |
 Ed> |  Hinxton, Cambridge CB10 1HH                                           |
 Ed> |                                                                        |
 Ed> | email: edgrif sanger ac uk  Tel: +44-1223-496844  Fax: +44-1223-494919 |
 ------------------------------------------------------------------------


 Ed> -- 
 Ed> The Wellcome Trust Sanger Institute is operated by Genome Research
 Ed> Limited, a charity registered in England with number 1021457 and a
 Ed> company registered in England with number 2742969, whose registered
 Ed> office is 215 Euston Road, London, NW1 2BE. 


--
Daniel M. German                  "Much of my work has come
   John W. Backus ->               from being lazy"
http://turingmachine.org/
http://silvernegative.com/
dmg (at) uvic (dot) ca
replace (at) with @ and (dot) with .