Re: an interesting test of gnumeric
- From: Miguel de Icaza <miguel nuclecu unam mx>
- To: tlewis mindspring net, federico metropolis nuclecu unam mx
- CC: gnome-list gnome org
- Subject: Re: an interesting test of gnumeric
- Date: Thu, 7 Jan 1999 21:33:01 -0600
> Now you have a big fibbionacci spreadsheet with lots of recursive
> addition on it. Set the bottom right selector thingie to average.
> Now start selecting a single cell and dragging your collection to cover
> a large rectangle. The expansion of the selection region is very
> slow for me.
Yes. The problem is not with the computation engine, but with the
code redraw engine in the canvas.
My guess is that I am invalidating the whole region instead of the
changes (this is just a theory though, I believe I invalidate the
proper spots in the cell marking code) and this might be causing the
slowness.
This first appeared on the Dec 15 hacking that Fedeico did to the
canvas, I am not sure what might be causing this.
> 1) Perhaps these values could be cached rather then being recomputed
> on the fly;
They are not recomputed at all unless you change a cell.
> He glanced at gnumeric for about 5 seconds and said "I'm
> installing GNOME." Keep up the great work!
I am glad to hear that. I will look into this problem, it might very
well be my rectangle-invalidate code.
Miguel.
[
Date Prev][
Date Next] [
Thread Prev][
Thread Next]
[
Thread Index]
[
Date Index]
[
Author Index]