Re: printing (was Re: [Rosegarden-devel] Hullo. Quick question.)

[Guillaume L. <gla...@te...>]

On Saturday 27 March 2004 14:50, Chris Cannam wrote:
> On Friday 26 Mar 2004 9:50 pm, Guillaume Laurent wrote:
> > On Friday 26 March 2004 10:26, Chris Cannam wrote:
> > > The latter point probably means that using a full-size canvas
> > > isn't fatal.
> >
> > If that was true, then commenting out the call to
> > QCanvas::drawArea() wouldn't speed things up.
>
> Not true, it could (and I believe does) just mean that the time
> drawArea() takes is related to the number and complexity of items on
> the canvas rather than its size.

I don't think so. I've also tried leaving the canvas to its default, much 
smaller size and again the printing of my sample 5 pages doc went down to a 
few seconds.

> If you leave in the drawArea() call 
> but omit to render the canvas items first, it takes no time at all.

Of course, there's nothing to draw.

> So using a large canvas but only filling in and copying a small part
> of it shouldn't be too dreadfui an overhead.

No, it forces the canvas to allocate huge amounts of memory.

Tell me when you're done with your experiments on the canvas, then I'll try 
printing with a page-sized canvas. I'm quite sure it will solve the problem.


> There are quite a few things conspiring to make
> the process slow

No argument here. But I maintain that the primary one is the size of the 
canvas.

-- 
						Guillaume.
						http://www.telegraph-road.org