[cairo] Re: [poppler] Re: poppler profiling on geode with cairo backend

[Behdad Esfahbod]

On Thu, 2006-11-09 at 03:42 -0500, Carl Worth wrote:
> [I'm taking this conversation over to the cairo list---please drop the
> poppler list if you reply.]
> 
> On Wed, 8 Nov 2006 21:51:36 -0500, Jeff Muizelaar wrote:
> > So I had a look at why pixman_compositeGeneral is being hit during text
> > rendering. It looks like this gets called for text when the document is
> > being rendered at weird scaling factors.
> 
> The short answer is that any time compositeGeneral is called there's a
> bug somewhere. (That function is so slow, we know we never want to use
> it).
> 
> > For example with the pdf you provided: calling
> >   poppler_page_render_to_pixbuf(..., scale=1.0, ...)
> > does not trigger any calls to pixman_compositeGeneral.
> ...
> > However,
> >   poppler_page_render_to_pixbuf(..., scale=0.5, ...)
> > does.
> ...
> > I will follow up with the cairo people about why this happens and if
> > anything can be done to improve the situation. In the mean time it may
> > be possible to work around the problem by avoiding bad scaling factors.

Is it a Type3 font?  Type3 fonts are drawn as path in poppler currently
IIRC.  Can you post a test PDF and driver code?

behdad


> That looks really odd to me. In general, text should be rendered
> through the same code paths whether or not there is scaling in the CTM
> I would think.
> 
> If you can construct a pretty small test case demonstrating this
> behavior then that would be useful.
> 
> Thanks in advance,
> 
> -Carl
> _______________________________________________
> cairo mailing list
> cairo at cairographics.org
> http://cairographics.org/cgi-bin/mailman/listinfo/cairo
-- 
behdad
http://behdad.org/

"Commandment Three says Do Not Kill, Amendment Two says Blood Will Spill"
        -- Dan Bern, "New American Language"