[cairo] generation of inefficient postscript?

[Jeff Muizelaar]

On Tue, Feb 12, 2008 at 08:15:43AM +1030, Adrian Johnson wrote:
> Jeff Muizelaar wrote:
> > On Sun, Feb 10, 2008 at 08:55:00PM -0500, Adam Goode wrote:
> >> It takes 3.9s to print from xpdf and about 7s from evince. Perhaps this
> >> is because xpdf prints as 1 bit and cairo expands to full RGB? Still, an
> >> amazing improvement!
> > 
> > Poppler always gives cairo RGB24 data. Do the ps/pdf backends act
> > differently if given A1 data? If they do poppler could probably be
> > changed to give A1 data.
> 
> The PS/PDF backends both accept A1/A8 images but the end result is
> always a color image. The PDF surface embeds a black image with the
> A1/A8 data in the SMask. The PS surface blends the image over a white
> RGB24 surface to eliminate the transparency.

So to be clear, currently there is no advantage to using A1 data?

-Jeff