Re: RFS: xpdf (updated package)
On Fri, 4 Jun 2010 11:56:00 +0400, Stanislav Maslovski wrote:
> On Fri, Jun 04, 2010 at 12:29:07AM -0400, Michael Gilbert wrote:
> > In terms of xpdf performance, can those concerned please try files that
> > they consider big with the poppler-ized version and compare that to the
> > original xpdf so we can actually quantify the impact (if there even is
> > one). Speculating doesn't really get us anywhere. We need a
> > quantified impact.
>
> I have not tested the poplerized version of xdpf yet. Because you
> mentioned speculations, let me in turn say that I have not seen
> your answers yet on the questions I raised in my first mail (besides
> some speculations, of course). Rogério mentioned some facts (thanks!)
> but could you please provide more details?
I can't say anything definitive, but I can speculate that it will
not be a problem, and here is the logic:
xpdf's rendering code is itself essentially an older version of
poppler. I doubt the poppler developers have intentionally made that
code itself any slower, and it was possible to interface xpdf to
poppler without a significant rewrite (a couple minor few-line patches
and a bunch of variable renaming that doesn't have any impact). Hence
any performance differences to be found between evince and xpdf will
reside in the non-poppler code. Restated; even with switch to poppler,
the original xpdf zoom/display logic remains, and that will be the
primary driving factor for performance.
But even so, it should be tested.
Best wishes,
Mike
Reply to: