El Dilluns, 20 de juliol de 2015, a les 15:26:42, Jeremy Maitin-Shepard va escriure: > On Sat, Jul 18, 2015 at 7:27 AM, Albert Astals Cid <[email protected]> wrote: > > El Divendres, 17 de juliol de 2015, a les 12:52:30, Jeremy Maitin-Shepard > > va > > > > escriure: > > > Forwarding to the list, since I sent my original reply incorrectly... > > > > > > On Fri, Jul 17, 2015 at 12:51 PM, Jeremy Maitin-Shepard < > > > > [email protected] > > > > > > wrote: > > > > > > > > On Thu, Jul 16, 2015 at 3:16 PM, Albert Astals Cid <[email protected]> > > > > wrote: > > > > [snip] > > > > > > > >> > Any advice about how to deal with this redraw problem? > > > >> > > > >> Trigger the redraw yourself? See Saheeb's branch for optinal content > > > >> (origin/gsoc2015_layer_feature), I'd say it probably makes sense for > > > > you > > > > > >> to > > > >> base your work on his branch. > > > > > > > > Thanks, I'll take a look. > > I see that the gsoc2015_layer_feature branch just triggers a redraw of > every page in the document any time there is any change to optional content > group visibility. Merging my changes on top of this branch indeed makes > the animation kind of work. However, there are some serious performance > issues. Redrawing the whole document for each frame of the animation kills > performance if there are many pages. To avoid this though would seem to > require poppler providing an interface for determining which pages are > affected by each optional content group.
That'd be ideal, yes. > (This is needed for a proper > implementation of the Javascript OCG interface anyway.) Alternatively a > hack might be to redraw only the current page; not sure how this would > interoperate with the layer panel. > > Even in single page documents, though, redrawing the whole page is often > too slow to achieve reasonable frame rate. Somehow acrobat reader is > significantly faster. It would be nice if there were a way to prerender > each optional content group individually so that toggling the visibility > wouldn't require rerendering. That's another option, though it probably increses considerably the memory usage it's worth exploring. Cheers, Albert _______________________________________________ Okular-devel mailing list [email protected] https://mail.kde.org/mailman/listinfo/okular-devel
