On Thursday 04 March 2010 07:36:22 Antti Koivisto wrote: > On Thu, Mar 4, 2010 at 1:05 AM, Zack Rusin <[email protected]> wrote: > > 3) it's 2x memory usage vs qpixmap. > > By 2x you mean the original bitmap + the texture cache in GL context?
Correct. But not only that, every single platform needs 2x mem, not only GL. The data in QImage needs to be uploaded to be displayed, so whether one is using X11, VG, GL, with or without caching when rendering QImage will be 2x mem no matter what. > The cache presumably has limited size and mostly contain textures for > images that have been displayed recently. Typically only a small > portion of all images on a web page are visible at a given time. The > real memory use should not be anywhere near 2x assuming the cache > works sensibly. Of course once that limit hits it means that the data will have to be uploaded on every rendering meaning that rendering will be exponentially slower. Anyway, it's a purely academic discussion at this point. Like I mentioned as long as X11 works the way it does and X11 engine is the default (and as Girish pointed out NPAPI works the way it does) there's no way we could switch to QImage's globally. z _______________________________________________ webkit-qt mailing list [email protected] http://lists.webkit.org/mailman/listinfo.cgi/webkit-qt
