On Wed, Sep 10, 2008 at 11:46 PM, Erik Garrison <[EMAIL PROTECTED]> wrote: > On Tue, Sep 09, 2008 at 12:10:41PM +0200, Tomeu Vizoso wrote: >> On Mon, Sep 8, 2008 at 3:44 PM, Eben Eliason <[EMAIL PROTECTED]> wrote: >> > >> > Should we verify that the activity is the active one, and that the >> > desktop is not shown, before taking an invalid screenshot? Also, when >> > is compositing potentially coming? Is it necessary to do this for >> > 8.2.1 if we'll have a far better solution which handles all cases in a >> > release or two? >> >> We have a chance of improving screenshot taking without composition, >> see https://dev.laptop.org/ticket/8303 . > > I'm slightly confused about what distinguishes this approach. You are > moving the screenshot processing to the server side, correct? If you > have time please clarify.
We wait a long time for the server side pixmap to be copied to the client side. If we scaled in the server side instead, we transfer a much smaller pixmap. We may save as well some implicit format conversion that may happen inside gtk. > Have you tested the change on an XO running a 8.2 release candidate? You mean scaling in the server? I wrote the code attached to the ticket about one year ago and then ran out of time before managed to have it working reliably. May have been a bug in the driver (that may have already been fixed), or just me not getting what X hacking is about. If someone could give it a try, may be worth it. Regards, Tomeu _______________________________________________ Sugar mailing list [email protected] http://lists.laptop.org/listinfo/sugar
