On Thu, Apr 10, 2008 at 12:14 AM, D. Michael McIntyre <[EMAIL PROTECTED]> wrote: > I haven't tried on other machines yet, but I'm thinking if it really is some > kind of half-a-pixel problem there's no simple solution.
I'll try to look into this. > What if we switch the thing to a diamond shape, so it's less obvious that > it's > misaligned by one pixel under certain conditions? I don't really care about > one pixel if it isn't so obvious that it's one pixel off. The diamond shape would have the blocky aspect of the little flag, as long as I don't figure out how to handle shape dithering. On Thu, Apr 10, 2008 at 2:25 AM, D. Michael McIntyre <[EMAIL PROTECTED]> wrote: > On Wednesday 09 April 2008, Yves Guillemot wrote: > > > RG revision is 8796 (I don't know if +1 is set or no). > > I distinctly see (mostly with zoom) the marker left edge aligned on the bar > > and the cursor left edge one pixel on the left. > > You probably have the +1 version. I can confirm. > I run 1280x1024, what about you two? Maybe it's an aspect ratio thing > factoring into the half-a-pixel problem. I'm running 1280x1024 too. By the way I tested it on another box, with the same result (NOK with the +1, OK without it). > Anyway, it definitely seems like we should get rid of the +1, since it breaks > for two people in order to fix it for one. We'd need some other cases to have better statistics. I for one would rather clarify how coordinates are calculated for the marker and the cursor. There may be some small differences that don't round well to integers. I'll try to debug this tonight. -- Best regards, Philippe. ------------------------------------------------------------------------- This SF.net email is sponsored by the 2008 JavaOne(SM) Conference Don't miss this year's exciting event. There's still time to save $100. Use priority code J8TL2D2. http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone _______________________________________________ Rosegarden-devel mailing list [email protected] - use the link below to unsubscribe https://lists.sourceforge.net/lists/listinfo/rosegarden-devel
