On 08/31/2012 09:18 AM, Michal Suchanek wrote: > On 31 August 2012 08:24, Jeremy Huddleston Sequoia > <[email protected]> wrote: >> Simon, >> >> Do you think you'll be able to come up with a fix for this regression >> tomorrow (ie before Saturday)? If not, we should revert the change from >> master before we tag 1.13 on Tuesday. I'll also revert it from >> server-1.12-branch if a fix isn't found. >> > According to the bug description this is a problem with some drivers > but not others. > > It looks like drivers can't agree if there should be this additional > pixel or not.
I have no idea how drawing XPMs and Bresenham lines are related, but if you could point me to the code, I could try to investigate. The alternative would be to draw the lines as before, but have an additional conditional in the error pixel section, which will prevent the overflow I addressed. However, that will have some (small) performance impact. cheers simon
signature.asc
Description: OpenPGP digital signature
_______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
