Graphical install and subsequent X configuration on some Mac hardware is broken.
This helps with the G3 PowerBook I tested, but we still don't ever seem to parse the server's output and get the panel size right. --- xserver.py.orig 2004-04-15 17:23:32.000000000 +0100 +++ xserver.py 2005-02-20 15:31:38.000000000 +0000 @@ -121,8 +121,8 @@ attempt = 'PROBED' # use the framebuffer server by default on ppc - if rhpl.getArch() == "ppc": - attempt = "FB" +# if rhpl.getArch() == "ppc": +# attempt = "FB" orig_hsync = monitor.getMonitorHorizSync() orig_vsync = monitor.getMonitorVertSync() @@ -199,6 +199,8 @@ probeflags=XF86HW_PROBE_NONE) hwstate.set_resolution(test_resolution) hwstate.set_videocard_card(card) + if rhpl.getArch() == "ppc": + hwstate.videocard_options.append(("UseFBDev", "True")) testx(hwstate, mouse, keyboard, logfile=logfile, xStartedCB = xStartedCB, xQuitCB = xQuitCB)
*** Bug 121265 has been marked as a duplicate of this bug. ***
*** Bug 136173 has been marked as a duplicate of this bug. ***
*** Bug 136168 has been marked as a duplicate of this bug. ***
*** Bug 136044 has been marked as a duplicate of this bug. ***
Users affected by this may want to use the Xautoconfig tool from http://www.bytebot.net/geekdocs/ibook/Xautoconfig-0.22-1a.ppc.rpm to automatically configure the X server.
Note that the ATI driver doesn't recognise the 'UseFBDev' option, and silently ignores it. It _ought_ to recognise and honour it though -- as it doesn't cause any harm at the moment I suspect we should continue to give it. We should probably revamp the screen detection code when the DDC probe fails too -- if there's a framebuffer we can use the ioctls on /dev/fb0 to get our screen mode. Xautoconfig has an example of this.
*** Bug 151380 has been marked as a duplicate of this bug. ***
*** Bug 155483 has been marked as a duplicate of this bug. ***
We try to do something along this line of thought now, but it's still not 100%. :/