On Wed, May 29, 2002 at 12:29:41AM -0400, Joey Hess wrote: > Thanks for the detailed report. > > Paul L. Rogers wrote: > > o Since this is a X window system/desktop environment > > install I would expect installation to complete with > > starting gdm/kdm/xdm. Perhaps a note should be added to > > the "Have fun!" message suggesting a reboot or how to > > manually get the system into the same state as a reboot > > would. > > Branden, what should I do here? Is there some reason xdm's postinst > doesn't start xdm? Yes. The postinst attempts to make an intelligent decision about whether or not to start the daemon. It will, but not if: # we are upgrading without stopping it # we're currently in X on the display it attempts to manage by default # it's already running # the options file says not to See /var/lib/dpkg/info/xdm.postinst for a breakdown of the logic. Note that the options file does say "not to" by default. In my opinion, it is good to not start display managers by default on installation. Most daemons don't immediately start an X server and force a VT switch on you, but display managers do. I think novice users would find this behavior surprising in the midst of a system update. > > o As part of the "Configuring Xserver-xfree86" process, I > > took the default <Yes> when asked "Use kernel > > framebuffer device interface". X would not start since > > the framebuffer drivers for the Matrox G200 video card > > are not included in the kernel that is installed as part > > of the installation process (vmlinuz-2.4.18-bf2.4). It > > would seem to me that at least on the Intel platform, > > <No> would be a better default. Perhaps more specific > > guidance than: > > In theory, either approach should work, but in > > practice, sometimes one does and the other does > > not. Enabling this option is the safe bet, but > > feel free to turn it off if it appears to cause > > problems. > > In my case, enabling this option was NOT the safe bet. > > Leaving this for Branden. The default was too aggressively set to "yes" in versions of the XFree86 packages prior to 4.1.0-17, which is now in unstable but not in testing. It is my intention to upload 4.1.0-17 to woody-proposed-updates as soon as I'm clear on how to do that. > Nope, it's too late for this kind of thing to go in. > > > o The default /etc/X11/XF86Config-4 file includes a > > reference to the device file "/dev/input/mice". When X > > is started, xf86OpenSerial reports as an error "Cannot > > open device /dev/input/mice No such device." (in > > /var/log/:0.log). No big deal though. > > Leaving for Branden.. If your mouse works, this is harmless error. The default server config file sets up two mouse devices to make life easier for people with USB mice. Failure to open a non-core pointer device is not fatal. -- G. Branden Robinson | Convictions are more dangerous Debian GNU/Linux | enemies of truth than lies. branden@debian.org | -- Friedrich Nietzsche http://people.debian.org/~branden/ |
Attachment:
pgpS8oOxwx6ZE.pgp
Description: PGP signature