On Fri, Apr 30, 1999 at 10:51:49AM +0300, Fabrizio Polacco wrote: > On Tue, Apr 27, 1999 at 01:09:39AM -0400, Branden Robinson wrote: > > > > I continue to believe > > that the default configuration for xdm (autostart on boot and manage :0 on > > vt 7) is the most sensible one, being the most popular configuration. > > I don't understand why the xdm behaviour was modifyed from hamm's "first > vt available" to a fixed vt7. This unexpectedly breaks systems where > /etc/inittab has vt7 spawning getty. They break without any advice and > at a later moment than the upgrade, as it shows up only after reboot. > It happened to me twice, and the second time made my daughters cry as > their system was unusable after a power down and they had to wait friday > for me coming back home. I have explained this time and time again. For many people, xdm grabs the "first vt available", but unfortunately getty hasn't yet claimed it, and xdm gets vt2. So you get the following problem, exactly: > The symptom of the problem is: after reboot, xdm starts and does no more > accept keyboard input, so no vc anymore, nor Ctrl-Alt-Bs. If your' not > on a net you have to power down and the reboot in single user. Yep. [...] > But the problem is now in slink and we can only put out notes on how to > face it. If I had discovered this before slink release I would have > raised a grave bug. > I wander what a newbie encountering this problem can say. "I wish xdm grabbed vt's in a deterministic fashion." Well, maybe, not but that is the solution. Until and unless we can find a way to tell xdm to wait for anything that's going to grab a VT at boot time to do so. And that's a really hard problem to solve. -- G. Branden Robinson | There's nothing an agnostic can't do Debian GNU/Linux | if he doesn't know whether he believes branden@ecn.purdue.edu | in it or not. cartoon.ecn.purdue.edu/~branden/ | -- Graham Chapman
Attachment:
pgpj1XTU7laLO.pgp
Description: PGP signature