[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#34074: xdm login loop failure has happend to a newbie



In article <199903032136.WAA08104@ludwigV.sources.org> you write:
>> Users are going to experience many many problems installing xdm on
>> their machine.  It is setup to start automatically (when there's a
>> valid parsable XF86Config file, I believe).
>
>Yes, if there is no XF86config, xdm is smart enough to give in.
>
>>  This can lead to the
>> notorious "infinite X11 reset loop" which is very very frustrating,
>> confusing, and infuriating for users.
>
>OK, xdm cannot run, so people cannot log in (we assume we talk about home 
>machines without a network access) to fix it. They don't know CtrlAltF1 
>either. Correct? May I suggest to document at least the CtrlAltF1 so they can 
>go back to a text console?

Just my two cents:

If a user is upgrading from Hamm to Slink, and has a XF86Config that
*appears* to be valid be isn't (maybe xdm has been disabled because of
problems with starting X), then updating xbase (and hence xdm) will
break the system.

eg: I have disabled xdm because my Super VGA monitor died, and I have
been forced to replace it with a VGA monitor. I don't know if installing
the latest version of xdm will automatically start it straight away, but
either way, I don't want X trying to display 1024*768 on my monochrome
VGA monitor.

Before I upgrade to slink, I will move XF86Config to something like
XF86Config.bad before hand (or just reconfigure xdm not to use the local
screen), but other people may not realize that.

BTW: If you do encounter the "infinite X11 reset loop", pushing
Ctrl+Alt+F1 will not help, as X will quit and reload before you get a
chance to do anything useful, hence switching the active TTY back to
that of the X screen. Network logins will work though, if you have a
LAN... I remember that all this was discussed some time ago on
debian-devel, and I am surprised that I still hear it to be a problem.
The last I heard was the xdm has configuration resources to delay
restarting X if an error occured, but they didn't work as documented.

However, I understand that it is too late to do anything about this
for Slink now :-(

- --Boundary_(ID_9AuLaKIndFI2ZxJk3d9jAg)--

------- End of Forwarded Message






Reply to: