On Tue, Jan 31, 2006 at 12:27:27PM +0530, Ajith Vargese Thampi wrote: > > Yes, tty2 problem is at stage2. I have considered trying with bash, a > static bash and ash shells but to no avail. Is there any need for a udev > filesystem over devfs filesystem for /dev ? i am presuming as No, but > does it? > I am currently looking into the init code. So i should try to put in > more details. There is no need for a udev filesystem. The required tty device nodes are created by the install init process (see anaconda/loader2/devices.h and init.c) > I see that there is something called a build-locale-archive trying to be > used from anaconda. My distro does not have such a locale archive. Right. > What exactly is anaconda trying to get from locales? where does it look > for for and what comprises of a correct locale/locale archive in terms > of anaconda? maybe i can hack out little details of it? It needs locale data so that can properly display text. When you are dealing with multibyte encodings (such as UTF-8) for messages, you have to know things like how to sort characters and how wide (in screen cells) the characters take up with displayed. That information (and more) comes from the locale data. > But why does anaconda's gui mode exit abruptly? Any package > dependencies? How can i debug this problem? Normally an early crash has a stack trace. > Only if i could get tty2 to work, which would make things much simpler. Yes, getting tty2 to work is very important. Then you will be able to hit Ctrl+S as anaconda is starting on tty1 (so that it doesn't exit when anaconda crashes) and then start anaconda by hand on tty2 to see if you get more messages. Cheers, Matt -- Matt Wilson rPath, Inc. msw@xxxxxxxxx
