Bug 11587

Hi,

First off, it would be nice if there was a way to dump the errors from an
install to a floppy disk on failure. Writing down the error and carrying it
across the room to type it back in is a rather primitive method of doing
things.

lilocheck is failing with the following error:

File: "/usr/lib/anaconda/kickstart.py", line 84, in doLiloCheck drives =
isys.hardDriveList ().keys() AttributeError: hardDriveList

I am going to guess that someone typo'd here, since this is what I see when
I cd to the anaconda dir on my updated (patched) cd directory and

grep -i harddrive * */* 2>/dev/null
fstab.py:       drives = isys.hardDriveDict().keys()
fstab.py:       return isys.hardDriveDict()
isys.py:def hardDriveDict():
kickstart.py:        drives = isys.hardDriveList ().keys()
kickstart.py:                "harddrive"        : None                  ,
raidstart-stub:drives = isys.hardDriveDict().keys()
todo.py:        d = isys.hardDriveDict()

All the other routines use sys.hardDriveDict().

I modified the kickstart.py to test this theory.

I changed drives = isys.hardDriveList ().keys() to
drives = isys.hardDriveDict().keys()

After making that change, the kickstart reboots the system without any
message as to why. I am guessing that is what it was supposed to do, but a
message of some kind, like: "LiloCheck: Lilo detected in boot sector,
probable existing install, kickstart aborting." would be helpful.

On a slightly more advanced note:

Is it possible for the kickstart to simply boot off the hard drive at that
point, instead of rebooting? The reason I am asking is that I have built
custom CD's that contain a ks.cfg in the initrd.img in the etc directory,
and the default in the syslinux.cfg is to boot the kickstart. If the
lilocheck routine could switch the boot to the hard drive, the CD could be
left in the drive after the install without having to switch the boot
device, and the installed rpm's could be verified against the files on the
CD.


Thanks,
Sean Utt
sean