Comment 10 for bug 693671

bcbc (bcbc) wrote :

@Barry, from my understanding, wubildr.mbr is grub4dos's bootloader, and it looks on the root of all partitions until it finds the 'rest of itself': the wubildr file. (Standard grub4dos is the grldr.mbr and grldr but you can rename these, as has been done for wubi).

The "Try (hd0,0): <file system>: " is grub4dos' standard output that it gives as it tries each partition, looking for the wubildr in the root of that partition. If it doesn't find the wubildr it outputs "No wubildr" and moves on to the next partition "Try (hd0,1):... etc.

The fact that it hangs there... this is known to occur when the partition's file system type is ext4 (because the version of grub4dos Wubi uses is old)... but in this case it appears to be something to do with the wubildr itself. If you replace it with one from a wubildr from a fresh Maverick install, it continues to start the installation.

I don't know a whole lot about wubildr, but I guess it chains grub4dos to grub2 with the core.img and some grub2 instructions to find the root.disk, loop mount it and load the grub.cfg. You can see how it gets created in the script grub-mkconfig (the version in package lupin-support), but this rebuilt wubildr doesn't contain all the grub2 instructions to support the initial installation from the windows host. It's missing this part:
if [ ${show_panic_message} = true ]; then
    if search -s -f -n /ubuntu/install/boot/grub/grub.cfg; then
        if configfile /ubuntu/install/boot/grub/grub.cfg; then
            set show_panic_message=false