Comment 30 for bug 671027

Attaching a patch which I have tested extensively which restores the -efikamx subarch and also makes the script work to our basic specifications. I did capitulate to the new "boot.script" method but we can properly handle kernel versions. It is amazing what a little bit of 'sed' can do.

Also fixes
 * No uInitrd would be generated if one didn't already exist. The check has been changed to check for an existing initrd.img of the correct version, and then if a uInitrd of that version already exists, back that one up
 * Restores the kernel description 'comment' for uImage generation
 * Lots of anal retentive curly brackets
 * Picks a rootfs properly but this is a bit of a hack as I don't understand the initramfs hook or the difference between the default and override (I guess, my script will explode if run from a udeb since /dev/root is the livecd and not the target)
 * Stop re-compressing the uInitrd. It is already gzip compressed, so compressing it again in from mkimage is redundant
 * Needs a /boot/boot.script which is NOT shipped in this patch

If you could apply this to the Maverick flash-kernel I would much appreciate it. For Natty we need to think bigger, or we can fix up this patch some more.

One thing I am really concerned about is where the proper place to ship the boot.script source for generating the boot.scr - is this meant to be in some hwpack metapackage or something? How do we assure it's installed?