Totally, me too. Sadly there’s a chunk missing for a Silicon Mac machine. Notably the stuff about case sensitivity and the bin formats. Still haven’t got it built - tar of all things is misbehaving. I’ve tried switching to BSD tar with no improvement (quite the opposite, it’s missing an option that’s needed).
GNU tar gives this: “Directory renamed before its status could be extracted”. That’s after about an hour’s run of the make file, annoyingly.
Onwards, or perhaps downwards into this spiral.
EDIT: Okay, the tar thing seems to be fixed by adding --absolute-namesto the invocations in config/functions/build-rootfs. Some sort of interaction between GNU tar and the Mac filing system, if other people’s attempts to build things on Macs are anything to go by.
I’ll write this up as a single top-level post if and when it actually builds.
EDIT: nope, sad to say, I’m giving up for now. The Mac filesystem is putting up too much of a fight - extended attribute support is now causing issues. One last try, without the bind mounts, so everything is “inside” Docker.
EDIT: image created, but there’s some bodging. The final step looks like:
Info: About to run /home/khadas/fenix/build/utils-*/aml_image_v2_packer -r /home/khadas/fenix/build/images_upgrade-*/Amlogic/package.conf /home/khadas/fenix/build/images_upgrade-*/Amlogic /home/khadas/fenix/build/images/vim4-ubuntu-24.04-minimal-linux-5.15-fenix-1.7.5-260216-emmc-develop.img
qemu-i386: unable to find CPU model 'cortex-a53'
(I added the info line to show me the raw command line it was struggling with). If I run the command manually then it successfully creates an image, I’m guessing since QEMU_CPU is not set.
Final EDIT: I’ll leave all this up for posterity, but I’ve not managed to make a bootable image. I’m sure I’m doing something dumb (using a Silicon Mac to build a VIM4 image primarily).