Anonymous person... I like that monicar lolTurns out devs who write web based email systems never read the RFC for SMTP so they all reject my other email address ^_^@xganon.com lol.
On 2025-12-10 14:33, Roland Clobus wrote:
Hello anonymous person, On 10/12/2025 17:15, donotemailme@xganon.com wrote:Can someone explain why uefi boot is not a valid option when telling it to build an hdd image?Because the focus for the live images have been on ISO files and generating other formats such as hdd images or tarballs has no test coverage.There are attempts to support UEFI for hdd images, but that needs a suitable test coverage before it can be accepted. See https://salsa.debian.org/live-team/live-build/-/merge_requests/429
Ok, i have a fully working function in the dev branch of my repo, in the file make_image.sh look for build_img(). This uses the contents of <live-build>/binary to build a hybrid legacy/uefi disk image file.
I tend to test on my private cloud using a vm setup for legacy boot and one setup for uefi. I then will write it to an sd card and/or usb thumb drive and boot it on a hacked chromebook where I replaced the firmware with open source firmware.
It has worked on every test Ive done.
I would also note the following The documentation suggests to cp iso to /dev/sdX Then add a partition to /dev/sdX for persistent storage.The test coverage for the persistent feature is rather minimal.
I ran into that, but I just designed a custom dialog menuing system that is inside a custom initramfs to manage all the persistent storage setup and Luks2 functions. As there is no networking setup in that initramfs so it allows for setup and key management in an air gapped environment.
This does not work. If you cp or dd the iso to the block device the block device is identified as an ISO and you can not add a partition to it. If you partition the block device first then cp or dd the iso to the / dev/sdX1 it sometimes works and boot and other times it fails to find the uefi partition. Not exactly surprising as writing partitions to a partition is a very edge case thing.There are plans to create installation tests (meaning the installation of the ISO file onto a USB stick, not the installers from the live system). They will use other (well known) tools such as dd, Rufus, Ventoy and BalenaEther to create USB media.That will take some development time. Do you volunteer?
I would be interested in testing as time permits. Your welcome to reuse any code in my repo if it is of any help.
I tried a lot of things to produce a usb disk where the iso and the persistent storage were both on the same device and ended up adjusting my script to partitioning out a loop device then using the files in <live-build>/binary to build out a properly laid out img file that can be written to a usb stick and boot legacy/bios or uefi. This is all done outside of live-build.If your computer does not require legacy/bios boot, but is able to boot with UEFI, it is sufficient to copy all files from the ISO image onto your USB stick.Additionally, the mailing list archive shows that several other people attempted this and succeeded.
Ive done a lot of testing. Mentioned earlier.
Is live-build still being actively maintained?I'm sorry, but I feel offended... Take a look at the mailing list archive and the git history.
Sorry I didn't mean to offend. I did not have the link to the git repo. Ive book marked it.Im still new to Debian, I spent over 35 years working in Unix/Linux and most of the linux has been RHEL. Now that I am retired I get more time to expand and play with other things. To put it into perspective I ran Minix 1.5 on a 286, worked managing SCO Unix, AT&T Sysv, BSD, SunOs 4.1.4, then Solaris and HP-UX, moved into linux and ended up working only with RHEL in corporate and DOD subcontractor environments.
IMHO, Debian is nicer and more reminiscent of Slackware, at least what I feel it would have evolved into.
Is there a place I should report problems?You are already in the right place. This mailing list and the Debian bug tracker are the primary sources regarding live-build.Apart from reporting problems, it is highly appreciate to also report solutions.With kind regards, Roland Clobus
I can definitely report solutions.Though right now Im working on integrating a smartcard reader into an initramfs that can be used to validate a chain of trust boot using a JCOP4 openpgp smartcard. So the image my live-build scripts produce can be cryptographically validated on top of the SHA512 hash validation it does. When finished it should also use challenge/response from the card to login and unlock the luks persistent storage.
Thank you for the response. You can email me directly if you want to know who I am. ;) -- -- Creator of GhostDisk https://github.com/cyberlink1/GhostDisk
Attachment:
0xBE355809.asc
Description: application/pgp-keys
Attachment:
signature.asc
Description: OpenPGP digital signature