Closed tdewey-rpi closed 6 months ago
Wait, what? I can't see why this has brought in those other two commits....
Paging @XECDesign for a view.
Confirmed I see this more gracefully handling fatal errors
Is off_chroot
needed?
I believe the unmount "${WORK_DIR}/${STAGE}"
line should already do everything the off_chroot
function would've done.
The on
in on_chroot
maybe should've been in
. The idea is that you pipe commands into it and it runs in the chroot environment.
Or have I missed something?
Is
off_chroot
needed?I believe the
unmount "${WORK_DIR}/${STAGE}"
line should already do everything theoff_chroot
function would've done.The
on
inon_chroot
maybe should've beenin
. The idea is that you pipe commands into it and it runs in the chroot environment.Or have I missed something?
I found without this, I was still getting sub-mounts mounted on abort - which then caused trouble when I tried delete the workdir and re-try my build.
RE: naming:
Completely my fault for misreading. Happy to rename.
This version, which also includes #771, works for me without off_chroot
with failures introduced in various sub-stages and also in export-image while the img file is attached.
Rebased off the master
branch, since that's what gets merge into arm64
The current git history is just for testing and clarity, so I've converted it to a draft for now
Also, thanks! This has been a long overdue addition.
@XECDesign Yours to merge, I lack the powers for now.
I was waiting to hear back from #771. To keep a sensible git history, I want to squash those commits first, then rebase and squash this.
Will be sorted a bit later today.
Add a function to unmount the bindmounts in the chroot, and unmount the image and stage in progress, in the event of a trap.