Thank you for the discussion! I found it educational and motivating, and hope that everyone found the same. It's really refreshing to see this. Sorry for the naivety in my analysis and in the delay replying; I've been drained/busy, but I read each of your emails carefully, wanted to reply to each of them, and chose not to CC everyone because that would be noisy. Steve, thank you for the quick fix :-) Reply follows inline: Steve McIntyre <steve@einval.com> writes: > On Sat, Dec 04, 2021 at 10:42:29PM +0000, Steve McIntyre wrote: >>On Sat, Dec 04, 2021 at 11:37:28PM +0100, Pascal Hambourg wrote: >>>Le 03/12/2021 à 22:08, Nicholas D Steeves a écrit : >>>> >>>> >>>> c) parse /target/etc/fstab, and attempt to mount other partitions >>> >>>The rescue system already offers to do it for separate /boot and /boot/efi, >>>so why couldn't it do for separate /usr too ? >> >>That's exactly what I'm about to do. > > In fact, it needed more work than that - the code chrooted into > /target and ran mount there. That didn't work for a separate /usr. But > I've refactored the code and made things work more cleanly inside d-i. > I had also attempted a fix at the same time, but my approach wasn't nearly as nice as yours. In particular, I appreciate how your fix produces insight into debian-installer/rescue function, architecture, and style, and I hope to say thank you more substantially with an eventual MR for btrfs subvol support :-) It will be much less ugly and "tacked on" as a result of studying your changes! Regards, Nicholas
Attachment:
signature.asc
Description: PGP signature