Topics of interest until bookworm is released (12.0)

Things that we should keep an eye on (not necessarily fixing all of them), with the 12.0 goal in mind.

RC 2

  • cryptsetup upstream fighting for the memory cost thing [RC2]
  • partman-efi vs. (plain) LVM on UEFI systems (#1033913) [RC2]
  • udeb tweaks in src:linux [RC2]
  • base-installer vs. zstd (bwh, ema) [RC2]
  • disabled os-prober (#1033985) [RC2]
  • enable <codename>-updates before stable is released (#991947) [RC2]

RC 3

  • kernel hang on ppc64el, linux regression (#1033058) [RC3]
  • missing modesetting_drv.so (#1035014)
  • hostname=unassigned-hostname (#1035349)
  • live vs. LUKS (#1035360) [confirmed in weekly builds]
  • kernel size on arm64 (#1035398)
  • fun with file descriptors, for firmware license acceptance (#1033921)
  • X vs. default driver under QEMU+UEFI (#1036019) — workaround included
  • support for powers-of-two units (#913431)
  • grub-installer patches (pham)

RC 4

  • udeb += staging r8188eu (#1035824) — if src:linux gets uploaded
  • “dosfstools for EFI partition”
  • installation-guide unblock request
  • final l10n uploads

RC 5

  • X vs. default driver under QEMU+UEFI (#1036019) — real investigation
  • debian-cd: check/refresh firmware sof hardcoded list
  • debian-cd vs. firmware: https://bugs.debian.org/1036828
  • investigate memory size requirements
  • investigate broken pipes in debootstrap, mostly seen with netboot-gtk and encrypted LVM (on 2023-05-25, before RC 4)

Others

  • snapshot status (missing dinstalls)
  • hw-detect patches for even more firmware support: #1033679 (pham) → post-bookworm
  • hw-detect patches for better external storage support: #1029962 (pham) → post-bookworm, but might be backported in some point release
  • possible race condition between “firmware install+module modprobe” and NIC checking (#1034801)
  • missing RST support (#1035101)
  • spaces in filenames, the Raspberry edition (#1029843)
  • cros/efi as a subarch
  • memory requirements vs. TFTP booting (#1035854): the dwarves update seems to have helped (-20M for mini.iso on amd64 and arm64 after 6.1.27-1 was built against it)

(Thanks to @zeha for the checklist-in-an-issue idea.)

Edited by Cyril Brulebois