E.g. it references a version of squashfs-tools (1:4.2+20130409-2) that we probably don't use anymore.
build-manifest-extra-packages.yml: remove squashfs-tools version we don't use anymore (refs: #12684).
Apparently our apt-get/debootstrap wrapper tricks are enough to detect the
version of squashfs-tools we actually install and use: I see it in
tails-amd64-3.0.build-manifest, while it's only installed by lb_binary_rootfs..
#1 Updated by intrigeri over 2 years ago
- Status changed from Confirmed to In Progress
- Target version changed from Tails_3.1 to Tails_3.2
- % Done changed from 0 to 10
- Feature Branch set to bugfix/12684-update-build-manifest-extra-packages
Indeed, at first glance it seems we have a problem here:
reprepro-tagged-snapshots@apt:~/repositories/3.0$ find . -name squashfs* ./debian/pool/main/s/squashfs-tools ./debian/pool/main/s/squashfs-tools/squashfs-tools_4.2+20130409-2.dsc ./debian/pool/main/s/squashfs-tools/squashfs-tools_4.2+20130409-2_i386.deb ./debian/pool/main/s/squashfs-tools/squashfs-tools_4.2+20130409.orig.tar.xz ./debian/pool/main/s/squashfs-tools/squashfs-tools_4.2+20130409-2.debian.tar.xz ./debian/pool/main/s/squashfs-tools/squashfs-tools_4.2+20130409-2_amd64.deb
OTOH we don't include packages coming from our custom APT repo in tagged snapshots, so as long as we use a custom
squashfs-tools package, we're good. And even once we switch to a
squashfs-tools from Debian, we're good as well since the version we really use is actually identified by our
apt-get wrapper, and included in the build manifest:
- arch: amd64 package: squashfs-tools version: 1:4.2+20130409-2 - arch: amd64 package: squashfs-tools version: 1:4.3-3.0tails4
I think this demonstrates that the "pulled by lb_binary_rootfs, outside of the reach of our apt-get wrapper" comment and the rationale for 0dcf467e25dda7d2ee33142806e45ca236b35927 are actually wrong: we wrap
apt-get before live-build creates the chroot, so it'll save whatever packages are installed in there, including those installed outside of the chroot copy (
chroot/chroot), e.g. by
lb_binary_iso. I admit I'm afraid I might be missing something though: I'm pretty sure these changes did fix a real problem when I applied them a year ago, but OTOH I don't think reverting them can break anything on the short term, as we'll likely keep using a custom
squashfs-tools for a while, so I'll go ahead.
#3 Updated by intrigeri over 2 years ago
- Assignee changed from intrigeri to anonym
- % Done changed from 10 to 50
- QA Check set to Ready for QA