pack-licenses.sh packs the image's license tree into one tar.gz and
removes the tree. It piped tar into gzip under /bin/sh (dash, no
pipefail): the pipeline's status is gzip's, and tar's is lost. The
check that follows proves only that the bundle holds license.manifest.
So a member tar cannot read leaves a bundle without that text, the
build green, and the tree removed, the only copy of the text with it.
tar now runs the compressor itself (-I 'gzip -9 -n'), so its exit
status covers the read and the write. A failed pack removes the
partial bundle and the script refuses, with the tree left for the
build log; pack_licenses in forgefirm-image.bb fails do_rootfs on the
refusal. The bundle's bytes do not change: tar feeds the same stream
to the same gzip, and a tree packed by both forms gives one sha256.
Proof: forgetest-ci "Unit tests" (`python -B -m unittest discover -s
tests -v` in forgetest), on Ubuntu 24.04 with python 3.12 as an
unprivileged user: test_image.py's PackLicensesTests gain
test_a_license_text_tar_cannot_read_is_refused_and_the_tree_left (a
text under zlib/ with mode 0: a nonzero exit, a message on stderr, no
bundle, the manifest left); the class's 5 tests OK. Negative control
on a scratch copy: the new case against the script without this change
fails at assertRefused with an exit status of 0. At the shell, on a
fixture tree with the unreadable member, the script without this
change exits 0, leaves the bundle without the text and removes the
tree; with it, the script exits 1, leaves no bundle and leaves the
tree. On a readable tree both forms write byte-identical bundles (one
sha256 for both). pyflakes reports no undefined name.
Acceptance: every catalog test. pack-licenses.sh is under
meta-forgefirm, whose content hash is part of the platform in every
fingerprint, so the image that carries this change needs the whole
catalog on the bench.
The host proofs of image.license-bundle's pack half and of
image.network-boot's interfaces check. No catalog test is removed here;
suite/image.py is unchanged.
pack-licenses.sh, next to forgefirm-image.bb, is the license step.
pack_licenses calls it with the common-licenses directory and the
output tar path (forgefirm-image-dev.bb requires the release recipe, so
both images run the one call), and do_rootfs's file-checksums name it.
Before it writes the tar or removes the tree, it exits non-zero with a
message on stderr when the directory is missing, when the manifest is
missing, or when the manifest lacks any of the newline-terminated lines
"RECIPE NAME: forgectrl", "RECIPE NAME: grblhal-glowforge" and "RECIPE
NAME: kernel-module-glowforge". On success it packs as before: tar
--sort=name --mtime=@0 --owner=0 --group=0 --numeric-owner, gzip -9 -n,
the member common-licenses/license.manifest, and the loose tree removed
once the bundle holds the manifest.
test_image.py's PackLicensesTests drives that script on a fixture: a
missing directory, a missing manifest, and a manifest missing each one
of the three lines are refused with a message on stderr and no tar, the
tree left; a manifest naming all three packs, the tar holds
common-licenses/license.manifest, and the tree is gone.
NetworkBootTests reads the BSP layer's interfaces file from the
meta-openglow checkout beside forgefirm's (forgetest-ci checks out its
scarthgap branch there) and finds no line that starts a DHCPv6 client.
A missing file fails the case.
meta-forgefirm's content hash moved (forgefirm-image.bb and the new
script are layer content, and the platform identity of
forgefirm-image-manifest.bbclass leaves out only *.md, *-pin.inc and
forgefirm-release.inc): the next image cannot inherit a campaign.
Proof: python3 -B -m unittest discover -s tests -p test_image.py on
Linux (Ubuntu 24.04, GNU tar 1.35, meta-openglow at f012d8e): 14 OK, the
five new cases among them; pyflakes names no undefined name.
ScottW514