The automake test harness[1] already dumps the contents of
`test-suite.log` when `VERBOSE=1` is set, so we don't need to add
special handling for it.

1. https://www.gnu.org/software/automake/manual/html_node/Parallel-Test-Harness.html

This refreshes the build dependencies installed for the GitHub Tests
workflow based on the Build-Depends in the upstream packaging. The
handling is now more explicit about any deviations and any release
differences.

Rather than use the release codename tags, use the release stage tags.
This way the configuration (theoretically) doesn't need to be updated
when new Debian and Ubuntu releases are made.

For Debian stable is used instead of buster and a testing (bullseye)
build is added. For Ubuntu, latest is used instead of focal for the
current LTS and rolling is used instead of groovy for the latest
release. This actually changes the Ubuntu build from groovy to hirsute.

Since #2377 was merged, this is in libtest-core.sh, which is sourced by
libtest.sh.

Signed-off-by: Simon McVittie <smcv@collabora.com>

libtest.sh: Remove duplicate ERR trap and report_err()

This tweaks `ostree_repo_write_commit` so that it checks for the
envinroment variable `SOURCE_DATE_EPOCH` as a way to override
the current time, which is used as the commit timestamp.

Ref: https://reproducible-builds.org/docs/source-date-epoch/
Ref: https://reproducible-builds.org/specs/source-date-epoch/

lib/commit: respect SOURCE_DATE_EPOCH for commit timestamp

The `systemd-rfkill.*` service falls in the category of early things
that need write access to `/var`, so we need to make sure we run before
or it might hit the read-only sysroot.

The long-term fix for this is
ostreedev/ostree#2115.

Closes: coreos/fedora-coreos-tracker#746

ostree-remount: Order before systemd-rfkill.*

Prow is not managing merges on this repo.

OWNERS: remove

This fixes the ci-release-build.sh script to directly source
and evaluate 'package_version' from its m4 definition, without
requiring a fully configured source tree.

ci/release-build: evaluate package_version from m4 definition

Release 2021.3

A statoverride file written in the obvious way will produce incorrect
results for two independent reasons.  Document them.

man: improve statoverride description

Currently the verifier only imports all the GPG keys when verifying
data, but it would also be useful for inspecting the trusted keys.

In order to use the GPG verifier, it needs to be seeded with GPG keys
after instantation. Currently this is only used for verifying data, but
it will also be used for getting a list of trusted GPG keys in a
subsequent commit.

Currently the verifier decides whether to include the global keyrings
based on whether the specified remote has its own keyring or not. Allow
callers to exclude the global keyrings even when that's not the case.
This will be used in a subsequent commit in order to get the GPG keys
only associated with a remote.

This function enumerates the trusted GPG keys for a remote and returns
an array of `GVariant`s describing them. This is useful to see which
keys are collected by ostree for a particular remote. The same
information can be gathered with `gpg`. However, since ostree allows
multiple keyring locations, that's only really useful if you have
knowledge of how ostree collects GPG keyrings.

The format of the variants is documented in
`OSTREE_GPG_KEY_GVARIANT_FORMAT`. This format is primarily a copy of
selected fields within `gpgme_key_t` and its subtypes. The fields are
placed within vardicts rather than using a more efficient tuple of
concrete types. This will allow flexibility if more components of
`gpgme_key_t` are desired in the future.

This provides a wrapper for the `ostree_repo_remote_get_gpg_keys`
function to show the GPG keys associated with a remote. This is
particularly useful for validating that GPG key updates have been
applied. Tests are added, which checks the
`ostree_repo_remote_get_gpg_keys` API by extension.

This will be used to implement the PGP Web Key Directory (WKD) URL
generation. This is a slightly cleaned up implementation[1] taken from
the zbase32 author's original implementation[2]. It provides a single
zbase32_encode API to convert a set of bytes to the zbase32 encoding.

I believe this should be acceptable for inclusion in ostree. The license
in the source files is BSD style while the original repo LICENSE file
claims the Creative Commons CC0 1.0 Universal license, which is public
domain.

1. https://github.com/dbnicholson/libbase32/tree/for-ostree
2. https://github.com/zooko/libbase32

Calculate the advanced and direct update URLs for the key discovery
portion[1] of the OpenPGP Web Key Directory specification, and include
the URLs in the key listing in ostree_repo_remote_get_gpg_keys(). These
URLs can be used to locate updated GPG keys for the remote.

1. https://datatracker.ietf.org/doc/html/draft-koch-openpgp-webkey-service#section-3.1

If the key UID contains a valid email address, include the GPG WKD
update URLs in GVariant returned by ostree_repo_remote_get_gpg_keys().

Update to upstream version '2021.3'
with Debian dir 8860039737458a32b15fcdbfddb9cd869e19da99

Otherwise, some tests will think we're downgrading, and fail.

Some distributions set this during build in order to have reproducible
builds from the same source code: for example, Debian uses the date
from debian/changelog.

However, some of our tests assume that `ostree commit` will result in
a commit with the current date/time, and SOURCE_DATE_EPOCH breaks that
assumption. Unset it for our build-time tests.

Resolves: ostreedev/ostree#2405
Signed-off-by: Simon McVittie <smcv@collabora.com>

tests: Unset SOURCE_DATE_EPOCH

Pointer command is dangerous if there is no assignment.

Log: Avoid wild pointers

fix: Avoid wild pointers

Move the existing docs permissions stanza to the top of the workflow for
consistency.

workflows: bump lint toolchain; restrict repository access

We need to check all signatures for one which passes, not just fail on
the first one.

Reported-by: Seth Arnold <seth.arnold@canonical.com>

Reported-by: Seth Arnold <seth.arnold@canonical.com>

Reported-by: Seth Arnold <seth.arnold@canonical.com>

A couple of Coverity fixes

This explicitly checks for commit command options asking for both
non-zero UID/GID and canonical permissions at the same time,
which are incompatible.

This reduces the usage of goto cleanup logic by porting the commit
modifier pointer to autoptr.

builtins/commit: check for conflicting permissions options

builtins/commit: move commit modifier to auto-cleanup

This adds a new `OSTREE_CHECKSUM_FLAGS_CANONICAL_PERMISSIONS`
checksumming flag, which is needed in bare-user-only mode
to ignore local IDs.

This automatically enables canonical permissions for checkouts in
bare-user-only mode.

lib/checkout: use canonical permissions in bare-user-only mode

This reverts commit d18bfcc.

Recent debhelper installs units into /usr/lib/systemd, so adjust our
override to match either way.

This tweaks commit logic to detect bare-user-only repositories and
canonicalize permissions automatically.

Remote GPG key info

lib/commit: autofix permissions for bare-user-only

This fixes the logic to detect whether xattrs should be automatically
ignored when diffing.

lib/diff: ignore xattrs if disabled on either repos

As pointed out in the original review, `gpg-list-keys` fits better
alongside the existing `gpg-import`.

Changes were done with:

```
git grep -l list-gpg-keys | xargs sed -i 's/list-gpg-keys/gpg-list-keys/'
for src in $(git ls-files '*list-gpg-keys*'); do
  dst=${src/list-gpg-keys/gpg-list-keys}
  git mv "$src" "$dst"
done
```

bin/remote: Rename list-gpg-keys to gpg-list-keys

This detects bare-user-only mode and automatically enables a
commit modifier with relevant flags.

commit: automatically skip xattrs in bare-user-only mode

This fixes some aspects of OstreeRepoAutoTransaction and re-aligns
it with the logic in flatpak. Specifically:
 * link to the underlying repo through refcounting
 * bridge internal errors to warning messages
 * verify the input pointer type

This is a preparation step before exposing this logic as a public API.

lib: improve transactions auto-cleanup logic

Hit this while working on some Rust code.

This is really the standard best practice, matching how
e.g. dpkg/rpm work, as well as most local development
environments (including mine) with e.g. `toolbox`.

I was seeing an `EPERM`  here which was confusing.
It turned out the real error was `EEXIST`.

Since we're referring to the original error, but we do a
lot of computation in the middle, we need to save errno.

lib: Change read_commit_detached_metadata to be nullable

ci: Run GH action CI build+test as non-root

Followup to PRs related to ostreedev/ostree#2410

Since the test suite now covers this the test was failing on
a Fedora SELinux enabled host where we see `security.selinux`
even if not in the commit.

checkout: Save errno when re-throwing

checkout: Also ignore xattrs for union in bare-user-only mode

We have a bunch of APIs to do GPG verification of a commit,
but that doesn't generalize to signapi.  Further, they
require the caller to check the signature status explicitly
which seems like a trap.

This much higher level API works with both GPG and signapi.
The intention is to use this in things that are doing "external
pulls" like the ostree-ext tar import support.  There we will
get the commitmeta from the tarball and we want to verify it
at the same time we import the commit.

Add an API to verify a commit signature explicitly

It cannot work to use `--no-xattrs` when SELinux is enabled
because we get a `security.selinux` attribute on created files
regardless.  So just skip this test if true.

Also add some `ostree fsck`s in here which helped me debug
this.

This avoids possible issues when trying to chmod, tweaking
permissions instead.

There are some existing issues around fsck in unprivileged bare mode,
so this test does not really work at the moment. Leaving it as a FIXME
for the moment.

tests: fix bare mode unprivileged 'make check'

We're waaay overdue for this, it's been the default
in rpm-ostree for years, and solves several important bugs
around not capturing `/etc` while things are running.

Also, `ostree admin upgrade --stage` (should) become idempotent.

Closes: ostreedev/ostree#2389

for deltafiles the legacy_transaction_resuming flag is not used,
which will mark the commit as done, even if files are missing.
using already existing commitstate_is_partial function as fix

upgrade: Stabilize deployment staging