We now require Go 1.18. As of that version:
- GO111MODULE=on is implied by having a go.mod file
- -mod=vendor is implied by having a vendor directory
so just remove both options everywhere
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
gopkg.in/check.v1 hasn't had any commit since Nov 2020.
That's not a immediate issue for a test-only dependency, but
because it hides access to the standard library *testing.T,
eventually it will become limiting.
Also, using the same framework for unit and integration tests
seems practical.
This is mostly a batch copy&paste job, with a fairly high risk
of unexpected breakage.
Also, I didn't take much time at all to carefully choose between
assert.* and require.*; we can tune that as failures show up.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
Use (podman unshare) as already suggested, it is necessary for an unprivileged
user to remove the temporary c/storage state. OTOH it doesn't work with Docker at all.
Don't use the - prefix, it only works at the _start_ of a rule, not in the middle of
a multi-line shell script.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
Typically, the compat with earlier versions causes us to use
newer versions of dependencies, which can only be a good thing.
Over time, the 1.17 version reference is just going to become obsolete.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
This can happen when building RPMs out of tarballs (which don't contain
the .git repository).
To test:
> make -n /bin/skopeo; mv .git ../.git ; make -n bin/skopeo
Fixes#1707 .
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
The new file `skopeo.spec.rpkg` along with a webhook will automatically
build rpms on every PR merge on the main branch.
Run `rpkg local` or `make rpm` to generate the rpm.
Known issue: Doesn't yet build for EL8 environments.
Signed-off-by: Lokesh Mandvekar <lsm5@fedoraproject.org>
The {a,b} syntax is not POSIX compatible. The Makefile should run with
all POSIX shells so we cannot use shell specific features like this.
Fixes#1657
Signed-off-by: Paul Holzinger <pholzing@redhat.com>
As of the just-updated github.com/proglottis/gpgme 0.1.2,
the gpgme subpackage uses CGo's native #cgo pkg-config support
to find the relevant libraries, and we no longer need to manually set
CGO_CFLAGS and CGO_LDFLAGS. So stop doing that.
Note that the proglottis/gpgme update (implied by vendoring c/image)
means the minimal supported version of GPGME is 1.13.0.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
In order to meet achievable deadlines converting from Travis to Cirrus
CI, one significant artifact was carried forward (instead of fixing):
Depending on a `--privileged` container to execute all/most automated
checks/tests.
Prior attempts to remove this aspect resulted in several test failures.
Fixing the problems was viewed as more time-consuming than simply
preserving this runtime environment.
Time has passed, and the code has since moved on. This commit removes
the legacy need to execute CI operations in a `--privileged`
container, instead running them directly on the host. At the same time,
the necessary test binaries are obtained from the same container used
for development/local testing purposes. This ensures the two
experiences are virtually always identical.
Signed-off-by: Chris Evich <cevich@redhat.com>
This is the script that runs 'skopeo COMMAND --help' and
cross-checks that all the option flags are documented
in man pages, and vice-versa (all options listed in man
pages appear in COMMAND's --help message).
Copied from podman, with changes for skopeo-land (removing
the rst checks, and conforming to skopeo conventions).
Signed-off-by: Ed Santiago <santiago@redhat.com>
Add new script, hack/man-page-checker, copied from podman. Run it
in 'make validate-local' target.
This is NOT the checker requested in #1332 (verify that flags
listed in 'skopeo foo --help' are documented in man pages and
vice-versa). This is a much simpler script that merely looks
for very basic typos or discrepancies between skopeo.1.md
and skopeo-foo.1.md.
The next part (cross-checking flags) is in progress but will
require a huge number of changes to the man pages. I'm submitting
this now because it's easy to review.
Signed-off-by: Ed Santiago <santiago@redhat.com>
This partially reverts commit a81cd74734
so that most path variables like PREFIX and BINDIR refer to paths on
the installed system, and DESTDIR is used only in (make install), following
the philosophy of the GNU coding standards for path variables. (But not precisely
the variable names, which are lowercase in the standard, nor the principle that
even SYSCONFDIR should be under $PREFIX.
Keep the use of ?= instead of = because it somewhat better expresses the idea
that the values can be overridden.
Use ${DESTDIR}${BINDIR} instead of ${DESTDIR}/${BINDIR} etc, so that a plain
(make install) does not use paths like //usr/bin/... ; strictly speaking they
are IIRC reserved by POSIX, and more importantly it just looks untidy :)
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
- Bugfix `make nixpkgs` which pin with branch `nixos-21.05`
- Code lint with `nixpkgs-fmt`
- Code sync between x86\_64 and aarch64
Signed-off-by: Wong Hoi Sing Edison <hswong3i@pantarei-design.com>
In 18e6c6f17b there where changes to
better handle PREFIX and DESTDIR in the Makefile. "CONTAINERSCONFIGDIR"
was renamed to "CONTAINERSCONFDIR" without ensuring it was correct in
the install block. This results in policy.json being isntalled into
"${DESTDIR}/" which is "/policy.json" for Linux distros.
Signed-off-by: Morten Linderud <morten@linderud.pw>
- Bugfix `make nixpkgs` which pin with branch `nixos-20.09`
- Code lint with `nixpkgs-fmt`
- Code sync between x86\_64 and aarch64
Signed-off-by: Wong Hoi Sing Edison <hswong3i@pantarei-design.com>
Without this env. var. being set from hack/make.sh, many/most
integration tests will `SKIP`. Fix this by notifying the user
and setting the magic `SKOPEO_CONTAINER_TESTS` variable.
Signed-off-by: Chris Evich <cevich@redhat.com>
These tests need to operate as part of the c/image repository CI to
verify downstream-usage. That environment is already inside the
container built from the Dockerfile (here). Support this use-case by
adding 'local' targets which bypass the container build. Also,
simplify the "in-container" check to more specifically verify the exact
container image it's operating under.
Signed-off-by: Chris Evich <cevich@redhat.com>
PREFIX should default to /usr/local according to GNU coding standards.
DESTDIR should just be used for install
This will allow us to specify
make PREFIX=/usr DESTDIR=/tmp/build/ install
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
By default we should build bin/skopeo locally
and build docs locally.
Show output when doing make docs.
Add description in `make help` to explain default
behaviour.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Add a make target that cross-compiles for a handful of the possible
targets that `go tool dist list` can tell us about.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
This is conceptually consistent: First change the set of
dependencies, then update the vendored copy.
(Due to (go mod verify) afterwards, and CI running this again,
this should not make a difference in practice, so this is just
a clean-up.)
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
(export a=b command args) does not run (command args) with a=b,
it sets $a to b, and marks variables $a $command $args as exported,
i.e. (command args) is not run.
So, before https://github.com/containers/skopeo/pull/888 we were not actually
running (go mod tidy), and now we are not running (go mod vendor).
Just use $(GO), which already sets GO111MODULE=on, without the extra export.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
Add a vendor-in-container make target to allow for executing make vendor
in a golang:1.13 container. The CI is currently enforcing golang 1.13
which has a different vendoring behavior than previous versions which
can lead to failing tests as some files might be added or deleted. The
new make target will help users who are not using 1.13 to vendor their
changes.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Experimentally, this seems to help with localhost access inside that
container (but I have no idea what's the reason for that).
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
The nested podman tries to write to it. This primarily only
removes noise from logs, it does not seem to significantly change
behavior.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
This allows using the vendored dependencies instead of
searching for them in $GOPATH and elsewhere.
This does not necessarily matter for skopeo itself, but
the test-skopeo Makefile target in containers/image uses
(go mod edit -replace) to replace the vendored c/image with
a locally-edited copy; skopeo's (make check) then runs tests in
a container which does not have access to this locally-edited
copy, and since Go 1.13 this causes (go {list,test,vet})
to fail if -mod=vendor is not used.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
drop support for the ostree backend. The only known user for the
backend is the atomic CLI tool that is not maintained anymore
upstream.
Closes: https://bugzilla.redhat.com/show_bug.cgi?id=1766404
Signed-off-by: Giuseppe Scrivano <gscrivan@redhat.com>
Move signature yaml file to point at /var/lib/containers/sigstore.
Change skopeo-copy.1 to use containers-storage and docker transports
rather then atomic.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Use GOPROXY=https://proxy.golang.org to speed up fetching dependencies.
Setting it makes `make vendor` six times faster in my local env.
For details please refer to https://proxy.golang.org/.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Go 1.13.x isn't sensitive to the `GO111MODULE` environment variable
causing `make binary-local` to not use the vendored sources in
`./vendor`. Force builds of module-supporting go versions to use the
vendored sources by setting `-mod=vendor`.
Verified in a `fedora:rawhide` container.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
containers/storage needs math/bits which has been added in go 1.9, so
this is now the lowest possible go version to build skopeo. We can also
remove the GO15VENDOREXPERIMENT variable since this has been enabled in
go 1.6 per default and removed in go 1.7.
Signed-off-by: Sascha Grunert <sgrunert@suse.com>
Turn of go modules to avoid breaking build environments to accidentally
try pulling the dependencies instead of using the ./vendor directory.
Signed-off-by: Valentin Rothberg <rothberg@redhat.com>
Skopeo CI tests run under podman; hence the registries
run in the tests will be podman-in-podman. This requires
complex muckery to make work:
- install bats, jq, and podman in the test image
- add new test-system Make target. It runs podman
with /var/lib/containers bind-mounted to a tmpdir
and with other necessary options; and invokes a
test script that hack-edits /etc/containers/storage.conf
before running podman for the first time.
- add --cgroup-manager=cgroupfs option to podman
invocations in BATS: without this, podman-in-podman
fails with:
systemd cgroup flag passed, but systemd support for managing cgroups is not available
Also: gpg --pinentry-mode option is not available on all
our test platforms. Check for it before using.
Signed-off-by: Ed Santiago <santiago@redhat.com>
Currently we are only installing the skopeo.1 man page. This
change will generate and install all man pages.
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
Create a different man page for each of the subcommands.
Also replace some krufty references to kpod with podman
Signed-off-by: Daniel J Walsh <dwalsh@redhat.com>
... which has, apparently, never worked, because the golang image
has neither the GOPATH nor the working directory the Makefile expects.
Rather than move all this configuration into the Makefile to be able
to work with the golang images, just always use the skopeobuildimage
path, and only override the tags, to minimize divergence.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
Minor change to allow passing the env TESTFLAGS to make. That's pretty
convenient to filter what tests to run.
E.g. run integration tests containing the substring `Copy`:
make test-integration TESTFLAGS="-check.f Copy"
Signed-off-by: Marco Vedovati <mvedovati@suse.com>
Replace the occurrences of `github.com/projectatomic` with
`github.com/containers` to ensure clean clones of the project are
building, travis badges on the README work as expected and other minor
things.
Signed-off-by: Flavio Castelli <fcastelli@suse.com>
These targets produce a pure-Go binary, without the following features:
* ostree
* devicemapper
* btrfs
* gpgme
Signed-off-by: Akihiro Suda <suda.akihiro@lab.ntt.co.jp>
The goal is to include the c/image documentation in a skopeo release,
so that RPMs and other distribution mechanisms can ship the c/image
documentation without having to create a separate package for c/image
(which would not otherwise be needed because it is vendored in users).
So, unify the updates of the "vendor" subdirectory as (make vendor),
and document it in README.md. Also drop hack/vendor.sh, we neither
use nor document it, so updating it as well seems pointless.
The security benefits of PIC binaries are quite well known (since they
work with ASLR), and there is effectively no downside. In addition,
we've been seeing some weird linker errors on ppc64le that are resolved
by using -buildmode=pie.
Signed-off-by: Aleksa Sarai <asarai@suse.de>
On macOS, (brew install gpgme) installs it within /usr/local, but
/usr/local/include is not in the default search path.
Rather than hard-code this directory, use gpgme-config. Sadly that
must be done at the top-level user instead of locally in the gpgme
subpackage, because cgo supports only pkg-config, not general shell
scripts, and gpgme does not install a pkg-config file.
If gpgme is not installed or gpgme-config can’t be found for other reasons,
the error is silently ignored (and the user will probably find out because
the cgo compilation will fail); this is so that users can use the
containers_image_openpgp build tag without seeing ugly errors
(and without the Makefile having to detect that build tag in even more
shell scripts).
This will allow compilation with a custom go binary,
for example /usr/lib/go-1.8/bin/go instead of /usr/bin/go on Ubuntu
16.04 which is still version 1.6
Signed-off-by: Lokesh Mandvekar <lsm5@fedoraproject.org>
Vendor containers/storage, and its dependencies github.com/pborman/uuid
and github.com/mistifyio/go-zfs, which we didn't already use.
Update the build Dockerfile to install their dependencies.
Add scriptlets that try to detect whether or not we need to use the
"libdm_no_deferred_remove" and/or "btrfs_noversion" build tags.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
Run the "go" command with the $(BUILDTAGS) makefile variable passed in
as build tags. We don't currently set it, but we'll need to eventually,
and adding it now does no harm.
Signed-off-by: Nalin Dahyabhai <nalin@redhat.com>
Both (make binary) and (make binary-static) compile the code and create
a skopeo binary, so (make all) should only depend on one of them.
Signed-off-by: Miloslav Trmač <mitr@redhat.com>
