Blueprint package set is now depsolved together with the OS package set
in a chain. The result is stored in the package specs sets under the OS
package set name.
In reality, the code was able to handle a `nil` package specs to be
passed to pipelines, however some parts were looking for the kernel
version in the blueprint package specs, which would be a bug.
Regenerated affected image test cases.
Introduce a new method `PackageSetsChains()` to the `ImageType`
interface, which returns a named lists of package sets, which should be
depolved together in a chain.
Extend all distro implementations with the new method.
Add a unit test ensuring that if an image type defines some package set
name chains, that all of the listed package set names are present in the
package set map returned by the image type.
The method is currently not used anywhere. This is a preparation for
switching from current way of depsolving to the chain depsolving.
Move the `kernelVerStr()` function duplicated in many
distro definitions to the `rpmmd` package as
`GetVerStrFromPackageSpecListPanic()`.
I could not come up with a better name, sorry.
This will prevent creating another copy of the code in rhel-84 for
the `gce` image.
This change initially exposed a bug in the original implementation of
`kernelVerStr()`. Since on the first line, we allocate an empty structure
into `kernelPkg` variable, it can never be `nil` and the function never
panicked even if there was no `kernel` package in the PackageSpec list.
Fix all unit tests to provide valid arguments when calling `Manifest()`
method of image types.
Signed-off-by: Tomas Hozza <thozza@redhat.com>
kernelVerStr fixup
Signed-off-by: Tomas Hozza <thozza@redhat.com>
The service is started via systemd activation sockets.
The service serves http POST requests, the same json as before is
expected as the body of the request, and the same json as before is sent
as the response of the request.
Generated image test case manifests for all supported distros, arches and
image-types are being tested as part of distro unit tests. However due
to time constrains, the unit test does not depsolve the image's default
package sets and thus does not check if they changed in the internal
osbuild-composer's representation, compared to the generated image test
case.
Extend the `TestDistro_Manifest()` function used by the unit test to
allow depsolving image's package sets.
Introduce a new test case binary `osbuild-composer-manifest-tests`
allowing to check the manifests generated by composer for all supported
combinations of images against generated manifests, including depsolving
image's default package sets.
Introduce a new CI test case `manifest_tests.sh` executing the
`osbuild-composer-manifest-tests` binary and testing all existing image
test cases. Run it in CI on RHEL-9 runner.
Modify SPEC file to ship the newly added test case.
Signed-off-by: Tomas Hozza <thozza@redhat.com>
To avoid packages specified in a blueprint from conflicting with exclude
lists, we depsolve blueprint packages separately and pass them into the
Manifest generator under the new "blueprint" package set key.
This approach has the added benefit that dependencies of packages
specified in the blueprint are not subject to exclusion in addition to
the explicitly named packages.
The OS pipeline which installs the packages for the base system merges
the two package sets before running the RPM stage. The signature of the
function is changed to explicitly require blueprint packages be
specified (though `nil` or empty slice is valid).
The kernel selection test is adapted to merge the package sets before
counting kernel package.
Adaptation of changes in
https://github.com/osbuild/osbuild-composer/pull/1349
My goal is to add a method to distroregistry to return Registry with
all supported distributions. This way, all supported distributions
would be defined only on one place.
To achieve this, the Registry must live outside the distro package
because the distro implementation depends on it and this would create
a circular dependency unsupported by Go.
Signed-off-by: Ondřej Budai <ondrej@budai.cz>
This replaces Packages() and BuildPackages() by returning a map of
package sets, the semantics of which is up to the distro to define.
They are meant to be depsolved and the result returned back as a
map to Manifest(), with the same keys.
No functional change.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Rather than setting this automagically, expose it to the caller. For
now the only caller we have simply passes it back in, so this is a
noop.
In follow-up commits this will be used to resolve the parent commit.
This is tested by verifying that the generated manifests do not
change.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Imagine this situation: You have a RHEL system booted from an image produced
by osbuild-composer. On this system, you want to use osbuild-composer to
create another image of RHEL.
However, there's currently something funny with partitions:
All RHEL images built by osbuild-composer contain a root xfs partition. The
interesting bit is that they all share the same xfs partition UUID. This might
sound like a good thing for reproducibility but it has a quirk.
The issue appears when osbuild runs the qemu assembler: it needs to mount all
partitions of the future image to copy the OS tree into it.
Imagine that osbuild-composer is running on a system booted from an imaged
produced by osbuild-composer. This means that its root xfs partition has this
uuid:
efe8afea-c0a8-45dc-8e6e-499279f6fa5d
When osbuild-composer builds an image on this system, it runs osbuild that
runs the qemu assembler at some point. As I said previously, it will mount
all partitions of the future image. That means that it will also try to
mount the root xfs partition with this uuid:
efe8afea-c0a8-45dc-8e6e-499279f6fa5d
Do you remember this one? Yeah, it's the same one as before. However, the xfs
kernel driver doesn't like that. It contains a global table[1] of all xfs
partitions that forbids to mount 2 xfs partitions with the same uuid.
I mean... uuids are meant to be unique, right?
This commit changes the way we build RHEL 8.4 images: Each one now has a
unique uuid. It's now literally a unique universally unique identifier. haha
[1]: a349e4c659/fs/xfs/xfs_mount.c (L51)
require.JSONEqf cannot handle diffs of such a big entity as a manifest is.
It just prints an empty string.
This commit unmarshalls the manifests instead and then uses the cmp library
to make a very nice and readable diff.
Signed-off-by: Ondřej Budai <ondrej@budai.cz>
Allow individual test-cases or sub-sets of test-cases to be generated
more easily.
We allow explicit skipping of image-info generation (and hence the
osbuild run), and also individual image types to be specified.
Also drop distros and image types that are no longer supported.
Signed-off-by: Tom Gundersen <teg@jklm.no>
Rather than Manifest() returning an osbuild.Manifest object, introduce a
new distro.Manifest object which represents it as an opaque, JSON
serializable object. This new type has the following properties:
1) its serialization is compatible with the input to osbuild,
2) any valid osbuild input can be deserialized into it, and
3) marshalling and unmarshaling to and from JSON is lossless.
This means that even as we change the subset of valid osbulid manifests
that we support, we can still load any previous state from disk, and it
will continue to work just as before, even though we can no longer
deserialize it into our internal notion of osbuild.Manifest.
This fixes the underlying problem of which #685 was a symptom.
Signed-off-by: Tom Gundersen <teg@jklm.no>
This is how it is used in the rest of the code, as a name to represent
the repository in the weldr API. Rename to match its use, and avoid
confusion with the ID passed to dnf-json, which is not the same.
Signed-off-by: Tom Gundersen <teg@jklm.no>
When generating an osbuild manifest for an image type, we take a
customizations struct, which specifies the image-type-independent
customizations to apply. We also take the size argument, which is
specific to the image build and not part of the blueprint.
Introduce a new argument ImageOptions, which for now just wraps the size
argument. These options are specific to the image build/type, and
therefore does not belong with the other customizations.
For now this is a non-functional change, but follow-up commits will
introduce more types of image options.
Signed-off-by: Tom Gundersen <teg@jklm.no>
