- We are modifying existing repo's .treeinfo:
-- Take info about included variants from iso's .treeinfo and put it to repo's .treeinfo
Change-Id: I29bf655d90994e8a1bda40ad04568dd7364f5dca
This is a workaround for modularity design issues and DNF bugs. If there
were gaps in contexts, DNF has trouble handling the upgrades. Thus we
may need to add module metadata for older versions of previously
released module streams and add the missing contexts.
JIRA: RHELCMP-982
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
Until now, the behaviour was that all debuginfo from a build would be
included if at least one package with the same arch was included.
This resulted in many debuginfo packages being included even though
their corresponding package was not present.
With this patch, we will always pull in debugsource, but foo-debuginfo
will only be included if foo is included for the same arch. As a
consequence, it is necessary to resolve dependencies of debuginfo
packages. There are cases where foo-debuginfo needs foo-debuginfo-common
for example.
This change means that DNF and YUM backends are no longer identical in
the output. The tests where this is demonstrated are duplicated and
their results are modified.
JIRA: COMPOSE-3823
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
Make dummy-bash -> dummy-glibc dependency require archful. This avoids
potential race condition where order of dependency processing can result in
different packages being pulled in. The tests where this could happen are
updated.
Make dummy-glibc-debuginfo depend on dummy-glibc-debuginfo-common.
The filenames for the repo no longer include hash, and sqlite databases are not
generated.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
It may make sense to break a big compose into smaller chunks that can be
done independently. This script allows describing the smaller parts,
runs them with correct dependencies and arranges the result to look like
a single big compose.
All parts use the same koji event, that is either obtained from Koji, or
from command line argument.
JIRA: COMPOSE-2654
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
The environment in comps for a variant can refer to groups in parent
variant (either for addons, or because of other configuration). We
should not remove the groups in this case.
This requires changes in two places:
* teaching `comps_filter` about groups that should not be removed
* fixing writing comps so that it does not actually change the data as
well
JIRA: COMPOSE-2612
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
The function loads existing treeinfo, removes reference to boot.iso and
adds [media] section. This is the basic tweak that should happen for all
ISOs. Additional changes depend on the actual content.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
The file is not being included in the source tarball, which breaks
internal jenkins and will eventually cause the test to fail at RPM build
time.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
This will be used by Anaconda to consume multiple repos for
installation.
JIRA: RCM-36970
JIRA: COMPOSE-2753
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
Fixup
If a package name contains leading or trailing whitespace, it will
eventually lead to issues: pungi will try to include that group, but
since it does not exist, the packages will not make it in.
The root cause is hard to find. Better report an error immediately.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
If the parent task is successful, there can still be failed child tasks
for failable arches. We need to log those and potentially mark the
compose as incomplete.
Fixes: https://pagure.io/pungi/issue/874
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
All use cases that are actually used by pungi-koji are tested. There is
missing coverage for
* keeping only items with matching arch
* not reindenting the file
These aren't currently used and should be removed in the future, but
there may be other tools depending on the comps_filter executable.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
When there is a typo in the comps file, instead of crashing with a
non-descript KeyError we should raise a nice error with details about
the problem.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
This adds a new test for Requires(pre) and (post). The general structure
of the test now makes it easy to use the same test scenarios for
different backend.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
When creating unified ISOs, the script will now also create one iso per
architecture containing a repo with debuginfo packages.
There is no switch to turn this off. The images can simply not be
shipped if not wanted.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
This a standalone script that will look into a compose and create
unified ISO for each architecture. The ISO contains RPM repositories for
all variants that have the arch.
Known issues:
* The filename does not respect settings. This is tricky because the
name could include variant name, which we don't have here (by design
of unified ISO).
* The same is true for volume id.
In order to test the feature without running actual compose, we need to
add essentially a big chunk of compose. Most of the files are empty, as
their content is never accessed.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
The check for arch compatibility should not be performed if the
debuginfo package is noarch. Such packages should be included
unconditionally.
Fixes: #450
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
These are tests written by Daniel Mach originally for Distill-NG. They
are ported to current Pungi. The test repositories are committed in the
tests/fixtures/ directory. This is the same data that is used for test
compose, but the actual RPM files are not present. Some tests are
adapted from dmach's fork of Pungi.
Some of the packages are marked with a comment saying they are
important. These are the packages that the test is specifically trying
to get included in the package set. There are also explicit tests for
packages that should not be included.
Two tests are skipped for now as there is a bug preventing them from
passing. This is related to fulltree being done for packages that are
explicitly multilib.
The depsolver is called by invoking a separate executable, so the
coverage data is wrong.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
There is a lot of mock objects needed: we bypass calls to Koji, use a
mock FileCache that does not need valid RPMs on disk and avoid any
multithreading.
The test data in tests/fixtures/tagged-rpms.json comes from Koji. It is
filtered down to only a few packages to make it manageable.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
There already were config options tree_arches and tree_variants, but the
filtering was done at the compose level and there was not interaction
between the two options. This led to problems when a variant would have
all its arches filtered out.
This patch moves the filtering to the variant loader. It adds better
logging, so whenever a variant is filtered (for any reason), it will be
explicitly stated in the logs.
Signed-off-by: Lubomír Sedlář <lsedlar@redhat.com>
